Teresa Halbach disappeared on Halloween Day, 2005. Her concerned family and friends contacted law enforcement after she did not show up at the photography studio where she worked and her voice mailbox was full. Law enforcement officers quickly zeroed in on the Avery Auto Salvage yard in Two Rivers, Wisconsin, as the *938last place she was known to have gone, and, in particular, on Steven Avery, the son of the salvage yard owner who lived in a trailer on the property. Earlier in the day, Avery called Auto Trader magazine, for whom Halbach sometimes took photographs, to request that she take photographs of a minivan that he wished to sell in its magazine. Eventually the police began to suspect that Avery’s 16-year-old nephew, Brendan Dassey, who also lived on the property, might have been a witness or had information about Halbach’s murder. After a few preliminary conversations, the investigators were concerned enough to call Dassey into the police station for a full interrogation. After many hours of questioning and interrogation spread over several days, Dassey confessed that he, along with Avery, had raped and brutally murdered Halbach and then burned her body in an on-site fire pit. By the time of the trial, Dassey had recanted his confession, and the State had failed to find any physical evidence linking him to the crime, but he was convicted and sentenced to life in prison nonetheless. After appeals and post-conviction proceedings in the state court failed to bring him relief. The state court on post-conviction review stated the generalized standard for evaluating the voluntariness of a confession— totality of the circumstances—but failed to note how that juvenile confession requires more care and failed to apply the standard at all. Dassey filed a petition for a writ of habeas corpus in the district court, claiming that he did not receive effective assistance of counsel and that his confession was not voluntarily given. The district court, concluding as we do that the state court did not apply the proper standard, granted the writ. Despite the limited role of a federal court on habeas review we must affirm. If a state court can evade all federal review by merely parroting the correct Supreme Court law, then the writ of habeas corpus is meaningless.
I.
The facts related to this case are expansive and convoluted, and those facts have been reported in various iterations throughout the decisions of the state courts of Wisconsin and in the district court. We borrow heavily from the district court and report just those facts needed for purposes of this appeal and refer the reader to the full district court opinion, Dassey v. Dittmann, 201 F.Supp.3d 963 (E.D. Wis. 2016) for further details.
Teresa Halbach was a 25-year-old sum-ma cum laude graduate of the University of Wisconsin-Green Bay who was running her own photography business. She was the second oldest of five children in a tight-knit family, and lived in a farmhouse a quarter mile from her parents. On October 31, 2005, she photographed three vehicles for Auto Trader Magazine. She took the third and final series of photographs at the Avery salvage yard. She never returned home. Her life and career were cut short by a heinous and senseless crime.
Her brutally burned body provided few clues about her death, but other investigative methods provided the state court with the following facts. Halbach had taken photographs at the Avery property on five prior occasions, and Avery called Auto Trader the morning of October 31 and requested that “the same girl who had been out here before” come and take pictures of a vehicle that was for sale. Just before 2:30 p.m., Halbach contacted Auto Trader Magazine and said that she was on her way to the Avery property. Sometime around 2:30 or 2:45 p.m., a neighbor of Avery’s saw Halbach photographing a minivan and then proceed toward Avery’s residence. The neighbor left home at about 3:00 p.m. and observed Halbach’s 1999 Toyota RAV4 still outside Avery’s residence but did not see Halbach. When he returned home at approximately 5:00 p.m., *939Halbach’s RAV4 was gone. Halbach was not seen or heard from after that time.
On November 5, 2005, volunteer searchers scoured the forty acre, 4,000 + vehicle salvage yard and found Halbach’s RAV4 partially covered by tree branches, fence posts, boxes, plywood, and auto parts. The license plates had been removed and the battery cables disconnected.
Based on that discovery, investigators obtained a search warrant for the entire salvage yard and, after a week-long search, found evidence that Halbach was the victim of a horrendous crime. Some of that evidence came from a burn barrel and a four-foot by six-foot burn pit near Avery’s trailer. In those burn areas, investigators found Halbach’s charred bone and dental remains, burned remnants of a cell phone and camera of the same make and model that Halbach used, and a zipper and rivets from a brand of women’s jeans that Halbach was known to wear. State crime lab experts later determined, based on the skull fragments, that Halbach had been shot twice in the head. Multiple witnesses reported seeing a large bonfire in the burn pit outside of Avery’s residence on October 31. The police arrested Avery after the discovery of this evidence.
Forensic investigators found a roughly six-inch blood stain in the rear cargo area of Halbach’s RAV4, and other smaller stains in and around the cargo area that matched Halbach’s DNA. Also in the RAV4, forensic examiners found very small blood stains that matched Avery’s DNA profile on the following locations: a panel just to the right of the ignition, a CD case, a metal panel between the rear seats and the vehicle cargo area, the driver’s seat, the front passenger’s seat, and the floor next to the center console. Avery’s DNA was also detected on the hood latch.
The investigation of Avery continued as he awaited trial. Investigators began interviewing family members, including Dassey and Avery’s niece, Kayla Avery. Kayla stated that her cousin Brendan Dassey had been “acting up lately,” that he was staring into space and crying uncontrollably, and that he had lost roughly forty pounds. Dassey later explained that the weight loss had been part of an effort to find a girlfriend and that the tears had been over a break up. But based on Kayla’s interview, and the fact that another witness reported seeing Dassey at the bonfire with Avery around 7:30 or 7:45 p.m. on October 31, investigators decided that it was necessary to re-interview Dassey.
Calumet County Sheriffs investigator, Mark Wiegert, and Wisconsin Department of Justice Special Agent, Tom Fassbender, travelled to Dassey’s high school on February 27, 2006, and, without his parents’ knowledge, met with him in a conference room for about an hour. Dassey was a sophomore who received special education services, and whose IQ had been measured at various times between 74 and 81, falling fairly far below an average range of intelligence. On the Wechsler scale of intelligence, Dassey’s score meant that 90% of adolescents his age would have performed intellectually better than he did, and on the Kaufman scale, 87% of adolescents his age would have performed better. R. 19-22 at 48-49. A psychological expert at trial described Dassey as highly suggestible, docile, withdrawn, with extreme social anxiety and social avoidant characteristics, and more suggestible than 95% of the population.
At that first interview with the officers, Dassey said that Avery had asked him to help load tires and an old van seat onto a bonfire near Avery’s trailer on the evening of October 31, but that he saw nothing unusual before going home. Because of the poor quality of the cassette tape recording of that interview, the prosecuting attorney requested that the investigators re-inter*940view Dassey to create a better record. Wiegert and Fassbender made arrangements to interview Dassey again later that same day at the local police station.
Wiegert and Fassbender contacted Das-sey’s mother, Barbara Janda, who met them at the school. The investigators drove Dassey and Janda to the police station. According to Wiegert and Fassben-der, Janda declined their offer to be present for the interview and instead remained in a waiting area of the police station. R. 19-19 at 71. According to Janda, the investigators discouraged her from attending the interview. R. 19-30 at 155. This second February 27 interview, which lasted less than an hour, began with a long monologue by Fassbender, who sat down with Dassey and said, “some people back there say no, we’ll just charge him. We said no, let us talk to him, give him the opportunity to come forward with the information that he has, and get it off his chest.” R. 19-24 at 5. Then, Fassbender set forth his role in the investigation and made what Dassey characterizes as the first of many assurances and promises:
Mark and I, yeah, we’re cops, we’re investigators and stuff like that, but I’m not right now. I’m a father that has a kid your age too. There’s nothing I’d like more than to come over and give you a hug cuz I know you’re hurtin.’ Talk about it ... I promise I will not leave you high and dry.
R. 19-24 at 5. After this assurance, Dassey began what would become a series of alterations in his story over time, increasing his culpability in response to suggestions by the investigators. The first such suggestion ■came after Dassey initially denied having seen anything but garbage and other detritus in the October 31 fire. The investigators insisted that Dassey must have seen something suspicious in the fire. Fassben-der set forth his suspicions as follows:
I’m more interested in what you probably saw in that fire or something. We know she was put in that fire, there’s no doubt about it. The evidence speaks for itself. And you were out there with him. And unfortunately, I’m afraid you saw something that you wished you never would have seen. You know, I mean that’s what we need- to know.... Did you see a hand, a foot, something in that fire? Her bones? Did you smell something that was not too right?
Id. at 5-6. Then, after Fassbender insisted several times that Dassey must have seen something in the fire, and suggesting the body parts that he had seen, Dassey admitted that he had seen those same body parts—fingers and toes, plus a forehead, and a belly in the fire. By the end of this interview, Dassey reported that he saw Halbach’s body parts in a fire, that he saw Avery burn clothing in a fire, and that Avery had confessed that he had stabbed Halbach, put her in the fire and hid her car in the yard.
Fassbender met with Dassey again that evening in a hotel room where Dassey told Fassbender, in an unrecorded interview, that he had stained his pants with bleach as he helped clean the floor of Avery’s garage. Wiegert testified that after those interviews he thought Dassey might have had some culpability in the criminal disposal of Halbach’s corpse. R. 19-12 at 18-21; R. 19-30 at 38.
On March 1, 2006, the officers returned to Dassey’s school for a fourth interview. They read Dassey his Miranda rights, and he again agreed to speak with them. Wie-gert and Fassbender first drove Dassey to his house on the Avery property to re*941trieve the bleach-stained jeans and then drove him forty-five minutes away to the Manitowoc County Sheriffs Department. The State asserted that it asked Janda for permission to interview her son. R. 19-19 at 12; 19-30 at 156. Janda claimed that the investigators never asked her if she wanted to be present for the interview. R. 19-30 at 156. This fourth interview produced a confession that became the key evidence against Dassey at his trial.
The March 1 interview lasted three hours, with one half-hour break, and then a second fifty-minute break at the end before Dassey was taken into custody. The interrogation was conducted in what is known as a “soft room” in the Sheriffs Department—one with a small couch, two soft chairs and lamps. Dassey was offered food, drink, and access to a restroom at the start and at various times throughout the interview. The investigators reminded Dassey of his Miranda rights, and the interview was audio and video recorded. No adult was present on Dassey’s behalf.
Dassey’s March 1 confession unfolded as follows in this very brief summary: Dassey first admitted only to helping Avery clean some fluid from the garage floor after Avery cut a line of the vehicle on which he was working. Eventually, after much encouragement, the story evolved to one in which Dassey saw Halbach’s already dead, clothed, and tied up body in the back of her RAV4 and helped Avery put her body in a bonfire. In the next iteration, he reported hearing screaming at Avery’s house as he brought Avery his mail. He entered and found a sweaty Avery and saw Hal-bach naked and handcuffed to Avery’s bed. Finally, Dassey admitted to a horrific series of crimes—raping Halbach, cutting her throat, tying her up, cutting her hair, and then taking her to the garage where Avery shot her in the head and the two of them disposed of her body in the fire. Although we report the evolution of his confession linearly, it is far from that. Dassey’s story changes; he backtracks; officers try to pin him down on time frames and details, but they are like waves on the sand. Even the State has trouble telling its version of the timeline of the story in any cogent manner due to the fact that it changed with each re-telling. See Brief of Respondent-Appellant at 9, n.3. Although the State presents a cogent story line in its brief on appeal, it does so by picking and choosing pieces from various versions of Dassey’s recitations.
At the very end of the confession, Das-sey’s mother entered the interrogation room and the following exchange occurred after the officers left the room:
Brendan: I got a question?
Barb Janda: What’s that?
Brendan: What’d happen if he says something his story’s different? Wh-he says he, he admits to doing it?
Barb Janda: What do you mean? Brendan: Like if his story’s like different, like I never did nothin’ or somethin’.
Barb Janda: Did you? Huh?
Brendan: Not really.
Barb Janda: What do you mean not really?
Brendan: They got to my head.
R. 19-25 at 148. At that point, one of the officers reentered the room and the conversation ended. We will fill in the remaining details of this confession as we discuss the voluntariness of ⅜ vel non, in the following sections.
Almost the entirety of the State’s case rested on these interviews and one phone call between Dassey and his mother after his final police interview which we describe below. There was no physical evidence linking Dassey to the murder of Halbach— investigators did not find any of Dassey’s DNA or blood on any of the many objects that were mentioned in his confession—the *942knives in Avery’s house, gun, handcuffs, bed, RAV4, key, or automotive dolly.
After his arrest, the state -public defender’s office appointed private attorney Len Kachinsky to represent Dassey. Kachinsky met with Dassey on March 10, 2006. Das-sey told Kachinsky that he was innocent, that his confession was not true, and that he wanted to take a polygraph test. After this meeting, despite Dassey’s claims of innocence, Kachinsky spoke to the media and described Dassey as sad, remorseful, and overwhelmed. The media reported that Kachinsky blamed Avery for “leading Dassey down the criminal path” and said that he had not ruled out a plea deal. R. 19-39 at 4, 9-11. Over the next few days, nearly all of Kachinsky’s work on Dassey’s case involved communicating with the local media, during which appearances he stated that “there is quite frankly, no defense,” and that all of the investigation techniques were standard and legitimate, despite the fact that Kachinsky had not yet watched the recorded police interview R. 19-26 at 142, 144-45, 153, 170. During each of Ka-chinsky’s media appearances he indicated that Dassey was guilty and would likely accept a plea. Kachinsky testified at a post-conviction relief hearing that one of his reasons for making these statements to the media was so that Dassey and his family would become “accustomed to the idea that Brendan might take a legal option that they don’t like.... ” R. 19-26 at 136-37. Eventually the prosecutor sent an email to Kachinsky expressing concern about the pretrial media appearances and referred Kachinsky to the relevant rules of ethics for attorneys.
In the meantime, Kachinsky hired investigator Michael O’Kelly, with whom he was not familiar, to help in the investigation of the case and to conduct the polygraph examination that Dassey had requested. Despite Dassey’s claims of innocence, Ka-chinsky and O’Kelly proceeded on the assumption that Dassey would plead guilty and assist the prosecution in Avery’s case. O’Kelly testified at the state post-conviction hearing that his goal was to uncover information and evidence that would bolster the prosecution’s case against Avery even if that “evidence would tend to inculpate Brendan,” R. 19-29 at 47, and that his “emotions sided with what happened to Teresa Halbach.” Id. at 96. Kachinsky and O’Kelly even sent information to the prosecution about the location of a knife they thought had been used in the crime, based on what they had cajoled from Dassey, but searches pursuant to those tips did not produce any evidence.
To effectuate his plan to garner Das-sey’s cooperation in Avery’s prosecution, Kachinsky decided that the investigator, O’Kelly, should re-interview Dassey and compel him to confess yet again, and should do so after the trial judge denied the motion to suppress his March 1 interview, when he would be most vulnerable. R. 19-26 at 244.
Shortly before interviewing Dassey, O’Kelly wrote to Kachinsky and referred to the Avery family as “criminals” and asserted that family members engaged in incestuous sexual conduct and had a history of stalking women. R. 19-29 at 93. He continued, “This is truly where the devil resides in comfort. I can find no good in any member. These people are pure evil.” Id. O’Kelly quoted a friend as having said, “This is a one branch family tree. Cut this tree down. We need to end the gene pool here.” Id. at 94. O’Kelly thought that Das-sey’s claim of innocence was an “unrealistic” “fantasy” that was influenced by his family. R. 19-29 at 83, 84, 86-88. On O’Kelly’s recommendation, Kachinsky canceled a planned visit with Dassey because Dassey “needs to be alone.” R. 19-26 at 248^49. O’Kelly said, “He needs to trust me and *943the direction that I steer him into.” R. 19-26 at 249.
O’Kelly began his interview with Das-sey, which he video recorded without permission from Dassey’s parents, by pointing to what he said were the polygraph examination results on a laptop computer screen and asking Dassey if he could read them. R. 19-38 at 1. Despite having previously told Kachinsky that the results of the polygraph examination were inconclusive R. 19-26 at 210,2 O’Kelly told Dassey that the polygraph indicated deception and that the probability of deception was 98%. R. 19-38 at 1. When Dassey asked what that meant, O’Kelly asked what he thought it meant. R. 19-38 at 1. Dassey responded, “That I passed it?” R. 19-38 at 1. “It says deception indicated,” O’Kelly responded, emphasizing “deception.” Id. After a long pause, Dassey asked, “That I failed it[?j” Id.
O’Kelly proceeded to harangue Dassey with photographs and personal effects of Halbach, threaten him with life in prison, and badger him to admit that he was sorry. Dassey continued to profess his innocence, insisting, “I don’t know [if I’m sorry], because I didn’t do anything,” to which O’Kelly responded, “If you’re not sorry, I can’t help you ... Do you want to spend the rest of your life in prison? You did a very bad thing.” R. 19-38 at 2. Das-sey responded, “Yeah, but I was only there for the fire though.” Id.
Eventually O’Kelly’s plan prevailed after he convinced Dassey that if he confessed he would be sentenced to only twenty years in prison and could someday be released and have a family. (The government had not, in fact, placed any plea deal on the table.) Otherwise, O’Kelly threatened, Dassey would go to prison for the rest of his life. After a grueling interrogation by O’Kelly, Dassey confessed, providing yet another version of the story. O’Kelly immediately telephoned Kachinsky who arranged for Dassey to undergo another police interrogation the next day, May 13. Kachinsky did not arrange for any immunity agreements, plea offers, or other safeguards. In fact, he agreed that the State would provide “no consideration” in exchange for a second chance to interrogate (the police considered this to be only the second interrogation because they considered the first few meetings to be “witness interviews.”) R. 19-26 at 80; R. 19-27 at 34-38. Kachinsky did not accompany Dassey to this meeting and allowed him to be interrogated without counsel. That interview differed in many significant ways from the story Dassey told on March 1, but it was never admitted or used at trial.
At the end of the May 13 interview, Fassbender and Wiegert advised Dassey that he should call his mother over the recorded jail telephone line and admit his guilt so that she would hear it from him first rather than from the officers. Das-sey’s mother was scheduled to visit him the following day, but the investigators told him that it would be a “good idea to *944call her before she gets here, tonight. That’s what I’d do. Cuz, otherwise she’s going to be really mad tomorrow. Better on the phone, isn’t it?” R. 19-34 at 69. The contents of that telephone call are set forth in the district court opinion. Dassey v. Dittmann, 201 F.Supp.3d at 980-81. In that call, Dassey explained why he was confessing (for a lower sentence), told his mother that he did “some of it” but denied having sexual contact with Halbach, denied seeing her in the fire, denied knowing if Avery Wiled Halbach but asked, “So if I was in the garage cleaning up that stuff on the floor, how much time will I get though for that?” R. 19-35 at 8. He described the liquid on the floor as “reddish-black stuff.” Id.
When the trial court learned that Ka-chinsky had allowed Dassey to be interviewed without counsel, it held a hearing on the effectiveness of Kachinsky’s counsel. The trial court concluded that Kachin-sky’s performance was indefensible and deficient under the standards set forth in Strickland v. Washington, 467 U.S. 1267, 104 S.Ct. 3562, 82 L.Ed.2d 864 (1984). The trial judge decertified Kachinsky from being appointed in most felony matters going forward, noting particularly the egregiousness of the fact that Kachinsky had “allowed his 16-year-old client, who previous testimony has disclosed to have cognitive ability within borderline to below average range, to be interviewed by law enforcement officials without his attorney present.” R. 19-14 at 22. The decertification was prospective only and thus did not directly apply to Kachinsky’s representation of Dassey. Nevertheless, Kachinsky moved to withdraw as Dassey’s counsel, and the court granted the motion.
The trial court never learned that Ka-chinsky and O’Kelly had worked to compel Dassey’s confession, videotaped O’Kelly interrogating Dassey, exchanged e-mails describing the whole family as “evil” and “criminals,” and, without Dassey’s knowledge or consent, sent an e-mail to prosecutors on May 5 indicating where they thought the murder weapon was hidden. No murder weapon was ever found. These facts did not come to light until the state post-conviction hearing.
The May 13 interrogation that grew from the poisoned tree of the O’Kelly interrogation was neither used nor discussed at trial, but the trial court never made any explicit ruling on its admissibility. At oral argument the State was unable to tell this court why the May 13 interview was not used at trial, but we will assume that based on what the State concedes was unacceptable representation by Kachinsky, the State recognized that the May 13 interview had been irreparably poisoned. But the May 13 phone call that resulted from the May 13 interrogation—the phone call the police had urged Dassey to make to his mother on the recorded jail telephone line—was used three times at trial: once to cross examine Dassey; once to cross-examine Dassey’s expert psychologist, and in closing argument to undermine Dassey’s alibi.
At trial, the centerpiece of the prosecution’s case was Dassey’s March 1 confession, in which he admitted to participating in the alleged sexual assault and murder of Halbach as well as the disposal of her body. Dassey’s defense was that his confession was not true or voluntary, that he accepted his uncle’s invitation to a bonfire and then helped him gather items from the salvage yard to burn before helping Avery clean up something that looked like automotive fluid from the garage floor, staining his pants with bleach in the process. Das-sey testified that he did not know why he had said the things that he did to the police investigators and that he thought that the investigators had promised that *945he would not go to jail no matter what he told them.
At trial, Dassey’s attorneys presented evidence that the answers in his confession came not from Dassey, but from ideas planted by the investigators, that the investigators continually linked the idea that if Dassey gave them the answers they wanted to hear, that he would be okay and set free, and that Dassey was extremely suggestible and would say things to please the investigators and avoid conflict.3 One example that the jury saw, as they watched the four hour interrogation, concerned Halbach’s shooting. By the time of the March 1 confession, forensic examiners had informed law enforcement that Hal-bach had been shot in the head, but this information was not yet public. If Dassey could tell the investigators that Halbach had been shot in the head, it would have been strong evidence of the veracity of his confession. Dassey had never mentioned that Halbach was shot. Consequently, the investigators repeatedly asked Dassey what else happened to Halbach. After many, many attempts at this, they became more specific and asked “What else did he do to her? ... Something with the head.” R. 19-25 at 60. But even this clue was not enough to elicit the information they wanted and instead triggered a litany of apparent guesses from Dassey that bordered on the absurd. Dassey guessed that her hair had been cut, that she had been punched, that her throat had been cut—each time being told by the investigators that was not what they were looking for, until finally, Wiegert became frustrated and asked, “All right, I’m just gonna come out and ask you. Who shot her in the head?” Id. at 63. This was one of the few scenarios that Dassey had not guessed at that point. As we will explore below, this pattern of suggestive questioning continued throughout the inteiTogation.
The defense also presented the testimony of a forensic psychologist, Dr. Robert Gordon, who testified that he reviewed many years of Dassey’s school records, performed a mental status examination of Dassey, and tested Dassey using various established psychological tests. R. 19-22 at 23-166. His ultimate conclusion was that Dassey had several characteristics likely to make him unusually suggestible in interrogation situations. Dr. Gordon described Dassey’s thought process as slow with a mild to moderate mental impairment. His test results demonstrated that Dassey performed on the extreme ends of the scales for social avoidance (being socially passive and withdrawn), social introversion, and social alienation (alienated from society and cut off from those with whom he interacts). Dassey scored in the 99th percentile for social avoidance, the 97th percentile for social introversion and 98.5th percentile for social alienation. On other tests, Das-sey’s results indicated that he was shy, passive, subdued and dependent—qualities that make one more susceptible to suggestion. Dr. Gordon also testified that Dassey had low average to borderline intelligence (IQ tests ranged from the low 70s to 84, or in the 10-13% percentile of intelligence). Gordon also administered the Gudjonsson Suggestibility Scales, a test developed by a forensic psychologist and a leading expert in confessions, which is designed to measure interrogative suggestibility. The results indicated that Dassey was more suggestible than 95% of the population. Dr. Gordon also explained how, based on all of his characteristics, Dassey would have been manipulable and vulnerable to the particular interrogation techniques used, including mild pressure and leading questions. He noted that a suggestible person would be particularly swayed by false information of guilt, minimization of the seri*946ousness of the crime, blaming other participants for their influence, or promises that family members will be spared trouble if the suspect confesses. Id. at 62. In a short rebuttal, the State presented psychologist Dr. James Armentrout, who expressed discomfort with the suggestibility testing and did not agree with the conclusion that Das-sey was particularly suggestible. Id. at 177-225.
After five and a half hours of deliberation, the jury found Dassey guilty on all counts. On August 2, 2007, the trial court sentenced Dassey to life in prison for first-degree intentional homicide, not eligible for release to extended supervision until November 1, 2048. R. 19-2 at 15-16. The court further sentenced Dassey to six years of imprisonment for mutilating a corpse, and fourteen years imprisonment for second-degree sexual assault, both to be served concurrently with the murder sentence. Id.) Dassey v. Dittmann, 201 F.Supp.3d at 985. Dassey appealed his conviction without success.
Dassey moved for post-conviction relief in the trial court claiming that his pre-trial and trial counsel provided ineffective assistance and that his March 1 confession was involuntary. Upon his motion, the Wisconsin state court held a five-day hearing, beginning January 15, 2010, which included the testimony of Dassey’s mother, his school psychologist, one of his trial attorneys, the prosecutor, a social psychologist, Kachinsky, O’Kelly, and Richard Leo, an expert on false confessions. The circuit court of Wisconsin denied Dassey post-conviction relief on December 13, 2010.
On appeal of the post-conviction ruling, the Wisconsin Court of Appeals stated that it was evaluating Dassey’s claim of involuntariness on the totality of the circumstances, “balancing the defendant’s personal characteristics against the police pressures used to induce the statements.” State v. Dassey, No. 2010AP3105, 2013 WI App 30, ¶ 5, *1, 2013 WL 335923 at *1, 346 Wis.2d 278,827 N.W.2d 928 (table) (Wis. Ct. App., Jan. 30, 2013).4 That evaluation boiled down to just a few sentences in the following two paragraphs:
¶ 6 The trial court found that Dassey had a “low average to borderline” IQ but was in mostly regular-track high school classes; was interviewed while seated on an upholstered couch, never was physically restrained and was offered food, beverages and restroom breaks; was properly Mirandized; and did not appear to be agitated or intimidated at any point in the questioning. The court also found that the investigators used normal speaking tones, with no hectoring, threats or promises of leniency; prodded him to be honest as a reminder of his moral duty to tell the truth; and told him they were “in [his] corner” and would “go to bat” for him to try to achieve a rapport with Dassey and to convince him that being truthful would be in his best interest. The court concluded that Dassey’s confession was voluntary and admissible.
¶ 7 The court’s findings are not clearly erroneous. Based on those findings, we also conclude that Dassey has not shown coercion. As long as investigators’ statements merely encourage honesty and do not promise leniency, telling a defendant that cooperating would be to his or her benefit is not coercive conduct. State v. Berggren, 2009 WI App 82, ¶ 31, 320 Wis.2d 209, 769 N.W.2d 110. Nor is professing to know facts they actually did not have. See State v. Triggs, 2003 WI App 91, ¶¶ 15, 17, 264 Wis.2d 861, 663 N.W.2d 396 (the use of a deceptive tactic *947like exaggerating strength of evidence against suspect does not necessarily make confession involuntary but instead is a factor to consider in totality of circumstances). The truth of the confession remained for the jury to determine.
State v. Dassey, 2013 WL 335923 at *2. Although the state appellate court listed Dassey’s characteristics and some of the circumstances of his interrogation, as we will describe in detail below, it did not do the one thing that the Supreme Court requires which is to use “special caution” when assessing the voluntariness of juvenile confessions. J.D.B. v. North Carolina, 564 U.S. 261, 269, 131 S.Ct. 2394, 180 L.Ed.2d 310 (2011); In re Gault, 387 U.S. 1, 45, 87 S.Ct. 1428, 18 L.Ed.2d 527 (1967); Gallegos v. Colorado, 370 U.S. 49, 53-54, 82 S.Ct. 1209, 8 L.Ed.2d 325, (1962); Haley v. Ohio, 332 U.S. 596, 599-601, 68 S.Ct. 302, 92 L.Ed. 224 (1948). Paragraph 6 of the appellate court decision lists Dassey’s age and intellectual limitations, but then, in paragraph 7, the only paragraph that analyzes whether Dassey’s confession was voluntary or coerced, it merely applies the same analysis that would apply to an adult with full intellectual capabilities. Specifically, the state appellate court concluded that tactics such as encouraging honesty and the use of deceptive practices that are not considered coercive when used with adults must not have been coercive when used on the intellectually challenged, 16-year-old Dassey. A state court’s evaluation need not be lengthy or detailed, but it must at the very least meet the bare minimum requirements of Supreme Court precedent. The admonition to assess juvenile confession with special caution has no meaning if a state appellate court can merely mention a juvenile’s age and then evaluate the volun-tariness of his confession in reference to the standard for adults of ordinary intelligence. And if a court can merely state the generic Supreme Court rule without any analysis, then no federal court could ever
find that “a decision ... involved an unreasonable application of clearly established Federal law” pursuant to 28 U.S.C. § 2254(d)(1).
In juveniles, the evaluation of the totality of the circumstances “includes evaluation of the juvenile’s age, experience, education, background, and intelligence, and into whether he has the capacity to understand the warnings given him, the nature of his Fifth Amendment rights, and the consequences of waiving those rights.” Fare v. Michael C., 442 U.S. 707, 725, 99 S.Ct. 2560, 61 L.Ed.2d 197 (1979); see also Murdock v. Dorethy, 846 F.3d 203, 209 (7th Cir. 2017); Hardaway v. Young, 302 F.3d 757, 762 (7th Cir. 2002). At no time did the state appellate court evaluate any of these factors, other than to merely list some of them. It did not provide any analysis of how Dassey’s personal characteristics played a role in the interrogation. It did not consider Dassey’s suggestibility, did not discuss the fact that he was unrepresented and without a parent’s assistance, and it did not consider whether Dassey’s low IQ and learning disabilities may have affected how he interpreted statements made by interrogators. The court never evaluated Dassey’s capacity to understand the warnings given him, the nature of his Fifth Amendment rights, and the consequences of waiving those rights. In short, the state appellate court did not identify the correct test at all and did not apply it correctly.
The state appellate court also declined to overrule the lower court’s decision denying Dassey’s claim of ineffective assistance of counsel. As for Kachinsky’s conceded deficiencies, the court stated that he was “long gone before Dassey’s trial or sentencing. Dassey has not convinced us that Kachinsky’s actions amounted to an actual conflict and that Kaehinsky’s advocacy was adversely affected, such that it was detri*948mental to Dassey’s interests.” Id. at *4. And in reference to trial counsel’s performance, the appellate court held that the trial court had not erred when it determined that each of Dassey’s claims of ineffective assistance of trial counsel was based on his attorneys’ reasonable tactical strategies. Id. at *6.
After the Wisconsin Supreme Court denied his petition for review, Dassey filed a petition for a writ of habeas corpus in the federal district court pursuant to 28 U.S.C. § 2254, claiming that he was denied his rights to effective assistance of counsel under the Sixth Amendment of the United States Constitution, and that his March 1, 2006 confession was obtained in violation of the Fifth Amendment. The district court concluded that although Kachinsky’s misconduct might support a claim for relief under Strickland, Dassey made his claims regarding Kachinsky under Cuyler v. Sullivan, 446 U.S. 335, 100 S.Ct. 1708, 64 L.Ed.2d 333 (1980), and case law demarcating the limits of the Sullivan test prohibit the court from granting Dassey’s ha-beas relief claim on that ground. Dassey, 201 F.Supp.3d at 991-92. It further concluded that the state court of appeals’ decision as to the admissibility of the May 13 telephone call between Dassey and his mother was not contrary to clearly established federal law or based on an unreasonable determination of the facts. Id. at 992. However, the district court concluded that “the confession Dassey gave to the police on March 1, 2006 was so clearly involuntary in a constitutional sense that the court of appeals’ decision to the contrary was an unreasonable application of clearly established federal law,” and that the admission of the confession was not harmless error. Id. at 1005-06. The district court ordered the State to release Dassey' from custody unless, within 90 days, the State initiated proceedings to retry him. Id. at 1006. On November 17, 2016, this court stayed the district court’s order releasing Dassey pending resolution of this appeal. Court of Appeals Record, R. 22.
II.
A. The AEDPA and habeas relief.
The Antiterrorism and Effective Death Penalty Act of 1996 governs our review of a state court conviction and limits it considerably. It “erects a formidable barrier to federal habeas relief for prisoners whose claims have been adjudicated in state court, requiring them to show that the state court’s ruling ... was so lacking in justification that there was an error ... beyond any possibility for fair minded disagreement.” Burt v. Titlow, — U.S. —, 134 S.Ct. 10, 12, 187 L.Ed.2d 348 (2013). “[W]e may not grant relief where reasonable minds could differ over the correct application of legal principles, and we must evaluate that application on the basis of the law that was ‘clearly established’ at the time of the state court adjudication.” Elmore v. Holbrook, — U.S. —, 137 S.Ct. 3, 7, 196 L.Ed.2d 272 (2016). A federal court reviewing a habeas petition must examine the decision of the last state court to rule on the merits of the issue, which in this case is the state appellate court ruling on post-conviction relief. Makiel v. Butler, 782 F.3d 882, 896 (7th Cir. 2015).
Under the AEDPA, Dassey must demonstrate that the state court proceedings “(1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or (2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d)(1) and (2). Under § 2254(d)(1), a state-court decision is contrary to Supreme Court precedent if it is inconsistent with the Supreme *949Court’s treatment of a materially identical set of facts, or if the state court applied a legal standard that is inconsistent with the rule set forth in the relevant Supreme Court precedent. Bell v. Cone, 535 U.S. 685, 694, 122 S.Ct. 1843, 152 L.Ed.2d 914 (2002) (citing Williams v. Taylor, 529 U.S. 362, 405-06, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000)). And a state-court decision constitutes an unreasonable application of Supreme Court precedent within the meaning of section 2254(d)(1) when, although it identifies the correct legal rule, it applies that rule to the facts in a way that is objectively unreasonable. White v. Woodall, — U.S. —, 134 S.Ct. 1697, 1705, 188 L.Ed.2d 698 (2014).
Under § 2254(d)(2), a state court’s decision involves an unreasonable determination of the facts if it “rests upon fact-finding that ignores the clear and convincing weight of the evidence.” Corcoran v. Neal, 783 F.3d 676, 683 (7th Cir. 2015), cert. denied, — U.S. —, 136 S.Ct. 1493, 194 L.Ed.2d 589 (2016); see also Miller-El v. Cockrell, 537 U.S. 322, 340, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003) (a federal court can, guided by AEDPA, conclude that a state court’s decision was unreasonable or that the factual premise was incorrect by clear and convincing evidence).
In granting the writ, the district court specifically noted that it did not reach its conclusion to declare the state court ruling unreasonable lightly. It was, as we are, mindful of the extremely restricted nature of habeas relief under the AEDPA, and that mindfulness was apparent from the great care the district court took in conscribing its ruling to the limited role a federal court can play in reviewing the petitioner’s writ. Dassey v. Dittmann, 201 F.Supp.3d at 986-87, 1005. The district court exhaustively surveyed Supreme Court precedent and continuously held its analysis up to the light of habeas restraint. See Id. at 986-87, 990-91, 1003-05. “Deference,” however, “does not by definition preclude relief.” Miller-El v. Dretke, 545 U.S. 231, 240, 125 S.Ct. 2317, 162 L.Ed.2d 196 (2005). Section 2254(d)(1) allows for a grant of relief when a decision involved an unreasonable application of clearly established Federal law. And if that section has any meaning, then it must mean that a state court evaluating the voluntariness of a juvenile confession must apply the factors that the Supreme Court has identified as relevant to juvenile confessions.
Moreover, the district court’s grant of the writ was firmly linked to its determination under § 2254 (d)(2) that “the state court’s finding that there were no promises of leniency was against the clear and convincing weight of the evidence.” Dassey v. Dittmann, 201 F.Supp.3d at 1003 (internal citations omitted). “Concluding that the investigators never made any such promises was no minor error but rather a fact that was central to the court’s voluntariness finding.” Id. The district court found that the determination was not merely incorrect, but unreasonable. Id. Secondly, the court concluded that the state court had unreasonably applied clearly established federal law by ignoring the totality of the circumstances in assessing the voluntariness of Dassey’s confession. Id. at 1004. The district court noted that although the state appellate court articulated the correct standard (but only as it applied to adults), it ignored several determinative factors outright and, most importantly, focused on the statements of the investigators in isolation rather than assessing them in view of Dassey’s personal characteristics or their cumulative effect on the voluntariness of Dassey’s confession. Id. at 1004.
We, like the district court, have kept the strict constraints of the AEDPA forefront in our minds as we proceed with our de novo review of the district court’s decision *950to grant the habeas petition. Rodriguez v. Gossett, 842 F.3d 531, 537 (7th Cir. 2016).
Yet even given the constraints of the AEDPA, we must conclude that the state court’s determination was an unreasonable application of Supreme Court precedent. Although it identified the general rule that a court must consider the totality of the circumstances, it failed to apply the “special caution” required in juvenile confessions and failed to evaluate the totality factors for juveniles as required. Furthermore, the state appellate court applied the generic totality of the circumstances test to the facts in a way that was objectively unreasonable. See 28 U.S.C. § 2254(d)(1). The trial court’s determination of the facts was also unreasonable as it ignored the clear and convincing weight of the evidence. See 28 U.S.C. § 2254(d)(2); Miller-El v. Cockrell, 537 U.S. at 340, 123 S.Ct. 1029. Although the state appellate court noted that it was obligated to consider the totality of the circumstances, it did not do so. As we noted, in juveniles, the evaluation of the totality of the circumstances “includes evaluation of the juvenile’s age, experience, education, background, and intelligence, and into whether he has the capacity to understand the warnings given him, the nature of his Fifth Amendment rights, and the consequences of waiving those rights.” Fare, 442 U.S. at 725, 99 S.Ct. 2560; see also Murdock, 846 F.3d at 209; Hardaway, 302 F.3d at 762. The state appellate court listed Dassey’s age, education and IQ, but it never, at any point, evaluated those factors to determine whether they affected the voluntariness of Dassey’s confession. Likewise the appellate court analyzed some of the investigators’ interrogation techniques, but it never evaluated or assessed how those techniques affected the voluntariness of an intellectually challenged juvenile’s confession. Instead, the state appellate court merely stated that, in eases involving adults of ordinary intelligence, encouraging honesty and using deceptive practices does not make a confession involuntary.
Moreover, the state appellate court ignored the many signs that Dassey was trying to please the interrogators and avoid conflict and a clear-cut pattern of fact-feeding linked to promises that, together, resulted in a situation where Das-sey’s will clearly was overborne. That pattern was as follows: the investigators emphasized, ad nauseum, that in order to be “okay” to “get things over with” to be “set free” Dassey had to be “honest.” Yet throughout the interrogation it became clear that “honesty” meant those things that the investigators wanted Dassey to say. Whenever Dassey reported a fact that did not fit with the investigators’ theory, he was chastised and told that he would not be “okay” unless he told the truth. And this pattern continued until Dassey finally voiced what the investigators wanted him to say, seemingly by guessing, or the investigators fed him the information they wanted. Once he spoke “correctly,” the investigators anchored the story by telling Dassey, “now we believe you” to signal to him that this was the version that would allow him to be “okay,” or “set him free.” By doing this— by linking promises to the words that the investigators wanted to hear, or allowing Dassey to avoid confrontation by telling the investigators what they wanted to hear—the confession became a story crafted by the investigators instead of by Dassey. And, as we will see, it was a confession that therefore cannot not be viewed as voluntary.
In this case the analysis of 2254(d)(1) and 2254(d)(2) overlap. The state court unreasonably applied the rule requiring it to consider the totality of the circumstances to the facts of the case, and those *951were the very same facts that the state court determined unreasonably.
B. Voluntariness in confessions.
1. The constitutional requirement of voluntariness.
False confessions are anathema to the judicial process. They are not beneficial to the prosecutor whose goal is to find, punish, and incapacitate the actual criminal, they are not beneficial to grieving relatives and friends who want to bring justice to the perpetrator of a crime, and, of course, they are of no benefit to a wrongfully accused defendant. For these reasons it is obvious why coercive tactics that lead to a false confession would be an affront to our judicial system. But the use of involuntary confessions violates the Constitution even when they are confessions of truth (where, in fact, it is possible to know such a thing). “The aim of the requirement of due process is not to exclude presumptively false evidence, but to prevent fundamental unfairness in the use of evidence, whether true or false.” Colorado v. Connelly, 479 U.S. 157, 167, 107 S.Ct. 515, 93 L.Ed.2d 473 (1986) (citing Lisenba v. California, 314 U.S. 219, 236, 62 S.Ct. 280, 86 L.Ed. 166 (1941)). The Supreme Court has long held that “certain interrogation techniques, either in isolation or as applied to the unique characteristics of a particular’ suspect, are so offensive to a civilized system of justice that they must be condemned under the Due Process Clause of the Fourteenth Amendment.” Miller v. Fenton, 474 U.S. 104, 109, 106 S.Ct. 445, 88 L.Ed.2d 405 (1985) (citing Brown v. Mississippi, 297 U.S. 278, 56 S.Ct. 461, 80 L.Ed. 682 (1936)). Coerced confessions also violate the Fifth Amendment’s right against self-incrimination. Withrow v. Williams, 507 U.S. 680, 688, 113 S.Ct. 1745, 123 L.Ed.2d 407 (1993). As the Supreme Court noted, “[A] criminal law system which comes to depend on the confession will, in the long run, be less reliable and more subject to abuses than a system relying on independent investigation.” Berghuis v. Thompkins, 560 U.S. 370, 403-04, 130 S.Ct. 2250, 176 L.Ed.2d 1098 (2010) (internal citations omitted).
“[T]he ultimate issue of ‘voluntariness’ is a legal question requiring independent federal determination.” Arizona v. Fulminante, 499 U.S. 279, 287, 111 S.Ct. 1246, 113 L.Ed.2d 302 (1991); Miller v. Fenton, 474 U.S. at 110, 106 S.Ct. 445. And under the AEDPA, this court must ask whether the Wisconsin appellate court’s decision concluding that Dassey’s confession was not involuntary “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States,” (28 U.S.C. § 2254(d)(1); Bobby v. Dixon, 565 U.S. 23, 27, 132 S.Ct. 26, 181 L.Ed.2d 328, (2011)), or whether it was based on an unreasonable determination of the facts in light of the evidence presented in the state court proceeding. 28 U.S.C. § 2254(d)(2).
2. The risks of coercion on voluntariness.
Historically, courts have looked at traditional modes of coercion in evaluating whether the defendant voluntarily confessed—that is, whether the suspect was tortured, beaten, or deprived of sleep, food or water. The Supreme Court and the community of experts on confessions have long recognized, however, that psychological coercion can be as powerful a tool as physical coercion. Fulminante, 499 U.S. at 287, 111 S.Ct. 1246.
The primary cause of police-induced false confessions is the use of psychologically coercive police interrogation methods. These include methods that were once identified with the old “third degree,” such as deprivation (of food, sleep, water, or access to bathroom fa*952cilities, for example), incommunicado interrogation, and extreme induced exhaustion and fatigue. Since the 1940s, however, these techniques have become rare in domestic police interrogations. Instead, when today’s police interrogators employ psychologically coercive techniques, they usually consist of implicit or explicit promises of leniency and implicit or explicit threats of harsher treatment in combination with other interrogation techniques such as accusation, repetition, attacks on denials, and false evidence ploys.
Jon B. Gould & Richard A. Leo, One Hundred Years Later: Wrongful Convictions After A Century of Research, 100 J. Crim. L. & Criminology 825, 846 (2010).
In closing arguments at trial, the state argued that “people who are innocent don’t confess.” R. 19-28 at 144. We know, however, that innocent people do in fact confess and do so with shocking regularity. The National Registry of Exonerations has collected data on 1,994 exonerations in the United States since 1989 (as of February 26, 2017), and that data includes 227 cases of innocent people who falsely confessed.5 This research indicates that false confessions (defined as cases in which indisputably innocent individuals confessed to crimes they did not commit) occur in anywhere from 15-24% of wrongful convictions cases. Samuel Gross & Michael Shaffer, Exoneration in the United States, 1989-2012: Report by the National Registry of Exonerations, 60.6
3. The heightened risks of coercion for youth and the intellectually disabled.
Nowhere is the risk of involuntary and false confessions higher than with youth and the mentally or intellectually disabled. It is for this reason that the Supreme Court has cautioned courts to exercise “special caution” when assessing the voluntariness of juvenile confessions. J.D.B., 564 U.S. at 269, 131 S.Ct. 2394; In re Gault, 387 U.S. at 45, 87 S.Ct. 1428; Gallegos, 370 U.S. at 53-54, 82 S.Ct. 1209, (1962); Haley, 332 U.S. at 599-601, 68 S.Ct. 302.
Indeed, the pressure of custodial,interrogation is so immense that it “can induce a frighteningly high percentage of people to confess to crimes they never committed.” That risk is all the more troubling—and recent studies suggest, all the more acute—when the subject of custodial interrogation is a juvenile.
J.D.B. 564 U.S. at 269, 131 S.Ct. 2394 (internal citations omitted). In one of the seminal juvenile coerced-confession cases, the Court noted that interrogators must treat minors more carefully when questioning them as “[t]hat which would leave a man cold and unimpressed can overawe and overwhelm a lad in his early teens.” Haley, 332 U.S. at 599, 68 S.Ct. 302.
As the amicus curiae and related articles demonstrate, data supports the Supreme Court’s admonition for special care. A survey of false confession cases from 1989-2012 found that 42% of exonerated defendants who were younger than 18 at the *953time of the crime confessed, as did 75% of exonerees who were mentally ill or mentally retarded, compared to 8% of adults with no known mental disabilities. Samuel Gross & Michael Shaffer, Exoneration in the United States, 1989-2012: Report by the National Registry of Exonerations, 58.7 Overall, one sixth of the exonerees were juveniles, mentally disabled, or both, but they accounted for 59% of false confessions. Id. In another study of those exonerated by DNA, juveniles accounted for one third of all false confessions. Brandon L. Garrett, The Substance of False Confessions, 62 Stan. L. Rev. 1051, 1094 (2010). Indeed, age and intellectual disability are the two most commonly cited characteristics of suspects who confess falsely. Samuel R. Gross, Kristen Jacoby, Daniel J. Matheson, and Nicholas Montgomery, Ex-onerations in the United States 1989 through 2003, 95 J. Crim. L. & Criminology 523, 545 (2005).8 Dassey suffered under the weight of both youth and intellectual deficit and thus the state court was required, by a long history of Supreme Court precedent, to .assess the voluntariness of his confession with great care, yet the state appellate court did not do so. Although it mentioned Dassey’s age and low IQ it never made any assessment about how the interrogation techniques could have affected a person with these characteristics.
4. The totality of the circumstances requirement for assessing voluntariness.
There is no magic formula or even an enumerated list for assessing the voluntariness of a confession. Such an assessment depends, instead, upon the totality of the circumstances. Withrow, 507 U.S. at 693, 113 S.Ct. 1745; Schneckloth v. Bus-tamonte, 412 U.S. 218, 226, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973). An incriminating statement is voluntary “if, in the totality of circumstances, it is the product of a rational intellect and free will and not the result of physical abuse, psychological intimidation, or deceptive interrogation tactics that have overcome the defendant’s free will.” Carrion v. Butler, 835 F.3d 764, 775 (7th Cir. 2016). Police conduct may be unduly coercive because of the inherent nature of the conduct itself or because “in the particular circumstances of the case, the confession is unlikely to have been the product of a free and rational will.” Miller v. Fenton, 474 U.S. at 110, 106 S.Ct. 445. “The admissibility of a confession turns as much on whether the techniques for extracting the statements, as applied to this suspect, are compatible with a system that presumes innocence and assures that a conviction will not be secured by inquisitorial means as on whether the defendant’s will was in fact overborne.” Id. at 116, 106 S.Ct. 445 (emphasis in original). In short, a court must look at the interplay between the characteristics of the defendant and the nature of the interrogation. A simple recitation of each, as the state appellate court did here, is not sufficient.
Factors that courts consider as part of the totality of the circumstances include the length of the interrogation, its location, its continuity, the defendant’s maturity, education, physical condition, mental health, and whether the police advised the defendant of his right to remain silent and have counsel present. Withrow, 507 U.S. at 693-94, 113 S.Ct. 1745. In juveniles, as we have noted, the evaluation of the totality of the circumstances “includes evaluation of the juvenile’s age, experience, education, background, and intelligence, *954and into whether he has the capacity to understand the warnings given him, the nature of his Fifth Amendment rights, and the consequences of waiving those rights.” Fare, 442 U.S. at 725, 99 S.Ct. 2560; see also Murdock, 846 F.3d at 209; Hardaway, 302 F.3d at 762.
The state appellate court did not give Dassey’s confession the consideration required when evaluating the voluntariness of a confession of an intellectually disabled juvenile.
5. Cases as guideposts for a voluntariness assessment.
By surveying the Supreme Court cases on the voluntariness of juvenile confessions one can see how much the unique characteristics of both the defendant and the interrogation play into the assessment of voluntariness. For this reason, other cases can only act as broad guideposts. “Determination of whether a statement is involuntary requires more than a mere color-matching of cases. It requires careful evaluation of all the circumstances of the interrogation.” Mincey v. Arizona, 437 U.S. 385, 401, 98 S.Ct. 2408, 57 L.Ed.2d 290 (1978) (internal citations omitted).
For example, in Haley, the Supreme Court held that the methods used in obtaining the confession of a fifteen-year-old boy could not be squared with the due process commanded by the Fourteenth Amendment. Haley, 332 U.S. at 599, 68 S.Ct. 302. Haley was arrested at midnight and interrogated for five straight hours by six officers in relays, after which time he confessed without being told his rights. Id. He was then informed of his rights and signed a written confession. Only after another three days of isolation did the police allow him access to his parents or a lawyer. Id. That confession, the court found, could not be deemed voluntarily made.
Likewise for fourteen-year-old Robert Gallegos, who was picked up by the police for assault and robbery and immediately admitted to a crime. Gallegos, 370 U.S. at 50, 82 S.Ct. 1209. He was locked in juvenile hall for five days without access to a lawyer or his parents, despite his mother’s attempts to see him, after which time he signed a confession. Id. The court concluded that a fourteen year old in those circumstances would have had no way to know what the consequences of his confession were without advice as to his rights. Id. at 54, 82 S.Ct. 1209.
In contrast, in Fare, a sixteen-year-old with rather extensive prior experience in the criminal system confessed to murder after being informed of his Miranda rights. Fare, 442 U.S. at 709-11, 99 S.Ct. 2560. The Supreme Court found that “there is no indication that he was of insufficient intelligence to understand the rights he was waiving, or what the consequences of that waiver would be. He was not worn down by improper interrogation tactics or lengthy questioning or by trickery or deceit.” Id. at 726-27, 99 S.Ct. 2560. And therefore, based on the totality of the circumstances, the confession was not coerced and thus admissible. Id. at 727, 99 S.Ct. 2560.
The cases from this circuit also demonstrate how we have applied Supreme Court precedent to determine the reasonableness of a state court’s determination of voluntariness. Derrick Hardaway was only fourteen years old when the police roused him from his sleep at 8:00 a.m., and took him to the police station without his parents. Hardaway, 302 F.3d at 760. He was not handcuffed and remained in an unlocked interrogation room until he was interviewed at 10:30 am and then interrogated for six hours, given a break for a few hours, and then interrogated again for another four hours. A youth advocate joined the interrogation but never once spoke up to aid Hardaway. A clearly torn panel of *955this court could not find that the state appellate court erred when it held that the confession was voluntary, even if we might have come to a different conclusion had we been deciding the matter ourselves in the first instance.
There is no doubt that Hardaway’s youth, the lack of a friendly adult, and the duration of his interrogation are strong factors militating against the vol-untariness of his confession; indeed, it seems to us that on balance the confession of a 14-year-old obtained in those circumstances may be inherently involuntary.
Id. at 767. Nevertheless, we concluded, the state court had considered the relevant factors and because “the weighing of factors under the totality of circumstances test is a subject on which reasonable minds could differ,” we could not hold that the state court had been unreasonable. Id. The state court, we explained, noted that the officers did not psychologically trick the defendant or misrepresent evidence, but rather Hardaway confessed after being confronted with truthful contradictory evidence. The state court carefully considered Hardaway’s nineteen previous encounters with law enforcement, the fact that the police not only read Hardaway his rights but that Hardaway was able to articulate them back in his own words, and that Hardaway did not have any mental incapacity or other mental infirmities. Id. at 767-78. Thus the state court seemed to have considered sufficiently the interaction between Hardaway’s limitations and the interrogation.
Similarly, in Carter v. Thompson, 690 F.3d 837, 844 (7th Cir. 2012), despite the fact that we were “unsettled” that a 16-year-old was in the police station for fifty-five hours without a blanket, pillow, change of clothes, or access to a shower, and without being told she could leave, we could not find that the state courts had been unreasonable in finding that her confession was voluntary. Id. The state court had considered all of these factors, along with the fact that the police read Carter her rights, her parents were with her for two of her three confessions, and her confession occurred impromptu, as she was on her way to the bathroom. Id.
Finally, in Etherly v. Davis, 619 F.3d 654, 662 (7th Cir. 2010), as amended on denial of reh’g and reh’g en banc (Oct. 15, 2010), we reversed a district court grant of a writ of habeas corpus, disagreeing with the lower court’s assessment that the Illinois appellate court had not properly addressed and considered all of the relevant factors in its analysis, noting that reasonable jurists could disagree about the weight to assign to each factor. Id. The Illinois appellate court, we concluded, evaluated and discussed the importance of the defendant’s age, whether a friendly adult was present, his intellectual disability, lack of criminal background, whether police engaged in physical or psychological coercion, and the defendant’s assertion that he understood his Miranda rights. Id. at 662. And despite agreeing that the state appellate court had been unreasonable in concluding that a fifteen-year-old, with no pri- or criminal experience, should be expected to seek the advice of a youth officer, this court concluded that this “lone error is not of such magnitude as to result in an unreasonable application of Supreme Court precedent under AEDPA.” Id. at 662-63.
In general, our eases demonstrate that we show great deference to state court adjudications where it is clear that the state court considered the totality of the circumstances cumulatively, in light of the defendant’s age and intellect, and without omitting or overlooking relevant factors bearing on the voluntariness of a juvenile confession. Murdock, 846 F.3d at 210-11; Gilbert v. Merch., 488 F.3d 780, 794 (7th *956Cir. 2007); Ruvalcaba v. Chandler, 416 F.3d 555, 561-62 (7th Cir. 2005).
Unlike in the cases above, where the state court sufficiently considered a totality of the circumstances, as cases like Fare and Carter require (Fare, 442 U.S. at 725, 99 S.Ct. 2560; Carter, 690 F.3d at 843), we see no similar evidence that the state court did so in Dassey’s case. For example, despite the Supreme Court’s emphasis on the importance of access to an adult ally in Gallegos, the Wisconsin state court in this case never discussed the fact that Dassey was alone, other than to note that “Das-sey’s mother, Barbara Janda, agreed to the second interview but declined the offer to accompany Dassey.” State v. Dassey, 2013 WL 335923 at *1.9
Moreover, in this case, in comparison to Fare and Hardaway (Fare, 442 U.S. at 725, 99 S.Ct. 2560; Hardaway, 302 F.3d at 767), the state appellate court did not view interrogation techniques as a totality factor overlaid with Dassey’s age and intellect. It merely looked at the investigators’ comments in isolation and opined, as it would with an adult of ordinary intelligence, that “[a]s long as investigators’ statements merely encourage honesty and do not promise leniency, telling a defendant that cooperating would be to his or her benefit is not coercive conduct.” State v. Dassey, 2013 WL 335923 at *2.
And unlike in Etherly where the state court made a single error—unreasonably concluding that the absence of a youth officer was inconsequential (Etherly, 619 F.3d at 662-63)—the state court’s error here was not a solitary one, but rather a failure of the very essence of Supreme Court precedént requiring a court to consider the totality of the circumstances and to consider juvenile confessions with special caution.
Where a determination of volun-tariness is so outside the realm of reasonableness, a federal court may grant the writ, as it did in AM. v. Butler, 360 F.3d 787, 801 (7th Cir. 2004). The court in A.M. recognized that “[e]ven in the context of federal habeas, deference does not imply abandonment or abdication of judicial review. Deference does not by definition preclude relief.” Id. (citing Miller-El v. Cockrell, 537 U.S. at 340, 123 S.Ct. 1029). And it concluded that the confession of an inexperienced 10-year-old who had no adult advocate was simply not reliable where the detective continually challenged the boy’s statements and accused him of lying—a legitimate interrogation technique in adults, but one likely to lead a young boy to confess to anything. AM., 360 F.3d at 800-01. And in fact, that is just what occurred in this case—detectives continually challenged Dassey’s statements and accused him of lying until, as we will describe, his confession became a litany of inconsistencies—shirts that changed color, fires that began and ended at different times, garbage bags that sat in burning fires without melting, trucks that were seen in garages and then not seen in garages, bloody crime scenes without a trace of blood remaining, metal handcuffs that left no marks on the bed posts, etc. But again we emphasize that because of the requirements of the totality of the circumstances, these cases provide only the broadest of guidelines on determining vol-untariness, see Mincey, 437 U.S. at 401, 98 S.Ct. 2408, and our full analysis of the voluntariness of the confession, toward the end of this opinion, will demonstrate why no reasonable court could have come to the conclusion that Dassey’s confession was voluntary. As will become clear through the entirety of this opinion, we can point to *957no solitary statement, factor, or interrogation question that rendered Dassey’s confession involuntary (although there were certainly some individual leading questions that came close), but rather it was death by a thousand cuts. Because of the cumulative effect of these coercive techniques— the leading, the fact-feeding, the false promises, the manipulation of Dassey’s desire to please, the physical, fatherly assurances as Wiegert touched Dassey’s knee etc.—no reasonable court could have any confidence that this was a voluntary confession.
6. No single factor is determinative.
a. Courts must pay dose attention to voluntariness when the defendant has no adult ally present.
As we have now concluded, the totality test prohibits any one factor from being determinative of voluntariness. Mur-dock, 846 F.3d at 209. Some courts, including this one, nevertheless have found particularly distressing the idea of minors waiving rights and confessing without an adult ally present. Those courts therefore have toyed with the idea of a per se rule that children under a certain age cannot waive rights or make a voluntary confession without a parent, guardian, or legal representative present. See e.g., Hardaway, 302 F.3d at 764. Our conclusion in Hardaway, however, was that there is no support in clearly established federal law for such a per se rule where Supreme Court precedent has been clear that courts instead must base their assessment on the “totality of the circumstances.” Id. (citing Fare, 442 U.S. at 726, 99 S.Ct. 2560). ‘Youth,” we concluded, “remains a critical factor for our consideration, and the younger the child the more carefully we will scrutinize police questioning tactics to determine if excessive coercion or intimidation or simple immaturity that would not affect an adult has tainted the juvenile’s confession.” Hardaway, 302 F.3d at 765. See also, J.D.B. v. North Carolina, 564 U.S. 261, 269, 280, 131 S.Ct. 2394, 180 L.Ed.2d 310 (2011); In re Gault, 387 U.S. at 45, 87 S.Ct. 1428.
The state appellate court applied no extra care to Dassey’s confession based on his lack of an adult advocate. Youth was not a “critical factor” in its analysis; indeed it was not a factor at all. It did not consider the interrogation techniques in light of Dassey’s lack of an adult advocate nor acknowledge how Dassey’s clear confusion during parts of the interview could have been aided by an adult ally who might have noticed Dassey’s confusion and the manipulation. It did not mention how, immediately after Dassey’s mother came to his side, he suddenly realized that the investigators “got to my head,” and he worried that he would be caught in a lie— having confessed to a crime he did not commit. He asks his mother, “What’d happen if he says something his story’s different. Wh-he says he, he admits to doing it? ... Like if his story’s different, like I never did nothin’ or somethin.’ ” R. 19-25 at 148.
b. Courts must pay close attention to voluntariness when manipulative interrogation techniques are used, particularly on the young and intellectually challenged.
Psychologically manipulative interrogation techniques, likewise, are not per se coercive, but among the circumstances that a court must evaluate in total to determine whether a particular defendant’s free will has been overcome. To be clear, many manipulative interrogation techniques, in and of themselves, are not unconstitutional. “Trickery, deceit, even impersonation do not render a confession inadmissible.” United States v. Villalpando, 588 F.3d 1124, 1128 (7th Cir. 2009) *958(citing U.S. v. Kontny, 238 F.3d 815, 817 (7th Cir. 2001)). The law permits the police to “pressure and cajole, conceal material facts, and actively mislead—all up to limits.” United States v. Rutledge, 900 F.2d 1127, 1131 (7th Cir. 1990). That limit is exceeded, however, when the government gives the suspect information that destroys his ability to make a rational choice “for example by promising him that if he confesses he will be set free.” Aleman v. Vill. of Hanover Park, 662 F.3d 897, 906 (7th Cir. 2011). And, as we describe further below, those limits depend on the characteristics of the defendant. False promises that a suspect will be treated leniently by the courts, we have noted, have “the unique potential to make a decision to speak irrational and the resulting confession unreliable ... because of the way it realigns a suspect’s incentives during interrogation.” Villalpando, 588 F.3d at 1128; United States v. Montgomery, 555 F.3d 623, 629 (7th Cir. 2009) (“a false promise of leniency may be sufficient to overcome a person’s ability to make a rational decision about the courses open to him.”). See also United States v. Nichols, 847 F.3d 851, 857 (7th Cir. 2017) (“a government agent’s false promise of leniency may render a statement involuntary.”); Montgomery, 555 F.3d at 629 (“[gjiven the right circumstances, a false promise of leniency may be sufficient to overcome a person’s ability to make a rational decision about the courses open to him.”); Hadley v. Williams, 368 F.3d 747, 749 (7th Cir. 2004) (police may not extract a confession in exchange for a false promise to set the defendant free).
We attach no nefarious purposes to the investigators who were using established interrogation techniques.10 And, in any event, the investigator’s purpose or subjective view of the coercive nature of the interrogation is not relevant. It is how those interrogation techniques interact with the defendant’s characteristics that determines the voluntariness of a confession. A seasoned criminal who has volleyed with interrogators many times before may not be swayed at all by an explicit but false claim of leniency, but a young, unsophisticated juvenile might believe, with just the slightest hint of an offer of leniency, that if he confesses to murder “God and the police would forgive him and he could go home in time for his brother’s birthday party.” A.M., 360 F.3d at 794.
The Constitution requires that a confession be voluntarily given. The dissent criticizes the panel opinion for relying on the subjective perception of a defendant in determining the voluntariness of his confession, but this is, in fact, what the totality of the circumstances test requires. A thirty-year-old with a law degree would not believe a police officer’s assurance that if he confesses to murder he will go punishment free, but yet the ten-year-old, A.M. did just that. Id. A consideration of the totality of the circumstances requires the court to consider “whether the techniques for extracting the statements, as applied to this suspect, are compatible with a system that presumes innocence and assures that a conviction will not be secured by inquisitorial means as on whether the defendant’s will was in fact *959overborne.” Miller v. Fenton, 474 U.S. at 116, 106 S.Ct. 445 (emphasis added). We need not accept a defendant’s after-the-fact proclamation of a lack of voluntariness, but the totality of the circumstances framework allows a court to consider the evidence about the defendant’s ability to comprehend and contemporaneous evidence of what he actually did or did not understand. If the Constitution requires that a confession be voluntary, then it can only be so if the particular defendant sitting in the interrogation was not, in fact, coerced.
In other words, the totality of the circumstances test dictates that coercive interrogation on the one hand, and suspect suggestibility, on the other, are on inverse sliding scales—the more vulnerable or suggestible a suspect, the less coercion it will take to overcome her free will. This is not a statement of a new test, but rather the logical conclusion of the totality of the circumstances review itself. Therefore, to determine whether a promise is coercive as a legal matter, a court cannot consider the promise alone, but rather the promise in conjunction with the characteristics of the suspect. Again, the Supreme Court’s seminal case advises, “[t]hat which would leave a man cold and unimpressed can overawe and overwhelm a lad in his early teens.” Haley, 332 U.S. at 599, 68 S.Ct. 302. And the Supreme Court precedent requires lower courts to consider interrogation techniques as applied to the particular defendant at hand. Miller v. Fenton, 474 U.S. at 116, 106 S.Ct. 445.
The dissent accuses us of redefining what counts as a false promise of leniency, noting statements by the police that passed muster with courts in other cases. The point of the totality test, however, is not to evaluate any promise of leniency in isolation, but rather in light of the specific characteristics of the defendant, that is, “as applied to this suspect.” Id. (emphasis added). The career criminal will not interpret a promise in the same manner as an inexperienced and intellectually disabled teen. The state court, however, did not view the coerciveness of the interrogation techniques in light of Dassey’s personal characteristics as the totality test requires.
The dissent states that the majority decision will make police investigations “considerably more difficult,” and asks “what should police do the next time an investigation leads to a teenager with some intellectual challenge?” (post at 984). To the extent that the result makes police investigations more difficult, it is not because of any change we have made to the law, but rather because the Supreme Court requires a totality of the circumstances framework that gives special caution to confessions of juveniles, the intellectually disabled and other defendants with vulnerable characteristics.
The benefits of the Supreme Court’s requirements expand beyond protecting the constitutional rights of defendants. It is of no help to the advancement of justice and to removing dangerous killers from the streets, if police coerce confessions from innocent suspects. Teresa Halbach and her family are not served if the wrong defendant spends his life in prison. Teresa’s family deserves to know that the police have found and incapacitated the right perpetrator—that no other family will be forced to grieve as they have because a brutal killer remains at large. The answer to the dissent’s inquiry about what police officers are to do in such a situation as Dassey’s, therefore, comes from a long line of requirements that courts have established for protecting the rights of defendants during police interrogations. Specifically, in such a case, the police should, as the Supreme Court requires, ensure that such a suspect “has the capacity to understand the warnings given him, the nature *960of his Fifth Amendment rights, and the consequences of waiving those rights.” Fare, 442 U.S. at 725, 99 S.Ct. 2560; see also Murdock, 846 F.3d at 209; Hardaway, 302 F.3d at 762. And a court reviewing a challenge to a confession must assess the totality of the circumstances to assure itself that the defendant voluntarily confessed. This the appellate court did not do.
7. The state court in this case did not apply a totality of the circumstances test.
The state court of appeals in this case affirmed the trial court’s determination that Dassey’s confession was not involuntary. State v. Dassey , 2013 WL 335923 at *2. As the last state court to speak to the issue, it is that court’s decision that we review. Makiel, 782 F.3d at 896. As set forth in the fact section above, after noting the requirement to consider the voluntariness of the confession using the totality of the circumstances test, the state appellate court addressed the voluntariness of the confession in two short paragraphs. The first paragraph (¶ 6) consisted of a list of Dassey’s characteristics and some general characteristics of the interrogation including: Dassey’s limited intelligence, the comfortable interrogation room, the Miranda warnings, his affect during the interview, the investigators’ normal speaking tones, the lack of “hectoring, threats or promises of leniency,” the pleas for honesty, and the investigators’ attempts to build rapport. State v. Dassey, 2013 WL 335923 at *2. In the second paragraph (¶ 7), the court of appeals concluded that the trial court’s finding of no coercion was not clearly erroneous. “As long as investigators statements merely encourage honesty and do not promise leniency,” the court reasoned, “telling a defendant that cooperating would be to his or her benefit is not coercive conduct. Nor is professing to know facts they actually did not have.” Id.
Although the statements in this second paragraph are accurate as applied to an adult of ordinary intelligence, they do not acknowledge the court’s obligation to consider juvenile confessions with caution and they do nothing to evaluate the totality of the circumstances. An evaluation requires that the court view the interrogation tactics in light of the defendant’s situation and characteristics. A court has not applied the totality of the circumstances test simply by stating its name and by noting that, in the ordinary course of dealings, a police officer may use deceptive techniques. Applying a rule of law does not require much, but it requires more than just parroting the words of the rule.
In addition to failing to consider the factors in light of the totality of the circumstances, the state appellate court failed to consider some key factors at all, even individually. The dissent correctly notes that a state court need not give all of its reasoning for its outcome. And the totality of the circumstances does indeed give state courts a somewhat wide berth for their considerations. It is true that “[t]he more general the rule, the more leeway courts have in reaching outcomes in case-by-case determinations.” Yarborough v. Alvarado, 541 U.S. 652, 664, 124 S.Ct. 2140, 158 L.Ed.2d 938 (2004). But the generality of the rule does not mean that a state court may forsake it completely, and it does not eradicate the general notion that “The standard [for habeas corpus relief] is demanding but not insatiable ... deference does not by definition preclude relief.” Miller-El v. Dretke, 545 U.S. at 240, 125 S.Ct. 2317.
If the totality of the circumstances standard means anything, it means that a state court must, at a bare minimum, do what the rule requires and consider the totality of the circumstances. A state court need not say much, but the less it says, the less *961a federal court can ascertain that the state actually applied a totality of the circumstances evaluation.
And at the very least a court assessing the voluntariness of a juvenile’s confession must evaluate whether deceptive interrogation techniques overcame the free will of this particular defendant. Missing entirely from the state appellate court’s analysis is any recognition that deception that is permissible when interrogating the average adult person of ordinary intelligence, might not be permissible with someone of Dassey’s age and intellect. For example, the state appellate court never considered whether the statement “the truth will set you free” would be considered idiomatically or literally by someone of Dassey’s age and limitations. Indeed if taken literally, that statement is the exact kind of promise of leniency that courts generally find coercive. Hadley, 368 F.3d at 749 (police cannot extract a confession in exchange for a false promise to set the defendant free); Rutledge, 900 F.2d at 1129 (same).
Nor was there any analysis of the key fact that Dassey had no adult ally with him during the interrogation. Although not dispositive, it is one of the most critical factors in evaluating voluntariness of juvenile confessions. Gallegos, 370 U.S. at 55, 82 S.Ct. 1209; Hardaway, 302 F.3d at 765 (noting that absence of a friendly adult is not dispositive of involuntariness, but a key factor that can tip the balance against admission). A friendly adult can ensure that a minor defendant can make critical decisions, for example, like the decision to waive Miranda rights. See Hardaway, 302 F.3d at 764. She could ensure that police do not take advantage of a minor’s youth or mental shortcomings. U.S. v. Bruce, 550 F.3d 668, 673 (2008). A friendly adult can level the playing field, help the child understand what the consequences of his confession might be, and help him understand his constitutional rights. Gilbert, 488 F.3d at 791-92.
Had Dassey’s mother been present in the room with him, she might have noticed if Dassey were guessing as to answers, alerted him to the consequences of incriminating himself, reminded her son that the investigators were not acting as his friends or advocates, and helped him distinguish between the actual truth and the information that the investigators were feeding him.
Obviously, we cannot know if she would have done any of these things, but we have one hint that she might have: At the end of the confession, after she was allowed to see Dassey and after he said “they got to my head,” she immediately asked the investigators, “Were you pressuring him?” R. 19-25 at 148. As we described above, Dassey became anchored and immediately realized, “They got to my head,” as soon as his mother entered the room. R. 19-25 at 148. But whether she would have helped Dassey or not, it confirms that Dassey had no protection against manipulation by the officers. The absence of Dassey’s mother or another friendly adult should have been a critical piece of the totality consideration by the state court and it was not even mentioned in the state court’s analysis of the voluntariness of Dassey’s confession.
Finally, the state appellate court did not consider Dassey’s suggestibility while assessing the coercive nature of the claim, despite the fact that one entire day of trial testimony consisted of experts assessing Dassey’s mental capacity and, in particular, his suggestibility. Given the instances we discuss below of investigators steering him to particular answers, this was a critical oversight.
The directive from the Supreme Court to consider the totality of the circumstances ensures that this particular defendant voluntarily confessed. It is no use to *962note that telling a defendant that cooperating would be to his benefit is not per se coercive, if the words used to convey that notion sound like a promise of leniency to this particular defendant. Likewise, falsely claiming to have knowledge is not per se coercive, unless it is used in a manner that overcomes the free will of this particular defendant. The state court did not, in any respect or manner, consider the interaction of the interrogation techniques with Dassey’s youth, intellectual limitations, suggestibility, lack of experience with the police, lack of a friendly adult, and naivete.
In sum, there was no “totality” in this “totality of the circumstances” test at all. There was no assessment of the cumulative nature of the interrogators’ promises, no assessment of the fact-feeding in light of Dassey’s limited intellectual abilities, no assessment of the absence of a friendly adult who could protect Dassey and advocate for his interests, no assessment of Dassey’s confusion in response to many questions, or his apparent desire to please the interrogators with his answers, no assessment of how his answers changed and why, and no assessment of his repeated statements that he expected that, in return for his statements, he would be “set free” to return to school at the conclusion of the interrogation. It is not that the state court did not do enough; we can have no confidence that it considered the totality of the circumstances at all.
Although different courts and judges might disagree as to “how much weight to assign each factor on facts similar to those in [any Petitioner’s] case” (Etherly, 619 F.3d at 662), a reasonable jurist must, in fact, consider the relevant facts surrounding a confession, and consider their combined and cumulative effect. Id. A consideration of the totality of the circumstances requires the court to consider “whether the techniques for extracting the statements, as applied to this suspect, are compatible with a system that presumes innocence and assures that a conviction will not be secured by inquisitorial means as on whether the defendant’s will was in fact overborne.” Miller v. Fenton, 474 U.S. at 116, 106 S.Ct. 445 (emphasis added).
C. The voluntariness of Dassey’s confession analyzed in light of the totality of the circumstances.
In addition to failing to apply a totality of the circumstances analysis to the facts of this case, as required by the Supreme Court, the state court acted unreasonably when it determined that—given the totality of the circumstances—Dassey’s confession was voluntary. The state appellate court’s finding that there were no promises of leniency or other factors that overcame Dassey’s free will was against the clear weight of the evidence. 28 U.S.C. § 2254 (d)(2); Ward v. Sternes, 334 F.3d 696, 704 (7th Cir. 2003).
Thus § 2254 (d)(2) requires a federal court on habeas review to look at those facts to determine whether the state court proceedings “resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” Id. Moreover, “[w]here a state court’s decision is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Harrington v. Richter, 562 U.S. 86, 98, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011). And a federal court reviewing a habeas petition under § 2254(d), “must determine what arguments or theories supported or, as here, could have supported, the state court’s decision; and then it must ask whether it is possible fairminded jurists could disagree that those arguments or theories are inconsistent with thé holding in a prior decision of this Court.” Id. at *963102, 131 S.Ct. 770. Such a determination does not turn habeas review to de novo review, as the dissent suggests. It is, to the contrary, precisely what the Supreme Court requires. Id. Because the state appellate court’s opinion failed to give any explanation other than a listing of Dassey’s characteristics and the circumstances of the interrogation, in reviewing the reasonableness of the determination of the facts in light of the evidence presented, we look to see what theories could have supported the state court’s conclusion.
1. The message sent to Dassey: “The ‘truth’ is what we want you to say, and that is what will set you free.”
Dassey’s interview could be viewed in a psychology class as a perfect example of operant conditioning. As we will demonstrate through myriad examples below, the theme set forth for Dassey was twofold, that “honesty is the only thing that will set you free,” R. 19-25 at 17, and that honesty would appease the investigators, avoid conflict, and allow them to be Dassey’s “Mend,” to “go to bat for [him]” to “be in his corner.” Id. at 16, 25. In other words, the key to walking out a free person, avoiding the conflict that his socially avoi-dant personality feared, and getting back in time for school lunch was “honesty.” But Dassey quickly learned that “honesty” meant telling the investigators what it was that they wanted to hear. When they did not like his answer, they told him things like “Come on Brendan. Be honest. I told you that’s the only thing that’s gonna help ya here;” and “[w]e don’t get honesty here, I’m your friend right now, but I gotta believe in you and if I don’t believe in you, I can’t go to bat for you.” Id. at 23. Every time the investigators said “tell us the truth” or “we know what the truth is,” Dassey altered his story just a bit. As Dassey got closer and closer to the answers the investigators were looking for, his statements were rewarded with affirmations like “that makes sense. Now we believe you,” and in doing so, they cemented that version of the facts. See, e.g., Id. at 73. But when Dassey deviated from the expected narrative, the investigators either offered no reward, ignored the comments, steered him away, or let him know that they thought he was not telling the truth. In short, as the examples clearly demonstrate, “be honest,” “tell the truth,” and similar pleas became code for “guess again, that is not what we wanted you to tell us.” And “now we believe you” and “that makes sense” became code for “that’s what we want to hear. Stop right there.” Dassey’s reaction to these cues is not unique. Experts on confessions have noted that “though courts are reluctant to find that police officers have overwhelmed a child’s will by repeatedly admonishing the child to ‘tell the truth,’ many children will eventually hear ‘tell the truth’ as, ‘tell me what I want to hear.’” Kenneth J. King, Waiving Childhood Goodbye: How Juvenile Courts Fail to Protect Children from Unknowing, Unintelligent, and Involuntary Waivers of Miranda Rights, 2006 Wis. L. Rev. 431, 472 (2006). Scholarly research such as this helps inform our understanding that the totality of the circumstances analysis means something different when applied to juveniles. It supports the reasoning behind the Supreme Court’s admonition to view juvenile confessions with special caution. See J.D.B., 564 U.S. at 269, 131 S.Ct. 2394.
The investigators’ “honesty is the only thing that will set you free” theme established a pattern whereby Dassey, seeking the promised result—freedom, or avoidance of conflict—searched for the narrative that the investigators would accept as “the truth.” Dassey found “the truth” either by stumbling upon it or by using the information the investigators had fed him. The promise of freedom became linked to the idea of truth which became defined as that which the investigators wanted to *964hear. Once this prompt-and-response pattern is noticed, it is impossible to read or view Dassey’s interrogation and have any confidence that Dassey’s confession was the product of his own free will rather than his will being overborne. Any reader who doubts that this pattern casts insurmountable doubt on the voluntariness of Dassey’s confession need only watch or read the interrogation with this “key” in hand.
The following exchange is a prime example of the investigators telling Dassey that he needs to change his story and how he should do it, followed by that exact change. Prior to the interaction below, Dassey confirmed approximately eight times, often insistently, that when he got home from school on October 31, he saw Halbach and Avery talking on Avery’s porch. R. 19-25 at 19-20, 27-28, 90. In fact, the officers grilled him asking “And you’re sure you saw that?” Id. at 20; “did you really see those two, talking on the porch” Id. at 27; “You’re 100% on that?” Id. at 28. And each time he answered affirmatively. Yet, once they repeatedly cued him that they did not like his answer and that he must “tell the truth”—in other words, tell them what they wanted to hear—he altered his message exactly as he was instructed:
Fassbender: OK, and you said you walked down th [sic] the road to your house, (Brendan nods “yes”) and you said that you saw Steven on the porch. Brendan: (nods “yes”) uh huh
Fassbender: Mark and I are havin’ a problem with that. Now if, I’m not, I’m not sayin’ that I’m gonna put words in your mouth so we’re havin’ a problem with that.... the time periods aren’t adding up. They’re not equaling out. We know when Teresa got there. (Brendan nods “yes”) Um, and, and I know I guarantee ya Teresa’s not standing on that porch when you come home from school. I ju [sic] I don’t see that.... Somethin’ is not adding up here and you need to tell us the truth. Did this all start right when you came home from school? You need to tell me, you need to be honest with me. I can’t tell ya, I I can’t tell ya these things. I can tell ya we don’t believe you because there’s some things that are wrong but you’ve gotta tell me the truth. This is you know getting’ serious here now, OK? (Brendan nods “yes”) Tell me what happened when you got home.
Brendan: I got off the bus. I walked down the road and when I got to that thing, ah, the other house I just sittin’ there for nothin’ [sic]. I could see her jeep in the garage just sittin’ there and I didn’t see Steven and her on the the porch.
Wiegert: You, you did or you didn’t?
Brendan: I didn’t.
Fassbender: Did not, OK.
R. 19-25 at 90-91 (emphasis added).11
The state presents these changes as a normal part of a confession. That is, that a *965defendant tells one version of events, backtracks as he is presented with inconsistencies and errors in his story, and reveals more and more as the interrogators coax the truth out of him. See, e.g., Reply Brief of Respondent-Appellant at 1 (“As with many difficult admissions, the truth did not come out,all at once, but little-by-little in fits of honesty.”) But again, a careful review of the confession does not reveal this to be a story gaining clarity over time. Unlike the ordinary course of a confession in which the narrative increases in clarity as the suspect reveals more information, this interrogation was just the opposite. Every time the interrogators protested the veracity of Dassey’s account or fed Dassey information, his story changed. If one sits in front of the taped confession with a legal pad and tries to sketch out the details and timeline of the crime, the resulting map is a jumble of scratch outs and arrows that grows more convoluted the more Dassey speaks. In fact, despite what the State describes as a detailed confession, it has never been able to map out a coherent timeline of the crime, or to figure out in what order or where many of the events occurred. See Brief of Respondent-Appellant at 9, n.3 (stating, in a footnote to the facts, “the narrative recounts details from Dassey’s confession in the most likely timeline, consistent with other evidence at trial. It is possible that some parts of the story are out of order,” and describing several items that are unclear).
Lest one think the details and timeline ever solidified, they did not. It only became more convoluted when Dassey appeared, without counsel, at the May 13 interrogation, after his lawyer’s own investigator, O’Kelly, had interrogated him. As we noted, that interrogation was not used at trial and the details are not discussed by the district court or by the parties. It was used, however, as part of the post-conviction hearing, and is part of the record. See R. 19-34. At the post-conviction hearing Wisconsin District Attorney Ken Kratz described that May 13 interrogation as a “fiasco” in which Dassey gave “inconsistent statements.” R. 19-26 at 97. Details both significant and insignificant changed, not only from the prior confession on March 1, but also within minutes of being disclosed at the May 13 interrogation. Das-sey changed details about things as benign as riding his bike to things as important as whether or not he cut Halbach’s throat. R. 19-34 at 7, 25. Dassey was inconsistent about how Halbach was restrained, about whether he saw Halbach’s vehicle, the order of events, facts about her body, where various events occurred, where the murder weapon came from, what it looked like, and what it was used for, where Halbach was stabbed, and, as we will see in a later example, whether he cut Halbach’s hair or not. Dassey is not merely a poor story teller who forgets details and orders, but rather the details and the order changes in ways that do not amount to confusion and error but rather a “fiasco” of a story— until, as we will see, the investigators steer him to the version of the story that fits their theory of the case.
For example, in the March 1 interrogation, on several occasions the investigators tried to pin down the constantly changing order of events. The events are gruesome, serious, and distinct, and the order is critical to how they were performed. For example, it is far different to choke a victim whose throat has been cut than to cut the throat of a victim who has been choked. Nevertheless, Dassey cannot keep these details straight. Initially Dassey said that Halbach was stabbed, tied up, and then choked R. 19-25 at 54-55. Moments later he stated that she was tied up, then stabbed, then choked Id. at 56, and a few transcript pages later he assures the investigators that he is “sure” that she was stabbed, choked, and then tied up Id. at 59; but a few pages after that he stated *966that she was stabbed, tied up and her throat was cut Id. at 64. Finally, he circles back to a re-telling in which he says that Halbach was tied up, stabbed and then cut Id. at 101. At one point the investigators are desperate to get the order right:
Fassbender: Brendan, we’re in the bedroom yet, OK? (Brendan nods “yes”) She’s handcuffed yet right? (Brendan nods “yes”) And you’re tellin’ me if, obviously correct me if I’m wrong, what we heard. (Brendan nods “yes”). While she’s handcuffed and alive, he stabs her. Brendan: (nods “yes”) mm huh. Fassbender: Chokes her? Right? (Brendan nods “yes”) Is that right?
Brendan: (nods “Yes”) mm huh. Fassbender: And then he has you cut her neck?
Brendan: Yeah.
Id. at 66. But just when the investigators thought that they had the order down, at the end of the interview they asked one more time to lock it in and the order falls apart again:
Wiegert: Well let’s, let’s just go back a little bit OK? Tell us what exactly happened to her, what order it happened in. You said there were basically three things prior to you guys shooting her. Explain those in, in the order that it happened.
Brendan: Starting with when we got in the room?
Fassbender: OK.
Wiegert: Yeah, what you guys did to her.
Brendan: We had sex with her
Wiegert: OK.
Brendan: Then he stabbed her.
Wiegert: Then who stabbed her?
Brendan: He did.
Wiegert: Who’s he?
Brendan: Steven.
Wiegert: OK, and then what?
Brendan: Then I cut her throat.
Wiegert: OK.
Brendan: And then he choked her and I cut off her hair.
Wiegert: OK. So he choked her after you cut her throat?
Brendan: (nods “yes”) mm huh.
Id. at 132-33. This is not a confession that becomes increasingly more coherent and clear over time, as the defendant reveals more and more of the truth. To the contrary, although Dassey’s culpability throughout these changes remains the same, the horrific story becomes less and less coherent until by the end Avery is choking a woman who has already had her throat cut. Yet through all of this tying, stabbing and throat cutting, Dassey insists he did not get any blood on himself:
Wiegert: You said that you had cut her throat. (Brendan nods “yes”) Here’s the thing Brendan, when you, cut somebody’s throat, they bleed a lot, (Brendan nods “yes”) OK? Am I right?
Brendan: (nods “yes”) Yeah.
Wiegert: She bleed a lot, (Brendan nods “yes”) so I know you had blood on ya, it’s pretty much impossible not to. Did you have blood on you?
Brendan: (shakes head “no”) No.
Wiegert: None at all?
Brendan: (shakes head “no”’) uh uh.
Wiegert: What about when you moved her?
Brendan: (shakes head “no”) No.
Id. at 116-17.
In short, a reasonable state court that had carefully reviewed the confession would have quickly determined that the interrogators pleas for honesty—irrespective of how they intended them—did not have the effect of eliciting honesty from Dassey, but rather had the effect of eliciting guesses from Dassey about what the investigators wanted to hear. In Dassey’s mind, the words “be honest” and the like *967came to mean “guess again until you say what we want to hear.” Consequently, the interrogation became not one of eliciting honesty through a voluntary confession, but one of leading Dassey into the story the interrogators wanted to hear. Nowhere is this more clear than in the following two examples below.
The first example comes from the key part of the interrogation. As we noted earlier, by the time of the March 1 interview, the investigators knew that Halbach had been shot in the head. They also knew that the battery had been removed from her Toyota RAY4. These two details had not yet been released publicly and thus Dassey’s knowledge of these details would be particularly inculpatory. It is a common investigative technique to hold back details of a crime from the media and public to test the validity of a confession. The following exchange demonstrates many of the totality factors and interrogation techniques we will describe below—Dassey’s naiveté, false information (“we already know”), minimizing Dassey’s role in the crime (“he made you do it”), and admonitions to “tell the truth.” But in particular it demonstrates how the interrogators’ admonitions to “tell the truth” cue Dassey to keep guessing, and most importantly, how the interrogators tainted the voluntariness of the interview by feeding Dassey the information that Halbach was shot in the head.
Wiegert: What else did he do to her? We know something else was done. Tell us, and what else did you do? Come on. (pause) Something with the head, (pause) Brendan?
Brendan: Huh?
Fassbender: ... can’t
Wiegert: What else did you guys do, come on.
Fassbender: What he made you do Brendan? We know he made you do somethin’ else
Wiegert: What was it? (pause) What was it?
Fassbender: We have the evidence Brendan, we just need you ta, ta be honest with us.
Brendan: That he cut off her hair.12
Fassbender: What else was done to her head, (pause)
Brendan: That he punched her.
Wiegert: What else? (pause) What else? (pause)
Fassbender: He made you do something to her, didn’t he? So he would feel better about not being the only person, right? Yea.
Wiegert: mm huh.
Fassbender: What did he make you do?
Wiegert: What did he make you do Brendan? (pause) It’s OK, what did he make you do? (pause)
Dassey: Cut her.
Wiegert: Cut her where?
Brendan: On her throat.
Wiegert: So Steve stabs her first and then you cut her neck (Brendan nods “yes”). What else happens to her in the head?
Fassbender: It’s extremely extremely important you tell us this, for us to believe you.
Wiegert: Come on Brendan, what else? (pause)
Brendan: That is all I can remember.
*968Wiegert: All right, I’m just gonna come out and ask you. Who shot her in the head?
Brendan; He did.
Fassbender: Then why didn’t you tell us that?
Brendan: Cuz I couldn’t' think of it.
R. 19-25, at 60-68.
This example demonstrates how critical the steering was to Dassey’s confession. Recall that the gunshot wounds to the head were unknown to anyone but the investigators and the real killer and thus were key to determining the veracity of the confession. Dassey “couldn’t think of it” and instead launched into a litany of other dubious guesses about actions that might have befallen Teresa. Shooting a living human in the head (or seeing it happen) is not something that a person is likely to forget. Indeed, Dassey later described how he could no longer shoot a gun or go hunting because he had been trau-matizéd by the shooting of his pet cat: “I couldn’t shoot no more ... cuz we used to have a cat that was like somethin’ was wrong with ‘em and we had to shoot ‘em because we didn’t want to pay for the bills ... and my mom told me not to watch when hers nows ex-boyfriend shot it, shot ’em and I couldn’t watch.” Id. at 65-66. But yet despite the impact of the cat incident, Dassey “could not think of it” when asked what was done to Halbach’s head. Clearly his inability to describe the shooting was not an effort to protect himself, as he had just admitted to slitting Halbach’s throat. After guessing many of the most common things that a person might do to a victim’s head—cutting hair, punching, cutting the throat—he simply “could not think of’ anything else that was “done to her head” until Wiegert says, “I’m just gonna come out and ask you. Who shot her in the head?” Id. at 63. Suddenly he “could[ ] think of it.” Id. And of course he had to “think of it” because Fassbender had just told Dassey that it was “extremely important for you to tell us this, for us to believe you.” In other words, finding the right answer was the key to freedom and pleasing the interrogators because “the truth”—meaning what the investigators wanted Dassey to say—would avoid conflict and “set him free.”
This example also reveals the power that the false assumption technique (described more below) had on Dassey. The “who shot her in the head?” question is the proverbial “when did you stop beating your wife?” assumption. And Dassey is quick to respond despite having no idea what happened to Halbach’s head just a few seconds earlier. Likewise, in the following exchange, a confused Dassey falls right into the trap again.13
Fassbender: The first time we talked to you or the second time you talked about cutting off her hair. Where did the hair go? Did you cut off her hair?
Brendan: Yeah.
Fassbender: Where did that happen
Brendan: In the, in the, bedroom.
Fassbender: What ya cut the hair off with?
Brendan: The knife.
Fassbender: The knife you found in the garage?
Wiegert: It doesn’t make sense.
*969Fassbender: It’s impossible. You took her out to the garage and that’s where you got the knife. Explain how that can be. (pause) Did you cut her hair off?
Brendan: No.
Fassbender: Then why did you tell us you did? Brendan?
Brendan: I don’t know.
Fassbender: Do you remember telling us prior? The last time that you saw that stuff in the burn barrel?
Brendan: Yeah.
Wiegert: So why did you do that? Brendan: I had too much stuff on my mind.
Wiegert: So now you remember a little more clearly? OK. How much of her hair did you cut off?
Brendan: A little bit.
Wiegert: You told me a couple of minutes ago you didn’t cut any off. What’s the truth? Did you cut some of her hair off?
Brendan: No.
Fassbender: ... did anyone cut her hair off that night?
Brendan: No. (shakes head no) Fassbender: Where did you get that from? (pause) I mean it seems kind of strange that you just all of a sudden told us you had cut her hair off. Where did you get that from, if it’s not true?
Brendan: I don’t know, I was just guessing.
Fassbender: Why, Did you think that was somethin’ we wanted to hear? Fassbender: Brendan, didn’t did someone some one [sic] cut her hair off that night? Truthfully, for Teresa? Brendan: No. (shakes head “no”)
R. 19-34 at 36-37, 65-66, 98 (emphasis added). In fact investigators never found any evidence of Halbach’s hair on Avery’s bed, his carpet, anywhere in his trailer or the garage.
Investigators also hoped that Dassey would reveal another detail unknown to the public—the fact that the car battery had been detached:
Wiegert: After he put the car there, what do you do next?
Brendan: We walk out.
Wiegert: With, how’s, the license plates were taken off the car, who did that?
Brendan: I don’t know.
Wiegert: Did you do that?
Brendan: (Shakes head “no”) No.
Wiegert: Did Steve do that?
Brendan: Yeah.
Wiegert: Well then why’d you say you don’t know?
Fassbender: Ok, what else did he do, he did somethin’ else, you need to tell us what he did, after that car is parked there. It’s extremely important. (pause). Before you guys leave that car. (pause)
Brendan: That he left the gun in the car.
Fassbender: That’s not what I’m thin-kin’ about. He did something to that car. He took the plates and he, I believe he did something else to that car. (long pause)
Brendan: I don’t know.
Fassbender: OK. Did he, did he go and look at the engine, did he raise the hood at all or anything like that? To do something to that car?
Brendan: (long pause) Yeah.
Fassbender: What was that? (pause)
Wiegert: What did he do Brendan? (Pause)
*970Fassbender: What did he do under the hood, if that’s what he did? (long pause)
Brendan: I don’t know what he did, but I know he went under.
R. 19-25 at 77-79. No reasonable court could read these exchanges and conclude that these ideas came voluntarily from Dassey’s mind.
Although these were the two most egregious, they were not the only examples of the investigators feeding Brendan answers. In the following exchange, Dassey insisted for some time that he had no idea what happened to Halbach’s personal effects. After some leading from Fassben-der, in which Fassbender initiated the idea that there must have been a purse, a cellphone and a camera in the burn barrel, Dassey was able to parrot that he saw these exact three items in the burn barrel. Even the investigators seem concerned about the veracity of his statements, asking him several times to verify the truth, particularly in light of his claim that he saw the items beneath a garbage bag, an item that would have melted within seconds in a fire. Once again, at this point Dassey has no motivation to lie or obfuscate facts about whether Avery burned Halbach’s property in the burn barrel, as he has already admitted several times that he and Avery killed Halbach and burned her body. Nevertheless he altered his answers in response to the cues from the investigators. They told him exactly which items were found in the burn barrel and then cued him to “tell the truth” which, we have established, had the effect on Dassey of meaning “tell us what we want to hear and keep guessing until you get it right.”
Fassbender: OK. We talked last er Monday we talked a little about some things a burn barrel out front do you remember anything about that burn barrel? It’s ah you might wanna be a little more truthful about now.14
Brendan: That it was full of stuff.
Fassbender: Was it burning?
Brendan: Yeah.
Fassbender: Did you put some things in that burn barrel that night?
Brendan: (shakes head “no”) No.
Fassbender: What happened to Teresa’s other personal effects? I mean ah a woman usually has a purse right? (Brendan nods “yes”) Tell us what happened ta that?
Brendan: I don’t know what happened to it.
Fassbender: What happened ta her ah, her cell phone? (short pause) Don’t try-ta ta think of somethin’ just.
Brendan: I don’t know.
Fassbender: Did Steven did you see whether ah a cell phone of hers?
Brendan: (shakes head “no”) No.
Fassbender: Do you know whether she had a camera?
Brendan: (shakes head “no”) No.
Fassbender: Did Steven tell ya what he did with those things?
Brendan: (shakes head “no”) No.
Fassbender: I need ya to tell us the truth.
Brendan: (nods “yes”) Yeah.
Fassbender: What did he do with her her possessions?.
Brendan: I don’t know.
*971Wiegert: Brendan, it’s OK to tell us OK. It’s really important that you continue being honest with us. OK, don’t start lying now. If you know what happened to a cell phone or a camera or her purse, you need to tell us. OK? (Brendan nods “yes”) The hard parts over. Do you know what happened ta those items?
Brendan: He burnt ’em.
Wiegert: How do you know?
Brendan: Because when I passed it there was like like a purse in there and stuff.
Wiegert: When you passed what?
Brendan: The burning barrel
Wiegert: Did ya look inside? (Brendan nods “yes”) Why did ya look inside?
Brendan: Cuz it was full.
Wiegert: What else was in there?
Brendan: Like garbage bags, some
Wiegert: Did you put those things in the burning barrel?
Brendan: (shakes head “no”) No.
Wiegert: Did you actually see those items in the burning barrel? (Wiegert emphasizes the word “see.”)
Brendan: (nods “yes”) Yeah.
Wiegert: Tell me what you saw in there exactly.
Brendan: Like they were buried underneath ah, garbage, a garbage bag that was
Wiegert: How do you know, or how could you see them if they were underneath a garbage bag?
Brendan: Because the garbage bag was like on top like that far off the top. Wiegert: OK. So we have the barrel, (Brendan nods “yes”) OK. Why don’t you look at me for a second, OK. We’ve got the barrel:
Brendan: (nods “yes”) mm huh.
Wiegert: OK and here’s is the top of the barrel (Brendan nods “yes”) and the garbage bag is on top?
Brendan: (nods “yes”) Yeah.
Wiegert: And where are those items you said you saw?
Brendan: Like right underneath there.
Wiegert: Underneath the bag?
Brendan: (nods “yes”)Yeah.
Wiegert: Well, how would you see that?
Brendan: Well, if the bags like that far off the you know the top of the thing you can see though underneath it.
Wiegert: You could see underneath it? (Brendan nods “yes”) What did you see?
Brendan: like a cell phone, camera, purse
Wiegert: Are you being honest with us?
Brendan: (nods “yes”)Yeah.
Wiegert: Did you actually see those items?
Brendan: (nods “yes”) Yeah.
Wiegert: When did you see them?
Brendan: When I came over there with the mail.
R. 19-25 at 95-98 (emphasis added).
Although the government concedes that “Who shot her in the head” was a leading question, it characterizes the rest of the interrogation as a litany of open-ended questions that were corroborated by other evidence. The many examples we have just cited belie that claim. As in the example above, after first denying that he knew what happened to Halbach’s personal effects, and after the investigators cued him, Dassey ultimately said that he saw no more nor less than precisely the three items they mention to him in their questions. But these are merely a few of many instances in which investigators explicitly told Dassey what facts he was to report: “We know the fire was going [when you arrived]” Id. at 23; “I think you went over to his house and then he asked [you] to get his mail.” Id. at 41; “You went inside, didn’t you?” Id. at 41; “Does he ask you [to rape Halbach]? He does, doesn’t he? We *972know. He asks you, doesn’t he?” Id. at 47; “You went back in that room.. .we know you were back there.” Id. at 48; “He asked if you want some, right? ... If you want some pussy?” M; “You were there when she died and we know that.” Id. at 54; “He did something else, we know that.” Id. at 54; “We know that some things happened in that garage, and in that car, we know that” Id. at 71.
The investigators even told Dassey what kinds of language he should use. When Dassey told the investigators that Avery had raped Halbach, Fassbender asked him, ‘What did he say? Did he use those words?” Dassey nodded affirmatively but Wiegert knew that did not sound accurate and cued him why it did not: “Are you sure cuz its usu, not usually the words he uses?” But Dassey nodded and said “yeah.” R. 19-25 at 36. But the next time they ask about the sexual assault, Dassey has figured out what they wanted to hear and they reward him by telling him that now they can start believing him:
Brendan: That he wanted to get some.
Fassbender: Some what?
Brendan: Pussy.
Wiegert: That’s what he said to you? (Brendan nods “yes”) OK.
Fassbender: Now I can start believing you, OK? (Brendan nods “yes”).
Id. at 46.
Now that we have set forth the pattern of questioning (the truth is what the investigators wanted Dassey to say and that truth was linked to pleasing the interrogators and his freedom), we turn to the remaining parts of the confession which likewise influence our decision that no reasonable court, having viewed the interrogation as a whole, could have found that Dassey’s confession was voluntary.
2. Dassey’s characteristics and limitations.
Sixteen-year-old Dassey walked into the interrogation room without a parent, a lawyer, or an advocate to look out for his rights. He had never had any contacts with law enforcement prior to his interviews in this case. As described in the fact section, he was passive, docile and withdrawn. He also suffered from intellectual deficits. His IQ was in the low average or borderline range. He was a “slow learner” with “really, really bad grades,” (R. 19-12 at 66), who received special education services and was the subject of at least three Individualized Education Programs, documents developed for children with special learning needs. Specifically, he had difficulty understanding some aspects of language and expressing himself verbally. He also had difficulties in the “social aspects of communication” such as “understanding and using non-verbal cues, facial expressions, eye contact, body language, tone of voice.” R. 19-12 at 91. Testing also revealed that he had extremely poor social abilities, that he was socially avoidant, introverted and alienated, and that he was likely to be more suggestible than 95% of the population.
3. Assurances and promises.
a. Paternalistic assurances
Sitting across from the young, socially and intellectually challenged Dassey were two seasoned police interrogators. Dassey had no adult advocate, but the investigators sought to fill that role and convince him that they were the adults who were on his side. During the first recorded interview, on February 27, Fassbender set the tone, saying,
I’ve got ... kids somewhat your age, I’m lookin’ at you and I see you in him and I see him in you, I really do, and I know how that would hurt me too.... *973Mark and I, yeah we’re cops, we’re investigators and stuff like that, but I’m not right now. I’m a father that has a kid your age too. I wanna be here for you. There’s nothing I’d like more than to come over and give you a hug cuz I know you’re hurtin’.
R. 19-24 at 5.
The paternal assurances and relationship building continued into the March 1 interview: “I wanna assure you that Mark and I both are in your corner, we’re on your side ... ” R. 19-25 at 16, and “... I’m your Mend right now, but I ... gotta believe in you and if I don’t believe in you, I can’t go to bat for you.” Id. at 23.15 Wiegert repeatedly touched Dassey’s knee in a compassionate and encouraging manner during the March 1 interview. See, e.g., R. 19-44, Ex. 48, Disc 1 at 11:20:28 a.m., 11:29:04 a.m., 11:37:32 a.m., 11:41:09 a.m. In one instance, Wiegert put his hand on Dassey’s knee, leaned forward, and said reassuringly and encouragingly, ‘We already know Brendan. We already know. Come on. Be honest with us. Be honest with us. We already know, it’s, OK? We gonna help you through this, aMght?” Id. at 11:29:04 a.m.; R. 19-25 at 37. He later did this again while saying, “Brendan, I already know. You know we know. OK. Come on buddy. Let’s get this out, OK?” Id. at 11:37:32 a.m.; R. 19-25 at 44. And within a few minutes of this knee touch and appeal to his “buddy,” Dassey confessed to raping Halbach. Id. at 50-51.
The government makes much of the fact that Wiegert stated at the beginning of the interview, “[w]e can’t make any promises,” but that one early admonition was countered by hours and hours of subtle and not so subtle declarations otherwise—the death by a thousand cuts. Moreover, Wie-gert’s full statement was: We can’t make any promises, but wé’ll stand by you no matter what you did.” R. 19-25 at 17. What would a reasonable person make of an admonition not to count on any promises, followed immediately by a clear, unconditional promise? More importantly, what would Brendan Dassey, with his limited intelligence and social skills, think of this admonition linked with a promise?
b. False 'promises of leniency.
After painting the “we’re on your side” backdrop, the investigators brought in the main scaffolding of their approach—the false promises that Dassey would be better off confessing than remaining silent. Some of these promises were problematic in and of themselves—for example, a promise that if Dassey told the truth, he would be set free. The other promises eroded volun-tariness because they were linked to a requirement to “tell the truth,” which, as we have established meant “the version of the story that the investigators wanted to hear.” By linking what the investigators wanted to hear with assurances that those versions would make Dassey “aMght” and “okay,” the confession became not one borne of Dassey’s free will but of the investigators’ wills.
The investigators began the interrogation with a monologue, the theme of which was that Dassey could improve his lot by telling the truth, and culminating in the statement, “Honesty is the only thing that will set you free.” R. 19-25 at 17. As the district court noted, this a biblical idi*974om that many adults would recognize as a figurative expression. Dassey v. Dittmann, 201 F.Supp.3d at 1002. Dassey, however, was not someone who understood idioms and subtle distinctions between literal and figurative language. His school special education reports (prepared long before the crime or trial, for use at school) noted in particular that idioms were an aspect of language that Dassey had trouble understanding. R. 19-20 at 79. This is a juvenile who, after all, when told that his polygraph showed a 98% probability of deception asked, “I passed?” R. 19-38 at 1. And when the tester repeated “It says deception indicated,” emphasizing the word deception, Dassey asked, “That I failed it?” Id. And when drawing pictures of the crime scenes for the detectives, he needed help spelling words like “rack” and “garage.” R. 19-25 at 124, 128. Likewise he was unlikely to understand that the other veiled, subtle promises of leniency were not actual promises. And as we know, a law enforcement officer may not promise a defendant that if he confesses he will be set free. Hadley, 368 F.3d at 749; Rutledge, 900 F.2d at 1129. See also Aleman, 662 F.3d at 906.
The investigators sounded the theme of “truth leads to freedom” again and again in that opening monologue:
• “It’s going to be a lot easier on you down the road, ah, if this goes to trial.”
• “[Hjonesty here Brendan is the thing that’s gonna help you. OK, no matter what you did, we can work through that. OK. We can’t make any promises, but we’ll stand behind you no matter what you did. Ok. Because you’re being the good guy here. You’re the one that’s saying, ’you know what? Maybe I made some mistakes, but here’s what I did.”
• “[T]he honest person is the one who is going to get a better deal out of everything.”
•“If, in fact, you did something, which we believe ... it’s OK, as long as you be honest with us, it’s OK. If you lie about it, that’s gonna be problems.”
R. 19-24 at 17. And of course, there was the most direct promise, “honesty is the only thing that will set you free,” (R. 19-25 at 17).
Promises come in many forms. It is true that the investigators never made the type of explicit and specific promise of leniency that an adult of ordinary intelligence might understand as a promise, such as “if you confess we will make certain that you will not be punished.” But to a suggestible suspect with poor social skills, low IQ, and a limited ability to understand idioms and metaphors, those implied promises, made over and over, had the same effect—an effect that could have been mitigated by the presence of a friendly adult.
Likewise, it was not just the promises to be “set free” that constituted a promise of leniency, but promises that Dassey, who was exceptionally introverted and socially avoidant, could escape the unpleasant conflict and social interaction, by providing what the interrogators wanted to hear. “Honesty” would allow them to be on his side and allow him to “get it all out ... and ... over with” and get out of that interrogation room. See R. 19-25 at 48. Although the furniture in the room may have been soft, the non-stop interrogation by two adults of authority would be very intimidating and anxiety producing to anyone but particularly for someone in the 95th percentile on the scale for social avoidance.
c. Coupling assurances and promises with false assertions of knowledge.
One form of a promise comes from coupling an acknowledgement of the facts *975with an assurance—in other words, stating “we already know everything you did and, even knowing all of that,' everything is going to be okay.” The investigators peppered the entire investigation with assurances that Dassey “was going to be alright,” coupled with acknowledgements that they were making these assurances notwithstanding all of the horrible facts that they already knew. Those pleas promised that the key to unlocking the “you’re going to be alright” result was honesty.
The assurances that Dassey would be “alright” came in many forms: “from what I’m seeing ... I’m thinking you’re all right. OK, you don’t have to worry about things.” R. 19-25 at 16; “[N]o matter what you did, we can work through that.” Id. at 17; “It’s OK. As long as you can, as long as you be honest with us, it’s OK. If you lie about it that’s gonna be problems. OK.” Id.-, “We already know. Just tell us. It’s OK.” Id. at 24; “It’s OK because he was telling you to do it.” Id. at 28; ‘We already know, it’s, OK? We’re gonna help you through this, alright?” Id. at 37; “It’s OK Brendan. We already know.” Id. at 41; “It’s OK, tell us what happened.” Id. at 46; “It’s not your fault.” Id. at 47; “Let’s get it all out today and this will be all over with.” Id. at 48; “It’s OK, what’d you do with it?” Id. at 76; “Brendan, it’s OK to tell us OK.” Id. at 96.
Again, the power came, not from the assurances alone, but the assurances coupled with the false information that the investigators “already knew everything.” The investigators were not merely telling Dassey, “Based upon what you have told us so far, we don’t think you have anything to worry about.” Rather, what they told Dassey was, ‘We already know what happened and you don’t have anything to worry about.”
Those assurances that they already knew everything, linked with the plea for “honesty” were plentiful: “We pretty much know everything!?] [TJhat’s why we’re talking to you again today.” R. 19-25 at 17. “[N]ow remember this is very important cuz we already know what happened that day.” Id. at 19; see also Id. at 23 (“We already know what happened!?]”); “We already know. Just tell us. It’s OK.” Id. at 24; “Come on we know this already. Be honest.” Id. at 26; “Remember we already know, but we need to hear it from you.” Id. at 28; “So just be honest. We already know.” Id. at 30; “We already know, be honest.” Id. at 36; “We already know Brendan. We already know. Come on. Be honest with us. Be honest with us. We already know, it’s, OK? We’re gonna help you through this, alright?” Id. at 37; “It’s OK Brendan. We already know.” Id. at 41; “Cuz, we, we know but we need it in your words. I can’t, I can’t say it.” Id. at 44; “Brendan, I already know. You know we know. OK. Come on buddy. Let’s get this out, OK?” Id. at 44; “Remember, we already know, but we need to hear it from you, it’s OK. It’s not your fault.” Id. at 47; ‘We know you were back there. Let’s get it all out today and this will be all over with.” Id. at 48; “We know what happened, it’s OK.” Id. at 50. (For a more complete list of these assurances, see Dassey v. Dittmann, 201 F.Supp.3d at 1002.)
In one instance, when asking Dassey if he helped Avery put Halbach in the back of her RAV4, Wiegert explicitly assured Dassey, “If you helped him, it’s OK, because he was telling you to do it. You didn’t do it on your own.” R 19-25 at 28. But of course, it would not be “okay” for Dassey to help mutilate and dispose of a corpse simply because Avery told him to do it. And likewise, it could not be any further from “okay” for Dassey to rape Halbach because Avery told him to do it. Yet as they were walking Dassey down the path, step-by-step, to admitting that he had raped Halbach, they stated,
*976Wiegert: What happens next? Remember, we already know, but we need to hear it from you, it’s OK. It’s not your fault. What happens next?
Fassbender: Does he ask you?
Wiegert: He does; doesn’t he?
Fassbender: We know.
R. 19-25 at 47 (emphasis added). And then, as he struggled to tell them the details of the alleged rape, they again assure him, “it’s not your fault, he makes you do it.” Id. at 50. The investigators assured Das-sey that once he revealed the details of the alleged rape, “this will be all over with.” Id. at 48. Similarly, just before Dassey stated that he cut Halbach’s throat, Wie-gert prompted Dassey by telling him, ‘What did he make you do Brendan? It’s OK, what did he make you do?” Id. at 62. Recall, of course, that because Halbach’s body was burned (adding even more atrocity to the crime) there was no forensic evidence that she had been raped or that her throat had been cut.
We can have no confidence that any person, but particularly one with Dassey’s IQ and suggestibility, would think that “you’re going to be alright” and “[l]et’s get it all out today and this will- be all over with” might lead to a life sentence in prison. A life sentence is neither “alright” nor something that would put the matter to rest and be “over with.” And in fact, we need not speculate as to how Dassey would interpret those promises, because we know exactly what Dassey made of them—that if he told the tale, as the interrogators had introduced it to him, he would be released. After confessing to the heinous crimes of raping Teresa Halbach, slitting her throat, and then burning her body, Dassey asked if he would make it to school by 1:29 p.m. so that he could turn in a project he had due in his sixth hour class. R. 19-25 at 89. And later he asked “Am I gonna be at school before school ends?” Id. at 148. When Fassbender asked him at the end of the interrogation if he knows what is going to happen next, Dassey says “I don’t know.” Id. at 144. When they tell him he will be arrested, he responds, “Is it only for one day?” Id. These lamentably naive questions suggest that Dassey counted on these assurances that he would be “okay” to mean that he had a free pass to say whatever he wanted (or, more accurately, whatever he thought the investigators wanted to hear) and would not go to jail. Certainly no adult had warned him otherwise.
Once again we recognize that false promises, like other interrogation techniques, do not, per se, make a confession involuntary. Villalpando, 588 F.3d at 1128. Promises, however, cannot be viewed in a vacuum, but rather assessed as they interact with a defendant’s unique characteristics. A mature adult of ordinary intelligence might always appreciate that regardless of any assurances he has been given that his incriminating statements might put him in prison. But the state appellate court viewed the words of the interrogators alone without reference to Dassey and without looking at their cumulative effect and concluded that those words by themselves did not promise leniency, but rather merely encouraged honesty. This is an unreasonable finding of fact and an unreasonable application of the federal law’s “totality of the circumstances” requirement to those facts.
If, in fact, the state court had looked at those promises, not as they stood alone, but cumulatively and in light of the fact that they were linked to the interrogators’ requirements that Dassey tell them what it was they wanted to hear, it could not have come to any other conclusion but that Dassey’s free will was overcome. And where a defendant’s will is overborne by the circumstances of the interrogation, due process precludes admission of a confes*977sion. Schneckloth, 412 U.S. at 225-26, 93 S.Ct. 2041.
d. The combined effect of the promises.
The false promises—that he will be “alright,” that “it is not his fault” that “the truth will set him free” clearly affected the voluntariness of Dassey’s confession. Villalpando, 588 F.3d at 1128 (“a false promise [of leniency] has the unique potential to make a decision to speak irrational and the resulting confession unreliable.”) The message Dassey heard loudly and clearly was that “the truth” was the key to his freedom, and “the truth” meant those things that the interrogators wanted him to say. Although the point has already been made, we include a few more examples to emphasize how readily apparent the involuntariness of Dassey’s confession ought to have been to any reasonable court reviewing the confession in its totality. Once again, these examples establish a clear pattern of the investigators subtly (or not so subtly) feeding options to Dassey and then admonishing him to “be honest” when his answers do not fit their theory of the case. When Dassey hits upon the correct facts however, interrogators lock in the story by telling him “now we believe you.” In the first example, Wiegert knew there were bullet casings found in the garage, but no bullet holes or shell casings found in Halbach’s vehicle, so he worked to bring Dassey’s answers in line with this evidence. Das-sey’s culpability does not depend on where Halbach was shot. His only stake is in determining what the investigators want “the truth” to be, because the “the truth” is the key to pleasing the interrogators, getting out of the interrogation room, and “setting him free.”
Fassbender: Tell us where she was shot?
Brendan: In the head.
Fassbender: No, I mean where, in the garage?
Brendan: Oh.
Fassbender: Outside, in the house?
Brendan: In the garage.
Fassbender: OK.
Wiegert: Was she on the garage floor or was she in the truck?
Brendan: In the truck.
Wiegert: Ah huh, come on, where was she shot? Be honest here.
Fassbender: The truth.
Brendan: In the garage.
Wiegert: Before she was put in the truck or after?
Brendan: After.
Fassbender: So she’s in the truck and that’s when he shoots her? .(Brendan nods “yes”)
Fassbender: And she was in the back of the truck or SUV the whole time that he shot her?
Brendan: She was on the garage floor. Wiegert: She was on the garage floor. OK.
Fassbender: Alright.
Wiegert: That makes sense. Now we believe you.
R. 19-25 at 72-73.
Similarly, Dassey had no real reason to fabricate what Halbach was wearing, as it neither increased nor decreased his culpability. He did, however, have an incentive to give the investigators the details they were looking for so that he could return to school and home. The investigators, on the other hand, had a description of what Hal-bach was last seen wearing—blue jeans, a white shirt, and a spring jacket, R. 19-18 at 6, and therefore had a weighty incentive to align Dassey’s descriptions with then-known facts. As in the previous example, this exchange contains fact-feeding and pleas for “honesty,” but it also includes a safety valve. When Dassey began making a mess of things, the investigators encour*978aged him to backtrack and say that he could not remember.
Fassbender: Do you remember what she was wearing? I know it’s a long time ago, don’t guess, if you remember, you can say it.
Brendan: (shakes head “no”) I don’t remember.
Id. at 20. Yet later when he receives the cue to “be honest,” and a set of options (t-shirt or button-up) he does seem to recall her clothes. When he gives a conflicting answer, and contradicts himself, he is told just to “say I don’t remember.”
Fassbender: Did she have clothes on? Now be honest. If she did, she did, and if she didn’t, she didn’t. '
Brendan: Sort of.
Fassbender: OK. What did she have on.
Brendan: Like a white T-shirt and that, pants.
Wiegert: What do you mean sort of? Either she had clothes on or she didn’t. It’s, was some of it on some of it off? What?
Brendan: It was ripped.
Wiegert: It was ripped (Brendan nods “yes”) Where was it ripped?
Brendan: Like right here, (pointing to chest)
Wiegert: Was it a T-shirt or button up shirt or what kind of shirt.
Brendan: A button up one.
Wiegert: What color?
Brendan: Like a black one.
Wiegert: OK, before you said there was a white T-shirt. She had that on too? Brendan: Yeah, (nods “yes”)
Wiegert: OK, and in the other interview you said it was blue. Do you remember what color it was? If you don’t remember, say you don’t remember.
Brendan: I don’t remember.
Id. at 31-32.
There is no reason to think that Das-sey’s pattern of guessing at “the truth” until he got it right was any different when the stakes mattered and his culpability was on the line. This is particularly true because the investigators had already assured him that, even knowing what they knew—that is, with “the truth” that they had—Dassey would be “okay.” The examples below demonstrate how these promises affected the voluntariness of Dassey’s confession of the most horrific acts of the crime.
Wiegert: What happens after you were done watching TV for 15 minutes. Brendan: I told him I had to leave cuz I had ta call Travis.
Wiegert: Brendan, be honest You were there when she died and we know that. Don’t start lying now. We know you were there. What happened?
Fassbender: He ain’t gonna lie to you, hey we know that OK.
Wiegert: We already know, don’t lie to us now, OK, come on. What happens next?
Fassbender: You’re just hurting yourself if you lie now.
Brendan: Then he went in, back in there and he stabbed her.
Wiegert: You were with him? (Brendan nods “yes”) Yes?
Brendan: Yeah.
Id. at 54.
Similarly, below, although Dassey had already denied that he had touched or sexually assaulted Halbach, he came to understand that his answer “I didn’t do nothing,’ ” was causing conflict, and was not “the truth” that the investigators want to hear and that would therefore “set him free.”
Wiegert: So you, he, he brings you back there and he shows you her (Brendan nods “yes”) and what do you do? Honestly. Because we think
Fassbender: Very important.
*979Wiegert: We know happened.
Fassbender: It’s hard to be truthful.
Wiegert: We know what happened, it’s
OK. (pause) What did you do?
Brendan: I didn’t do nothin’
Wiegert: Brendan, Brendan come on. What did you do?
Fassbender: What does Steven make you do?
Wiegert: It’s not your fault, he makes you do it.
Brendan: He told me ta do her ... Ta screw her.
Wiegert: Ok. Did you do that? Honestly?
Brendan: Yeah.
Id. at 50.
4. Examples of resistance.
The State makes much of the fact that Dassey resisted the interrogators on many occasions. In fact, the State counts eight occasions in which Dassey resists the interrogators’ suggested response. These exchanges differ markedly from the exchanges in which Dassey shifts or changes his answers. For example, in comparison to the example of the garage floor and seeing Halbach on the porch, the exchange between Dassey and the investigators regarding false information about a tattoo differs significantly in form, length and follow-up. Most importantly, it does not contain the pattern of continual pleas for honesty until the answer changes. In the exchange below, the investigators inserted the false notion that Halbach had a tattoo—-a tactic interrogators are trained to do to test a suspect’s honesty and suggestibility.
Fassbender: Probably when she was alive, did she have any scars, marks, tattoos, stuff like that, that you can remember?
Brendan: I don’t remember any tattoos
But then just seconds later, the following exchange occurs:
Fassbender: Ok. (pause) We know that Teresa had a, a tattoo on her stomach, do you remember that?
Brendan: (Shakes head “no”) uh uh.
Fassbender: Do you disagree with me when I say that?
Brendan: No but I don’t know where it was.
Fassbender: OK.
Id. at 137-39. Rather than explore the subject further, ask where the tattoo was and what it looked like, or admonish Das-sey to “be honest” to encourage him to guess again, Fassbender instead immediately moved on to a new subject. From the investigator’s perspective, no good could have come from further exploration after Dassey had demonstrated a willingness to go along with the idea that Halbach had a tattoo; he just doesn’t “know where it was.” Id. Moreover, Dassey was able to affirm that he did not disagree with the investigators so he was not forced to change his story to agree. Based on our prior examples, however, one can imagine that if Fassbender had continued as he did in other areas, and the next question he asked was “Be honest, did she have a tattoo of a butterfly or a tiger?” Dassey would have responded with one or the other until he found the correct answer.
Second, it is true that at first Dassey is firm about the location of the knife that Avery used to stab Halbach, but once the investigators use the code “tell us the truth” (in other words, “change your story to tell us what we want to hear”), he immediately caved to their suggestion. The State cites the initial response, but not the follow-up where Dassey succumbs. The initial exchange was as follows:
Wiegert: Where was the knife that he used, ’er you used. Where’s that knife go?
Brendan: He left it in the Jeep.
*980Wiegert: He what?
Brendan: He left it in the Jeep.
Wiegert: It’s not in the Jeep now, where do you think it might be?
Brendan: I sure [sic] it was.
Wiegert: Did you see it in the Jeep?
Brendan: Yeah, cuz he set it on the floor.
Wiegert: Where on the floor did he set it?
Brendan: In the middle of the seats.
Wiegert: Okay.
Id. at 80-81.
In that exchange, there was no admonition to tell the truth or inquiries about whether he was certain, as happened in the following exchange where he did, indeed change his answer about the location of the knife:
Wiegert: Wh-What about the knife, where is the knife, be honest with me, where’s the knife? It’s OK, we need to get that OK? Help us out, where’s the knife?
Brendan: Probably in the drawer.
Wiegert: In which drawer?
Brendan: His knife drawer;
Wiegert: And where’s that?
Brendan: In the kitchen.
Wiegert: Is it probably in there, or do you know it’s in there.
Brendan: That’s where I think it is.
Wiegert: Why do you think it’s in there?
Brendan: Cuz he wouldn’t let that knife go.
Wiegert: Cuz he wouldn’t let the knife go. How do you know that?
Brendan: Cuz it was a pretty nice knife.
Id. at 121 (emphasis added).
Third, the State argues that Dassey resisted changing his answer regarding when Avery started the fire despite many questions by investigators. But the conversation about the fire was the very exchange in which the investigators became stern with Dassey and set forth the “rules” for the interview—that is, if Dassey failed to tell them what they wanted to hear, the investigators would reprimand him until he guessed the correct answer. Fassbender tells Dassey precisely what the only acceptable answer will be:
Fassbender: What about the fire?
Dassey: Do you mean if it was started or somethin’? No it wasn’t (shakes his head “no.”)
Fassbender: Ok. We’re not going to go any further in this cuz we need to get the truth out now. We know the fire was going.... Let’s take it through honestly now.
Id. at 23 (emphasis added). The State argues that Dassey continued throughout the interrogation to state that the fire was going when he got there, but of course he did: Fassbender had made it clear from the very start that this was the only answer he would accept.
Dassey does indeed resist suggestions that he kept Halbach’s hair and does so many times. Id. at 102. The problem for the State is that the information about cutting Halbach’s hair came from some of the most suggestive questioning of the whole interrogation—when Fassbender was desperately trying to compel Dassey to tell him what the two of them had done “with [Halbach’s] head.” Id. at 60. In response to Wiegert’s eight questions in succession about what the two had done with Halbach’s head, Dassey says, with a rising intonation usually associated with asking a question, “That he cut off her hair [?[?] ” Id. (question mark added, see R. 19-44, Ex. 43, Disc 1 at 11:57:41 a.m.). It is not surprising that he would deny keeping Halbach’s hair when the notion that he cut her hair was simply one of his unsuccessful apparent' guesses at what had been done to Hal-bach’s head. Given the origin of the hair comment in the first instance and the recantation and then further confusion about *981the hair at the May 13 interview, it is difficult to make anything of Dassey’s comments about hair cutting at all.
Dassey also resisted the investigators’ several inquiries about whether the two of them had used some “wires hanging from the rafters,” in Avery’s garage to “do stuff’ to Halbach in the garage. R. 19-25 at 132-33. This is perhaps the State’s strongest evidence of resistance, as there is no readily apparent reason, apart from the truth, that Dassey resisted their questioning about these wires other than, perhaps, that he was too naive to think of an unimaginably horrible form of torture for which those wires could have been used.
Finally, it is also true that Dassey ardently resisted any suggestion that he shot Halbach or even touched the gun. But he had a reason to do so. He told the investigators that he had been traumatized when his mother’s boyfriend shot his cat and had decided he “couldn’t shoot no more” after that episode. Id. at 65-66. Having made a clear pronouncement to himself and others that he was a person who did not “shoot no more,” he would have been unlikely to have been as suggestible about such a fact.
It was not just Dassey’s ability to resist that the State used to support the volun-tariness of Dassey’s confession, but also the richness of the details he provided and the fact that physical evidence corroborated many of those details. As we noted at the outset, many false confessions contain intricate detail. Garrett, The Substance of False Confessions, 62 Stan. L. Rev. at 1054. And many of the elaborate details Dassey reported were available in the media reports. It had been widely reported in the media that Halbach’s RAV4 was found in the salvage yard partially concealed by branches and a car hood; her remains were found in Avery’s burn pit along with remnants of clothing; Avery burned tires on the night Halbach was last seen; eleven rifle casings were found in Avery’s garage; two rifles were recovered from Avery’s bedroom; a key to Halbach’s RAV4 was found in Avery’s bedroom; the key had Avery’s DNA on it; Avery’s blood was found in Halbach’s RÁV4; and Halbach’s blood was found in the cargo area of the RAV4. Dassey v. Dittmann, 201 F.Supp.3d at 997 (citing newspaper articles).
The State also argues that physical evidence corroborated many of the details to which Dassey confessed, but, in fact, the lack of physical evidence was the weakest part of the State’s case. There was no DNA or other physical evidence linking Dassey to this crime in any way—not a strand of his DNA in the garage, Avery’s bedroom, on the RAV4 or its key, on any knives, guns, handcuffs or any other relevant place. Despite descriptions of a gruesome killing with stabbing, throat cutting, hair cutting, rape, and a shooting, investigators never found a single drop of Hal-bach’s blood, hair or DNA in Avery’s not-so-tidy trailer and garage—not on- the sheets, mattress, carpet, walls, clothing, garage floor, mechanic’s creeper, gun, handcuffs, or bed posts. There was no forensic evidence supporting Dassey’s story that Halbach had been stabbed, raped, bound or cut. Investigators did find Hal-bach’s blood in her vehicle and her DNA on a bullet fragment in Avery’s garage. R. 19-16 at 62-66. The district court pointed out that some of the corroborative evidence had been challenged at trial as being the product of contamination and other unreliable methods. R. 19-27 at 210-32. And in any event, other purportedly corroborative evidence was as harmful as it was helpful. For example, investigators did find handcuffs and leg irons in Avery’s bedroom, but not a single scratch on the wooden bed posts as one would expect were Halbach handcuffed to the bed as Dassey described. R. 19-23 at 88. Other corroborative evidence supported both the state and Dassey’s theories of the events. *982For example, the bleach-stained pants supported the state’s version of the story in which Dassey knowingly helped clean Hal-bach’s blood from the garage, and also Dassey’s version of events in which Avery asked an ignorant Dassey to help clean from the garage floor something that appeared to be automotive fluid. The district court dismissed many of the state’s asserted corroborating details as unhelpful, and we need not repeat the district court’s explanations. See Dassey v. Dittmann, 201 F.Supp.3d at 998.
In sum, the investigators promised Das-sey freedom and alliance if he told the truth and all signs suggest that Dassey took that promise literally. The pattern of questioning demonstrates that the message the investigators conveyed is that the “truth” was what they wanted to hear. When he deviated, they told him he was lying and when he successfully parroted what they wanted him to say, either because he successfully guessed or the investigators had fed him the information, they patted him on the back for telling the truth and told him he would be “okay.” Dassey, however, had trouble maintaining a consistent story except when he was being led step-by-step through the facts, thus confirming that this confession emerged not from his own free will, but from the will of the investigators.
We are quite cognizant that our role in this habeas petition is limited. We have catalogued these parts of the confession not because we might have come to a different conclusion about Dassey’s guilt or innocence, but because they reflect on the totality of the circumstances that the state appellate court should have been considering when assessing whether Dassey’s confession was given of his own free will. By ignoring these false assurances and promises, steering, coaxing, and fact-feeding, the state court, although it knew it must address the totality of the circumstances, failed to apply that rule to these facts. See 28 U.S.C. § 2254 (d)(1). The requirement to view the totality of the circumstances, however, applies to adults and minors alike. See, e.g., Missouri v. Seibert, 542 U.S. 600, 608, 124 S.Ct. 2601, 159 L.Ed.2d 643 (2004). If the admonition to give extra care to juveniles’ confessions means anything, it must mean that a court must give extra scrutiny to a child’s confession. For example, it might ask if this youth was susceptible to steering. Was he fed information? Was he someone who needed an adult ally to explain the consequences of his Miranda waiver or his confession in general? Did he need someone to remind him not to guess at answers to please the interrogators? Did he need someone to remind him that the investigators were police officers with a different agenda than his? Had the state court given Dassey’s confession any of this required care, it simply could not have overcome the many doubts that his confession raises about vol-untariness. We have shown again and again a pattern of.steering, coaxing, fact-feeding and cueing followed by rewarding the “correct” answer, and we urge anyone with doubts about the voluntariness of Dassey’s confession to view the interrogation with this pattern in mind.
By determining that Dassey, under the totality of the circumstances, confessed of his own free will, the court ignored the clear and convincing weight of the evidence and thus made an unreasonable determination of the facts. See 28 U.S.C. § 2254(d)(2).
D. Harmless error.
Moreover, because the confession was essentially the only evidence the State presented against Dassey at trial, we, like the district court, must conclude that allowing its admission could not have been harmless error. Specifically, the violation *983of Dassey’s constitutional rights “had a substantial and injurious effect or influence in determining the jury’s verdict.” Brecht v. Abrahamson, 507 U.S. 619, 637, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993) (internal citations omitted). Indeed, as the district court pointed out, “Dassey’s confession was, as a practical matter, the entirety of the ease against him.” Dassey, 201 F.Supp.3d at 1006. Despite the intensity of the investigation, the brutality of the crime and the disarray of the premises, no one ever found a single hair, a drop of blood, a trace of DNA or a scintilla of physical evidence linking Dassey to this crime.
E. Ineffective assistance of counsel.
Because we affirm the grant of the writ of habeas corpus on these bases, we need not make a determination about the effective assistance of counsel. We note, however, that should the government decide to retry Dassey, the issue of the admissibility of the May 13 telephone call between Das-sey and his mother will require a fresh look to determine whether it is the fruit, so to speak, of an involuntarily-obtained confessional tree.
III.
Teresa Halbach’s family has now grieved for their painful loss through several trials, multiple state court appeals, state post-conviction relief appeals, and now the habeas proceedings in federal court. If only this court, through its many words, could re-write the tragic tale of that final day of Teresa’s life. But of course, we cannot. Dassey has successfully demonstrated that the state court decision resulted in a decision that was “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States” and that “resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d)(1) and (2). The decision of the district court is AFFIRMED in all respects. The writ of habe-as corpus is GRANTED unless the State of Wisconsin elects to retry Dassey within 90 days of issuance of this court’s final mandate, or of the Supreme Court’s final mandate.
. All record cites are to the record in the United States District Court for the Eastern District of Wisconsin, Case No. 14-CV-1310.
. Dassey's lawyer hired an expert who was prepared to testify that the polygraph showed no deception, but the state trial judge excluded any testimony about the polygraph. R. 19-30 at 231-233. The reliability and validity of polygraph evidence is hotly debated in the legal and scientific community. United States v. Scheffer, 523 U.S. 303, 309, 118 S.Ct. 1261, 140 L.Ed.2d 413 (1998). There is not a set standard of scoring for Polygraph examinations. In some numerical scoring systems, "the scores range from 3 for a dramatic reaction to a control question to -3 for the same type of reaction to a relevant question. Noticeable but smaller reactions are scored 1 or -1. A lack of a significant reaction is scored 0. Total scores of 6 or higher indicate truthfulness, while -6 or lower indicate deception. Scores that fall in between are considered inconclusive.” Paul C. Giannelli, Polygraph Evidence: Post-Daubert, 49 Hastings L.J. 895, 909 (1998). The record does not reflect what system O'Kelly used to score Dassey's polygraph examination. R. 19-29 at 21-22.
. Dassey and Avery were tried separately.
. We will refer to the state appellate court decision as “State v. Dassey” and the federal district court opinion on the writ of habeas corpus as "Dassey v. Dittmann."
. The National Registry of Exonerations, False Confessions, http://www.law.umich.edu/ special/exoneration/Documents/Exonerations-in_2016.pdf. at p.3; and http://www.law. umich.edu/special/exoneration/Pages/false-confessions.aspx.
The registry defines exoneration based on specific criteria available at http://www.law. umich.edu/special/exoneration/Pages/glossary. aspx. The summary definition is as follows: an exoneration occurs when a person who has been convicted of a crime is officially cleared based on new evidence of innocence. Id.
. https://www.law.umich.edu/special/ exoneration/Documents/exonerations_us_ 1989_2012_full_report.pdf.
. https ://www.law.umich.edu/special/ exoneration/Documents/exonerations_us_ 1989_2012_full_report.pdf.
. http://scholarlycommons.Iaw.northwestern. edu/iclc/voI95/iss2/5.
. As described in the facts, Janda claimed she was cajoled out of sitting in the interview. R. 19-30 at 155. She remained instead, in the waiting room of the police station.
. Apparently these techniques are not still de rigueur, as Dassey’s interrogation is now used as a "what not to do” in at least one certified interrogation course. See Brief of Amici Curiae, Juvenile Law Center, Wicklander-Zulaw-ski & Associates, Inc. and Professor Brandon Garrett, In Support of Appellee and Affir-mance, at p. 5-6 (citing https://www.wz.com/ 2016/08/19/netflixs-making-a-murderer-involuntary-confession-an-interrogators-perspective/#comment-1266). Of course our consideration of the constitutionality of the interrogation does not hinge on whether companies teaching these courses believe the technique to be effective or proper.
. Another prime example of the investigators telling Dassey exactly what he must say comes from the May 13 interview, which was not used at trial, but is part of the record.
Wiegert: Now where is her truck when you go into the garage.
Brendan: I didn't see it.
Wiegert: ... you can’t say you didn’t see the truck or know where the truck was because ... [t]hat’s just the way it is.
Following this exchange, the investigators launched into a long harangue threatening to leave the interview if Brendan was not "honest with us,” and beseeching Dassey to "do the right thing” for Teresa.
Wiegert: Ok. Then tell us the truth. Fassbender: Let's start with the truck. That’s a good place to stárt. There's other places we're going, but the truck is a good place to start. Tell us the truth about the truck.
Brendan: It was backed into the garage.
R. 19-34 at 21-22.
. We note that Dassey’s intonation rises at the end of this statement, as though he is asking a question. R. 19-44, Ex. 43, Disc 1 at 11:57:41.
. This exchange comes from the May 13 interview which, as we noted earlier, was not used at trial. It was admitted at the state post-conviction proceedings and is part of the record. R. 19-34. We highlight it only as an example of Dassey’s confused responses to leading questions. As a side note, this conversation also serves as a glimpse into the interrogators’ clear efforts to have Dassey move all of the events of the crime to the garage, as no forensic evidence was found in Avery’s trailer.
. In the earlier interview, on February 27, after asking about Avery burning clothes, Fassbender asked, "Did he tell ya anything about a, a, any of her other possessions like I imagine a woman would have a purse, she probably had her cell phone, a camera to take pictures. Did he tell you what he did with those things?” (R. 19-24 at 36). The transcript indicates no answer, but Fassbender follows up with "are you sure?” indicating that Brendan likely shook his head "no.” Id.
. The State portrays these statements as having been made during a time when the investigators still considered Dassey to be a witness rather than a suspect, but prior to this March 1 interview, the investigators thought "it was possible that Brendan might have been involved in the disposal of the corpse.” R. 19-30 at 38. And, as we set forth later, the investigators continued their assurances that Dassey would be "alright” throughout the interview, all of which had been prefaced and contextualized by the early assertions.