DISSENT
KAREN NELSON MOORE, Circuit Judge,dissenting.
There is a narrow question before this court: Should Gary Otte, Ronald Phillips, and Raymond Tibbetts have a trial on their claim that Ohio’s execution protocol is a cruel and unusual punishment, or should Ohio execute them without such a trial? The majority has concluded that there is no need for a trial on the merits of Plaintiffs’ constitutional claim. I disagree.
There is no dispute that the second and third drugs in Ohio’s execution protocol cause immense pain. There is significant evidence that the first drug, midazolam, cannot prevent someone from feeling that pain. After a five-day hearing on Plaintiffs’ motion for a preliminary injunction, the district court determined that there should *893be a full trial on the merits of Plaintiffs’ claim that Ohio’s use of midazolam as the first drug in a three-drug execution protocol creates a constitutionally unacceptable risk of pain. Despite the deferential standard of review that this court should apply, the majority casts aside the district court’s determination that Plaintiffs should have a trial before the state executes them. The majority also determines that despite Defendants’ unequivocal sworn testimony that they would no longer use pancuroni-um bromide or potassium chloride in executions, judicial estoppel does not prevent their renewed attempt to use those drugs. For the reasons discussed below, I would hold that Plaintiffs should have a trial on their Eighth Amendment and judicial-es-toppel claims, and I respectfully dissent.
I. BACKGROUND
Because a key issue in this case is whether the district court made the requisite findings of fact to support a preliminary injunction, I begin by discussing the evidence presented to the district court and the district court’s findings of fact based on that evidence. Over the course of the five-day hearing, the district court heard testimony from four experts: Dr. Craig Stevens, PhD., a Professor of Pharmacology at Oklahoma State University who testified as an expert witness for Plaintiffs; Dr. Sergio Bergese, M.D., a Professor of Anesthesiology and Neurological Surgery and practicing anesthesiologist at The Ohio State University Wexner Medical Center who testified as an expert witness for Plaintiffs; Dr. Joseph Antogni-ni, M.D., a retired anesthesiologist and faculty member at University of California, Davis who testified as an expert witness for Defendants; and Dr. Daniel Buffington, Pharm.D, a pharmacologist in private practice who testified as an expert witness for Defendants. The district court also heard testimony from: Edwin Voorhies, the Managing Director of Operations for the Ohio Department of Rehabilitation and Correction; Gary Mohr, the Director of the Ohio Department of Rehabilitation and Correction; and two Ohio Department of Rehabilitation and Correction Execution Team members (who testified anonymously). A reporter, Alan Johnson, testified as an eyewitness to the execution of Dennis McGuire by the State of Ohio. Five legal professionals testified as eyewitnesses to out-of-state executions in which midazolam was part of a multi-drug execution protocol.
In his 119-page Decision and Order Granting in Part and Denying in Part Plaintiffs’ Motions for Preliminary Injunction, the magistrate judge discussed this testimony and set out his findings of fact. First, the district court discussed the testimony of three eyewitnesses to Ohio’s execution of Dennis McGuire: ODRC Director Gary Mohr, Execution Team Member No. 10, and reporter Alan Johnson. All three testified that after McGuire appeared, to be unconscious, McGuire’s stomach began repeatedly to knot up and then relax, and McGuire began to snort. Decision & Order at 20-21. According to Johnson, “McGuire began coughing, gasping, choking in a way that I had not seen before at any execution.” Id. at 21. Johnson also testified that McGuire gasped “in a way that almost seemed to be choking,” clenched and unclenched his hands, and “attempted to kind of lift up off the table.” Id. Johnson testified that McGuire gasped fifteen or sixteen times, and that the gasping or choking went on for twelve to thirteen minutes. Id. Johnson has witnessed twenty Ohio executions, and had never previously seen anything like the intensity or duration of McGuire’s reaction. Id. Mohr has overseen eleven executions, and testified that he had not previously seen a reaction like McGuire’s. Id. at 20.
*894Next, the district court discussed the testimony of five eyewitnesses to midazo-lam-involved executions that took place outside of Ohio. Two of these out-of-state executions occurred after the Supreme Court’s Glossip decision. Spencer Hahn, an Assistant Federal Defender in the Capital Habeas Unit in the Middle District of Alabama, witnessed the December 8, 2016 execution of Ronald Smith by the State of Alabama. Like Ohio’s current execution protocol, the protocol used to execute Smith called for 500 milligrams of midazo-lam. It also called for a 600-milligram dose of a paralytic drug, and 240 milliequiva-lents of potassium chloride. Decision & Order at 22. Hahn testified that “[tjhere were two periods in which [Smith] appeared to rest somewhat briefly” but then he began “coughing, heaving, flailing, or attempting to flail arms, clenching and un-clenching of fists, movement of lips ... and then doing this asthmatic cough, barking-type cough.” Id. at 22. Terry Alang, an attorney employed as an investigator in the Capital Habeas Unit in the Middle District of Alabama, witnessed the January 20, 2016 execution of Christopher Brooks by the State of Alabama. Alabama used the same execution protocol that it used in the Smith execution, most notably 500 milligrams of midazolam. According to Alang’s testimony, after the execution team members administered midazolam, Brooks’s chest began heaving. Id. at 24.
The district court also discussed testimony about three executions that occurred before the Supreme Court’s Glossip decision. Id. at 22. First, Sonya Rudenstine, a Florida lawyer who specializes in capital post-conviction work, witnessed the execution of Paul Howell by the State of Florida. Like Ohio’s current execution protocol, the protocol used to execute Howell called for 500 milligrams of midazolam in two separate injections of 250 milligrams each. Id. at 23. The protocol then called for 200 milligrams of vecuronium bromide in two 100-milligram injections, followed by 240 milliequivalents of potassium chloride. Id. Rudenstine observed Howell open his eyes after the consciousness check. Id.
Second, Dale Baich, a supervisor in the Federal Defender Capital Habeas Unit in Arizona, witnessed the execution of Joseph Wood by the State of Arizona. Id. The protocol used to execute Wood called for injection of a mixture of 50 milligrams of midazolam and 50 milligrams of hydromor-phone. Id. During Wood’s execution, the State injected this mixture fifteen separate times. Id. “Wood continued to gasp and try to breathe until his death ... almost two hours after the process began.” Id. In a settlement agreement entered on December 19, 2016, Arizona agreed to “never again use midazolam, or any other benzo-diazepine, as part of a drug protocol in a lethal injection execution.” R. 976-2 (Stipulated Settlement Agreement at 2) (Page ID #36214); see also id at 23.
Third, Dean Sanderford, an Assistant Federal Defender in Colorado, witnessed the execution of Clayton Lockett by the State of Oklahoma. Decision & Order at 24. The protocol used to execute Lockett called for 100 milligrams of midazolam followed by a paralytic agent and potassium chloride. Id. According to Sanderford, three or four minutes after the administration of the paralytic, Lockett began writhing and attempted to speak. Id.
The district court then discussed, at great length, the testimony of the four expert witnesses. Dr. Stevens discussed sedation and general anesthesia. He explained that there are different levels of sedation: minimal sedation (i.e., the sedation that would be appropriate for a root canal); moderate sedation; and deep sedation. General anesthesia is beyond the deepest level of sedation, and is the state appropriate for surgery. Only at the level of general anesthesia is someone uncon*895scious. Dr. Stevens explained that midazo-lam can bring someone to the state of deep sedation, but not to general anesthesia or unconsciousness. Decision & Order at 78. Similarly, Dr. Bergese testified that he would never use midazolam alone as an anesthetic. He also testified that when mi-dazolam is used as an anesthetic, it is for relatively minor procedures, such as colo-noscopies, as opposed to more invasive surgeries. Id. at 47.
To explain why midazolam cannot render someone unconscious, Dr. Stevens explained midazolam’s ceiling effect. Id. at 31-32. Midazolam acts on a receptor called GABAa (GABA is short for gamma-amino-butyric acid), and can decrease neural activity only when GABAa is present. Once there is no GABAa left for midazolam to act on, midazolam cannot decrease neural activity anymore and the drug reaches its maximum potency, or ceiling. At this point, administering more midazolam does not increase midazolam’s effect.
Dr. Stevens explained that midazolam’s reliance on GABAa, and consequential ceiling effect, is a distinction between mi-dazolam and barbiturates like thiopental sodium. Id. at 31-32. Midazolam is a ben-zodiazepine, whereas thiopental sodium is a barbiturate. Although both benzodiaze-pines and barbiturates work on the central nervous system and can be used as sedatives, barbiturates can decrease neural activity without GABAa present. According to Dr. Stevens, because barbiturates do not depend on GABAa, they do not have a ceiling effect. Dr. Bergese agreed generally that midazolam has a maximum impact, but he emphasized that his main concern is that midazolam is simply the wrong drug to use. Id. at 87.
In response to Dr. Stevens’s discussion of ceiling effects, Dr. Antognini testified that midazolam’s ceiling effect is not germane. In his view, a 500-milligram dose of midazolam is sufficient to render a person unconscious. Whatever ceiling effect mi-dazolam may have beyond the amount necessary to render someone unconscious is irrelevant. Decision & Order at 70. He also testified that data on midazolam’s ceiling effect is unclear. Id. at 71. Taking an entirely different tack, Dr. Buffington disputed that midazolam has a ceiling effect at all. Id. at 93. However, he also testified that when midazolam is used alone, it is usually in situations where general anesthesia is not required, such as resetting bones, vasectomies, or placement of tubes or implanted devices. Id. at 92.
Dr. Antognini and Dr. Stevens disagreed strongly about whether midazolam possesses any analgesic (painkilling) properties. Dr. Antognini testified that midazo-lam does possess some analgesic properties, at least in massive doses. Dr. Stevens, by contrast, was adamant that midazolam does not treat pain. Id. at 75. Dr. Bergese agreed with Dr. Stevens. Id. at 47. Without addressing midazolam’s analgesic properties, Dr. Buffington said that mi-dazolam would sedate someone sufficiently to render them insensate to the pain caused by a paralytic and potassium chloride. Id. at 94. Dr. Antognini testified that the risk that someone would experience pain after receiving a 500-milligram dose of midazolam is “very, very low.” Id. at 66. Dr. Stevens, by contrast, concluded that “ ‘the use of midazolam as the first drug in a three-drug protocol is highly likely to cause intolerable pain and suffering,’ stemming from the administration of the second and third drugs.” Id. at 40. Again, Dr. Bergese agreed with Dr. Stevens. Id. at 47. The testimony of Dr. Stevens and Dr. Bergese that midazolam does not eliminate pain is unequivocal.
II. DISCUSSION
A. Legal Standards
“A plaintiff seeking a preliminary injunction must establish that he is likely to *896succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor, and that an injunction is in the public interest.” Glossip v. Gross, - U.S. -, 135 S.Ct. 2726, 2736, 192 L.Ed.2d 761 (2015). “The preliminary injunction posture of the present case thus requires petitioners to establish a likelihood that they can establish both that [Ohio’s] lethal injection protocol creates a demonstrated risk of severe pain and that the risk is substantial when compared to the known and available alternatives.” Id. at 2737.
An appellate court must review a district court’s decision granting or denying a preliminary injunction for an abuse of discretion. Ashcroft v. Am. Civil Liberties Union, 542 U.S. 656, 664, 124 S.Ct. 2783, 159 L.Ed.2d 690 (2004). “Under this standard, the court reviews the district court’s legal conclusions de novo and its factual findings for clear error.” Babler v. Futhey, 618 F.3d 514, 520 (6th Cir. 2010); see also Glossip, 135 S.Ct. at 2739. “[A] finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” Anderson v. City of Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 92 L.Ed. 746 (1948)). The clearly erroneous standard “plainly does not entitle a reviewing court to reverse the finding of the trier of fact simply because it is convinced that it would have decided the case differently.” Id. In particular, “when a trial judge’s finding is based on his decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.” Id. at 575, 105 S.Ct. 1504.
Recently, the Supreme Court reiterated the deference owed to a district court’s findings of fact, even when other trial courts have made different findings. “The rule that we review a trial court’s factual findings for clear error contains no exception for findings that diverge from those made in another court. Whatever findings are under review receive the benefit of deference, without regard to whether a court in a separate suit has seen the matter differently.” Cooper v. Harris, — U.S. -, 137 S.Ct. 1455, 1468, 197 L.Ed.2d 837 (2017) (citing Fed. R. Civ. P. 52(a)(6); Hernandez v. New York, 500 U.S. 352, 369, 111 S.Ct. 1859, 114 L.Ed.2d 395 (1991) (plurality opinion)). The Supreme Court continued by explaining that an appellate court “must ask not which court ... had the better view of the facts, but simply whether the court below’s view is clearly wrong.” Id. “[T]he very premise of clear error review is that there are often ‘two permissible’—because two ‘plausible’—‘views of the evidence.’ Even assuming [another] court’s findings capture one such view, the District Court’s assessment may yet represent another. And the permissibility of the District Court’s account is the only question before [an appellate court].” Id. (quoting Anderson, 470 U.S. at 574, 105 S.Ct. 1504).
Finally, the Supreme Court has also instructed appellate courts to err on the side of allowing trials in cases raising constitutional questions. “If the underlying constitutional question is close,” the Supreme Court instructs, an appellate court “should uphold the injunction and remand for trial on the merits.” Ashcroft v. Am. Civil Liberties Union, 542 U.S. at 664-65, 124 S.Ct. 2783.
This last instruction is perhaps the most significant for this case. But the majority *897heeds neither the instruction to review district court findings of fact deferentially, nor the instruction to err on the side of allowing trials in cases raising constitutional questions. The majority has decided to forego a trial on the merits of Plaintiffs’ constitutional claims and to allow the State of Ohio to execute Plaintiffs without such a trial. More egregiously, the majority has decided to forego a trial even though the district court, which has the better view of the evidence, determined that Plaintiffs should have a trial because they are likely to succeed on the merits of their constitutional claim.
B. Eighth Amendment Baze/Glossip Claim
1. Likelihood of success on the merits
I would affirm the district court’s judgment that Plaintiffs were likely to succeed on the'merits of their Eighth Amendment Baze/Glossip claim, which the district court supported with factual findings that, first, Plaintiffs established that Ohio’s mi-dazolam three-drug protocol creates a substantial risk of severe pain and, second, Plaintiffs presented an available alternative method.
a. Substantial risk of severe pain
Under Glossip, to establish that a method of execution violates the Eighth Amendment, prisoners must first “establish that the method presents a risk that is ‘sure or very likely to cause serious illness and needless suffering, and give rise to sufficiently imminent dangers.’ To prevail on such a claim, ‘there must be a substantial risk of serious harm[.]’ ” Glossip, 135 S.Ct. at 2737 (quoting Baze v. Rees, 553 U.S. 35, 50, 128 S.Ct. 1520, 170 L.Ed.2d 420 (2008) (plurality opinion)) (emphasis in original).
The majority is certainly correct that plaintiffs must meet a “heavy burden” to make the “rigorous showing” that a method of execution creates a substantial risk of serious harm. Maj. Op. at 886-87 (quoting Baze, 553 U.S. at 53, 128 S.Ct. 1520). The majority and I disagree about whether the determination that a method of execution creates a substantial risk of serious harm is a factual finding reviewed for clear error or a legal conclusion reviewed less deferentially. The majority characterizes the determination as application of a “legal standard” and reviews it without deference. Maj. Op. at 885-86. My view that this determination is a factual finding that must be reviewed for clear • error comes directly from the language of the Supreme Court. In Glossip, the Supreme Court said that the Oklahoma district court “did not commit clear error when it found that the prisoners faded to establish that Oklahoma’s use of a massive dose of midazolam in its execution protocol entails a substantial risk of severe pain.” Glossip, 135 S.Ct. at 2731. This statement is explicit that a district court’s determination as to whether midazolam entails a substantial risk of severe pain must be reviewed for clear error.
In this case, the magistrate judge—relying on his “superior! ] ... position to make determinations of credibility” and “experience” in “the determination of fact”—evaluated evidence from scientific experts, eyewitnesses to executions, and ODRC employees discussing Ohio’s current execution protocol. Anderson, 470 U.S. at 574, 105 S.Ct. 1504. Based on this evidence, the district court found that Plaintiffs were likely to succeed on their claim that the use of midazolam as the first drug in a three-drug protocol creates a substantial risk of severe pain. Decision & Order at 104-05. The district court found “from both the expert opinions and the lay descriptions comparing executions with a barbiturate as the first drug and midazo-lam as the first drug that the drugs do not *898produce the same effects in those being executed.... [T]hose administered mi-dazolam ... take longer to die and exhibit different bodily behaviors in the process.” Id. at 104. Evaluating the evidence presented to it during the hearing, including the eyewitness testimony and the opposing viewpoints presented by the experts, the district court “conelude[d] that use of mi-dazolam as the first drug in Ohio’s present three-drug protocol will create ‘a substantial risk of serious harm.’ ” Id. at 105.
For several reasons, the district court’s determination was not clearly erroneous. First of all, the district court did not base its opinion on uncertainty, as the majority asserts. Plaintiffs’ experts testified unequivocally that Ohio’s midazolam three-drug protocol is highly likely to cause intolerable pain. Id. at 40, 43, 47, 55. The district court was in the position to make credibility determinations about the competing experts’ testimony, and the district court’s discussion of the experts’ testimony indicates that it found Plaintiffs’ experts to be more credible than Defendants’. See Decision & Order at 103-05.
For example, the district court recognized that Defendants’ experts did not agree with each other about whether mi-dazolam has a ceiling effect, and Dr. Buff-ington did not appear to agree with Dr. Antognini that midazolam has analgesic properties (only that it would sedate someone sufficiently to make them insensate to pain, which is distinct from actually eliminating pain). Id. at 71-75, 93-94. By contrast, Plaintiffs’ experts were in agreement that midazolam does not have analgesic properties, and, although he was less adamant, Dr. Bergese generally agreed with Dr. Stevens that midazolam has a ceiling effect. Id. 31, 87. The specific points of disagreement between Defendants’ experts support the district court’s determination that Plaintiffs’ experts were more convincing.
It is also noteworthy that the district court in this case evaluated evidence that was not available to the Oklahoma district court in Glossip. The district court heard testimony from eyewitnesses to five executions. Two of those, the execution of Christopher Brooks and the execution of Ronald Smith, occurred after the Glossip decision. In addition to providing the district court with information about additional midazolam-involved executions, these recent executions also shed new light on earlier midazolam-involved executions. In Glossip, the Supreme Court noted that neither Lockett nor Wood received the dose of midazolam at issue in the case before it, and that there were problems with the Lockett execution that were not attributable to the drugs used (namely, “the execution team’s inability to obtain an IV access site”). Glossip, 135 S.Ct. at 2746. Taking into account these differences, the Supreme Court said that “[w]hen all of the circumstances are considered, the Lockett and Wood executions have little probative value for present purposes.” Id. This conclusion may have been reasonable given the circumstances at the time, but new circumstances entitle a district court to come to a different conclusion. The Brooks execution, and particularly the Smith execution, in which Smith coughed, flailed, and heaved for several minutes, cast the problems observed in the Lockett and Wood executions in a new light. Unlike Lockett and Wood, both Smith and Brooks were executed using 500 milligrams of midazolam followed by a paralytic drug and potassium chloride (like Ohio’s current protocol). Like Lockett and Wood, witnesses testified that Smith and Brooks moved and heaved during their executions. Witnesses’ testimony that Brooks was heaving and that Smith was heaving, coughing, and flailing could suggest that Lockett’s writhing and Wood’s gasping were attributable to midazolam’s *899inability to prevent the pain caused by paralytic drugs and potassium chloride, rather than to other circumstances.
The district court made the specific finding that these eyewitnesses were credible, even though many “were from legal practices devoted to representing capital clients.” Id. at 24. The district court noted that “their testimony was carefully confined to observations rather than opinions,” in “contrast[ ] with some press characterizations of some of these executions as ‘botched,’ ‘horrendous,’ ‘barbaric,’ and so forth. These witnesses were carefully professional in not adding advocatory characterizations to their observations.” Id. at 24-25.
The majority argues that the district court did not offer enough reasoning in support of its decision. Maj. Op. at 886-87. I agree that ideally the district court would have offered more reasoning in support of its findings. But it is clear that the district court’s specific findings were meant to be read in conjunction with its lengthy discussion of the testimony. And I do not agree that the district court’s 119-page opinion, which included a discussion of the testimony and specific findings of fact, did not provide sufficient reasoning to be entitled to the deference that we must give to district courts’ findings of fact. (The majority recognizes that the district court produced its opinion under tight timelines, but fails to acknowledge that these timelines were imposed by the State of Ohio.) In my view, the district court’s finding that “use of midazolam as the first drug in Ohio’s present three-drug protocol will create ‘a substantial risk of serious harm,’ ” Decision & Order at 105, was not clearly erroneous, and we are bound by this finding.
b. Availability of an alternative
To succeed on their Eighth Amendment claim, Plaintiffs must also show that there is “an alternative that is ‘feasible, readily implemented, and in fact significantly reduce[s] a substantial risk of severe pain.’ ” Glossip, 135 S.Ct. at 2737 (quoting Baze, 553 U.S. at 52, 128 S.Ct. 1520) (alteration in original). Glossip explicitly states that whether an alternative method of execution is available is a “factual finding” subject to the “clearly erroneous” standard of review. Glossip, 135 S.Ct. at 2738. Other than defining “availability” as a factual finding, the Supreme Court provides little guidance as to the definition of “availability.” As the district court observed, “In Baze and Glossip, the Supreme Court did not attempt to quantify how available the alternative method must be to qualify.” Decision & Order at 107.
Plaintiffs proposed two alternative execution methods. For different reasons, each of these alternatives warrants a remand for a trial on the merits. As to the first alternative, a one-drug protocol using compounded pentobarbital, the district court found that by proposing compounded pentobarbital, “Plaintiffs have met their burden to identify a sufficiently available alternative method of execution to satisfy Baze and Glossip.” Decision & Order at 107. Ohio does not currently have pento-barbital on hand and it cannot purchase pentobarbital to use in executions directly from drug manufacturers. However, according to the district court, Ohio has taken key steps toward acquiring compounded pentobarbital, including passing secrecy statutes “to protect the anonymity of potential suppliers and compounders,” and applying for the import license necessary to purchase pentobarbital’s active ingredient. Id. This court resolved litigation over the secrecy statutes and entry of a protective order in Ohio’s favor. See In re: Ohio Execution Protocol Litig. (Fears v. Kasich), 845 F.3d 231, 240 (6th Cir. 2016); Phillips v. DeWine, 841 F.3d 405, 420 (6th Cir. 2016). These favorable resolutions facilitate the State’s access to compounded *900pentobarbital. The district court also noted that Dr. Buffington, who helped develop Ohio’s current execution protocol and who testified about the content of an affidavit he submitted in an Alabama case, “stated in his affidavit in that case that since other state's had been able to procure compounded pentobarbital for their executions, he believed it could be obtained.” Decision & Order at 95.
Because Defendants’ own expert testified that compounded pentobarbital could be obtained, because Ohio succeeded in passing secrecy statutes and securing a protective order for the specific purpose of obtaining compounded pentobarbital, and because Ohio is currently undertaking the steps necessary to secure compounded pentobarbital, I am not “left with the definite and firm conviction” that the district court erred when it found that compounded pentobarbital is an available alternative. Anderson, 470 U.S. at 573, 105 S.Ct. 1504 (quoting United States Gypsum Co., 333 U.S. at 395, 68 S.Ct. 525). I would defer to the district court’s finding that compounded pentobarbital is available.
For the second alternative, Plaintiffs proposed a two-drug protocol using mi-dazolam and potassium chloride without a paralytic drug. Having decided that compounded pentobarbital is an available alternative, the district court did not make a finding as to whether the second alternative is available, or as to the more difficult question whether the second alternative would “significantly reduce[ ] a substantial risk of severe pain.” Glossip, 135 S.Ct. at 2737 (quoting Baze, 553 U.S. at 52, 128 S.Ct. 1520). Particularly if the majority is correct that compounded pentobarbital is not an available alternative, Plaintiffs are entitled to a finding as to whether the second alternative satisfies the Baze/Glos-sip standard. Instead, the majority has determined that Plaintiffs should be executed with Ohio’s three-drug protocol without a court ever deciding whether their proposed two-drug protocol would significantly reduce the substantial risk of severe pain.
There may be other possible execution methods that, if given a trial, Plaintiffs could prove are available and significantly reduce a substantial risk of severe pain. As anyone who reads the newspaper knows, the nation, and Ohio in particular, is in the midst of a virulent drug-death epidemic. See, e.g., Kristine Phillips, Drugs Are Killing So Many People in Ohio That Cold-Storage Trailers Are Being Used As Morgues, Wash. Post (Mar. 16, 2017), http:// wapo.st/2mNjFEp?tid=ss_mail& utm_term=.e29b9f46cfc8 (“As with much of the United States, Ohio is in the throes of a heroin and opioid epidemic that shows no signs of abating.... The drug epidemic also has caused [Stark] county to spend roughly $75,000 a year in toxicology tests alone.... ‘We’re just spending all kinds of money on lab work because there’s so many different drugs,’ [Stark County investigator Rick Walters] said.”); Kimiko de Freytas-Tamura, Amid Opioid Overdoses, Ohio Coroner’s Office Runs Out of Room for Bodies, N.Y. Times (Feb. 2, 2017), https://nyti.ms/2k0DV2Z (“On Thursday, only two days into February, the coroner’s office in Dayton, Ohio, had already handled 25 deaths—-18 caused by drug overdoses. In January, the office processed 145 cases in which the victims’ bodies had been destroyed by opioids.”); Katharine Q. See-lye et al., Inside a Killer Drug Epidemic: A Look at America’s Opioid Crisis, N.Y. Times (Jan. 6, 2017), https://nyti.ms/2k211F 0 (“Public health officials have called the current opioid epidemic the worst drug crisis in American history, killing more than 33,000 people in 2015. Overdose deaths were nearly equal to the number of deaths from ear crashes. In 2015, for the first time, deaths from heroin alone surpassed gun homicides.”); Courtney Astolfi, *901Report: Ohio Ground-Zero for Opioid Overdose Deaths, Cleveland.com (Dec. 1, 2016), http://s.cleveland.com/01L8JFD (“The Buckeye State topped the list of opioid overdose deaths among all 50 states, racking up 2,106 deaths in 2014.”). Given these reports, there are obviously multiple drugs that could be used to execute people.
The district court’s findings in this case were not clearly erroneous. The district court did not clearly err by finding that Ohio’s current three-drug protocol creates a substantial risk of severe pain, and it did not clearly err by finding that alternative protocols are available. Not only did the district court not clearly err in its factual finding that compounded pentobarbital is an available alternative, but there are other possible alternative protocols that no court has ruled on. Plaintiffs proposed a two-drug protocol, and the district court did not rule on this alternative. Moreover, there may be other drug protocols that Plaintiffs would propose if given an opportunity to litigate this case fully. Therefore, Plaintiffs have satisfied the first requirement for a preliminary injunction by demonstrating a substantial likelihood of success on their Eighth Amendment claim.
2. Likelihood of irreparable harm, balance of equities, and public interest
Having determined that Plaintiffs failed to show a likelihood of success on the merits of their Eighth Amendment claim, the majority does not assess whether Plaintiffs have satisfied the other requirements for a preliminary injunction. Because in my view Plaintiffs did show a likelihood of success on the merits, I address the other three requirements. Plaintiffs satisfy each one.
Plaintiffs would suffer irreparable harm if executed by a drug protocol later determined to be unconstitutional. “The key word in this consideration is irreparable.” Babler, 618 F.3d at 523-24 (quoting Sampson v. Murray, 415 U.S. 61, 90, 94 S.Ct. 937, 39 L.Ed.2d 166 (1974)). “A plaintiffs harm from the denial of a preliminary injunction is irreparable if it is not fully compensable by monetary damages.” Obama for Am. v. Husted, 697 F.3d 423, 436 (6th Cir. 2012) (quoting Certified Restoration Dry Cleaning Network, L.L.C. v. Tenke Corp., 511 F.3d 535, 550 (6th Cir. 2007)). Being executed by a method of execution that is later determined to be unconstitutional is quintessentially an injury that is not fully compensable. Or, as the district court noted, “[t]he irreparable harm to the named Plaintiffs if temporary injunctive relief is not granted is patent”; “[wjhether or not Plaintiffs’ claims survive their deaths, the injury would be irreparable.” Decision & Order at 116.
For the same reason, the balance of the equities favors Plaintiffs. Although “a State retains a significant interest in meting out a sentence of death in a timely fashion,” Nelson v. Campbell, 541 U.S. 637, 644, 124 S.Ct. 2117, 158 L.Ed.2d 924 (2004), the harm from a delay in meting out a death sentence is not an irreparable harm. As a result, balancing the equities counsels in favor of delaying executions until a full trial on the merits can be held on the method of execution.
Finally, the public interest favors Plaintiffs. The public has an interest in sentences being carried out, but it also has an interest in ensuring that those sentences are carried out in a constitutional manner. “[I]t is always in the public interest to prevent violation of a party’s constitutional rights.” Deja Vu of Nashville, Inc. v. Metro. Gov’t of Nashville & Davidson Cty., 274 F.3d 377, 400 (6th Cir. 2001) (quoting G & V Lounge, Inc. v. Mich. Liquor Control Comm’n, 23 F.3d 1071, 1079 (6th Cir. 1994)). As the district court stated, “[o]n balance, the public interest weighs in favor *902of granting temporary injunctive relief, but maintaining a fast track approach to adjudicating Plaintiffs’ claims on the merits.” Decision & Order at 118.
C. Judicial Estoppel
I would also hold that the district court did not err by deciding that Defendants were judicially estopped from reverting to an execution protocol that includes pancu-ronium bromide (a paralytic agent) and potassium chloride (which stops the heart). I begin by reviewing in detail the facts relevant to Plaintiffs’ judicial-estoppel claim.
Litigation challenging Ohio’s lethal injection protocol commenced in 2004. The first events relevant to Plaintiffs’ judicial-estoppel claim occurred in 2009. On October 19, 2009, the district court entered a stay of Kenneth Biros’s execution. R. 965-16 (10/19/2009 Order at 1-4) (Page ID #34294-97). A trial had been scheduled for November 2, 2009, and Biros’s execution’s date had been set for December 8, 2009. Id. at 1-2 (Page ID #34294-95). As of October 19, 2009, there was outstanding discovery, including discovery concerning the failed attempt to execute Romell Broom and the State’s consideration of a new execution protocol. Id. Because of the outstanding discovery, the district court postponed the trial date and entered a “stay of [Biros’s] execution.” Id. at 1-3 (Page ID #34294-6). The district court reasoned that “[gjiven the issues involved and the instruction of the appellate court, Biros is ... entitled to a stay affording him time for discovery and to be heard at trial on the merits of his claims.” Id. at 3 (Page ID #34296).
On October 27, 2009, the State filed a Notice of Appeal “from the Court’s Opinion and Order granting an injunction to intervenor Kenneth Biros, which was filed on October 19, 2009.” R. 965-18 (Notice of Appeal at 1) (Page ID #34304). Also on October 27, the State filed a motion asking this court to vacate the district court’s order delaying Biros’s execution, which the State variously referred to as a stay and a preliminary injunction. R. 965-19 (Defs-Ap-pellants’ Mot. to Vacate Prelim. Inj. Granted to Biros at 1-9) (Page ID #34307-15). In its motion, the State took issue with the district court making the determination that outstanding discovery necessitated a stay of execution without considering Biros’s likelihood of success on the merits. The State argued that “[a] condemned prisoner cannot obtain a stay of execution ... absent a finding by the court that the prisoner is likely to succeed on the merits of his claims.” Id. at 6 (Page ID #34312). Because “a party seeking a preliminary injunction must demonstrate, among other things, a likelihood of success on the merits!,] • • • [w]here a condemned prisoner seeks a stay of execution to permit litigation of a claim that the state’s method of execution will violate the Eighth Amendment, the likelihood or lack thereof of the prisoner’s success on the merits is not only a necessary consideration, but may well be sufficient to resolve the matter.” Id.
On October 29, 2009, the district court issued a second order which, “[i]n light of Defendants’ characterization of [the district court’s] actions and in an effort to assist the Sixth Circuit Court of Appeals in considering the appeal, ... further memorialize[d] the substance of the October 19, 2009 conference.” R. 966 (10/29/2009 Order at 1) (Page ID #34318). In the order, the district court noted that Defendants “helped develop, along with Plaintiffs’ counsel, proposed language to be included in the October 19, 2009 Order. In fact, Defendants’ counsel asked the Court not to characterize the stay as an injunction and explained that they did not want the court to make a finding of unconstitutionality in regard to the stay.” Id. at 1-2 (Page ID #34318-19). The district court surmised *903that Defendants did not want it to make a finding as to Biros’s likelihood of success on the merits of his constitutional claims because “[s]ueh Rule 65 injunctive relief analysis would have necessitated the Court discussing in detail in a written decision its review of the numerous deposition transcripts of witnesses involved in the attempted execution of Romell Broom.” Id. at 2 (Page ID #34319). The court also ordered that “all future conferences, except those dealing with protected discovery material, shall be held in open court and on the record” “[t]o avoid creating an incorrect impression of the events of this litigation and to facilitate clarity as to the parties’ public positions.” Id.
On November 13, 2009, the State announced its intention to change its execution protocol effective no later than November 30, 2009. In its News Release, the State, through ODRC Director Terry Collins, said, “ ‘The previous method of execution included a three-drug protocol applied intravenously. The first change to the execution procedure includes the adoption of a one-drug protocol, using thiopental sodium alone, applied intravenously. Pancuronium bromide and potassium chloride will no longer be used as a part of the process.’ ” R. 966-1 (11/13/2009 ODRC News Release) (Page ID #34322).
On the same day, the State filed in the district court a motion for summary judgment. In its motion, the State argued that, as a result of changes to the execution protocol, “Defendants have negated all of Plaintiffs’ claims” and “Plaintiffs’ challenges to defendants’ previous ‘three-drug protocol’ are moot.” R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 4) (Page ID #34328). The State explained,
It is readily apparent here that the recent changes to defendants’ execution procedures have rendered moot plaintiffs’ constitutional challenges to the “three-drug protocol” previously used by defendants to execute condemned prisoners. The issues presented by plaintiffs’ complaints stem from the alleged risk of severe pain which could be caused by the use of pancuronium bromide and potassium chloride, the second and third drugs in the so-called “three-drug protocol,” in the event that the first drug, thiopental sodium, is' not properly administered. In view of the new procedures’ elimination of the second and third drugs, the issues presented in plaintiffs’ suits are no longer actionable .... Moreover, there is no possibility here that the allegedly unconstitutional conduct will reoccur, or that there is any lingering effects of previous allegedly unconstitutional conduct. There is absolutely no reason to believe that defendants will reinstate the previous “three-drug protocol” if the plaintiffs’ suits were dismissed. And, more importantly, if defendants execute plaintiffs using the revised procedures, defendants cannot “go back to their old ways” and execute plaintiffs using the prior procedures.
Id. at 5 (Page ID #34329) (emphasis in original). The State attached to its summary-judgment motion an affidavit of Director Collins, in which he swore, “[Gjoing forward, pancuronium bromide no longer will be used as part of the lethal injection process. Also, potassium chloride no longer will be used as part of that process.” R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335).
In the summary-judgment motion, the State repeatedly argued that the claims of all Plaintiffs were moot as a result of the change to the execution protocol, R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329), and Collins’s affidavit stated that pancuronium bromide and potassium chloride would not be used “going forward,” R. 966-3 (Collins Aff. at ¶ 6) (Page ID *904#34335). The State sought judgment as a matter of law on the claims of all plaintiffs, not only Biros.
On November 16, 2009, the State filed in this court a reply in support of its October 27 motion to vacate the stay. R. 966-4 (Defs-Appellants’ Mem. Reply to Biros’ Mem. in Opp’n to Defs’ Mot. Vacate District Ct.’s Stay of Biros’ Execution, Sche’d for Dec. 8, 2009, and Defs’ Mem. in Opp. to Biros’ Mot. Dismiss Defs’ Appeal at 1) (Page ID #34338). In the reply, the State argued that “Biros’s lawsuit is moot.” Id. at 7 (Page ID #34344). The State explained that Collins, “has directed changes in the procedures used to carry out the execution of condemned prisoners. The changes include the discontinuation of the use of pancuronium bromide and potassium chloride in the execution process.” Id. The State argued that as a result of this change, “Biros’ suit no longer presents a case or controversey [sic], as the ‘three-drug protocol’ he challenges is no longer used.” Id.
Unlike in the district court summary-judgment motion, the State’s reply in our court argued that “Biros’ suit” was moot, but did not address claims of other Plaintiffs. This focus on Biros is in keeping with the narrowness of that appeal, in which the only issue was the stay of Biros’s execution, not the underlying merits of Plaintiffs’ challenge or any other Plaintiffs’ individual procedural claims. However, although the State did not mention the other Plaintiffs in the body of the reply, the State did attach its summary-judgment motion from the district court as an exhibit to the reply. R. 966-4 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched., filed as Ex. A to Defs-Ap-pellants’ Mem. Reply to Biros’ Mem. in Opp’n to Defs’ Mot. Vacate District Ct.’s Stay of Biros’ Execution, Sche’d for Dec. 8, 2009, and Defs’ Mem. in Opp. to Biros’ Mot. Dismiss Defs’ Appeal) (Page ID #34348-59). It also addressed Biros’s “suit,” as opposed to his claims, which may suggest that it had the entire lawsuit in mind, which involved multiple plaintiffs. As noted above, the summary-judgment motion argued that the claims of all the plaintiffs were moot, and sought judgment as a matter of law on all claims.
On November 25, 2009, a panel of this court vacated the district court’s stay of Biros’s execution. The panel held that “the district court’s stay order must be vacated because any challenge to Ohio’s three-drug execution protocol is now moot.” Cooey v. Strickland, 588 R.3d 921, 923 (6th Cir. 2009). The panel explained that, “the question at hand is whether Ohio will use the old procedure, or the new one, in executing Biros.” Id. In response to this question, the panel maintained that “[tjhere is no basis in the record or for that matter in common sense for assuming that the State will do anything other than what it has told us in court filings and what it has told the public at large: it has changed its execution protocol, and it intends to apply the substantially modified protocol to Biros.” Id.
On December 4, 2009, this court denied rehearing en banc. The concurrence with denial of rehearing en banc posited that, “At a minimum, the new protocol ‘likely’ moots the old challenge, and that is enough to create a likelihood-of-success problem for Biros when it comes to premising a request for a stay on orders related to a different protocol.” Cooey v. Strickland, 588 F.3d 924, 925 (6th Cir. 2009) (Sutton, J., concurring in denial of reh’g en banc). Dissents from denial of rehearing en banc pointed out that nothing prevented the State from going back to the prior execution protocol, which, they explained, fatally undermined the holding that the challenge was moot. Id. at 925-26 (Moore, J., dissenting from denial of reh’g en banc) *905(“Although there is little indication that the State will continue to use the initially challenged three-drug cocktail now that it has developed a new procedure, in analyzing whether Biros’s claim is moot, we must consider whether anything would prevent the State from doing so.... Although we have no reason to doubt Ohio’s sincerity, determining mootness based on a litigant’s statement that it has no reason to resume the challenged activity, no matter how earnest, is not part of the mootness analysis.”); see also id. at 928 (Martin, J., dissenting).
Neither the panel opinion nor the concurrence with denial of rehearing en banc clarified the breadth of the court’s holding. It is not clear whether the panel held that the challenge to Ohio’s lethal injection protocol was moot as to Biros or was moot as to all of the Plaintiffs. It is unclear, first, because in this court the State was ambiguous about whether it was arguing that the claims were moot as to Biros or all Plaintiffs. It is unclear, second, because neither the panel opinion nor the concurrence with the denial of rehearing en banc explicitly stated whether the claims were moot as to Biros or all Plaintiffs.
Subsequently, Biros’s execution was again set for December 8, 2009, and Biros challenged the November 30, 2009 one-drug execution protocol. On December 7, 2009, this court considered Biros’s challenge to the new protocol, and, affirming the district court, declined to stay his execution. Cooey v. Strickland, 589 F.3d 210, 221, 234 (6th Cir. 2009). The State executed Biros on December 8, 2009.
On December 9, 2009, the district court held a hearing. At that hearing, the district judge “suggest[ed] that all of these motions, as a result of the November 30, 2009, new protocol, are moot and should be withdrawn. I’m talking about the defendants’, the plaintiffs’ everything; that the plaintiffs should amend all of their complaints based upon the new protocol and we proceed from that standpoint.” R. 966-10 (12/9/2009 Hr’g Tr. at 25-26) (Page ID #34453-54). The district court added, “I actually can’t demand that you withdraw something, and wouldn’t do that, but I am suggesting that almost everything that’s been filed in this case up until now is moot.” Id. at 26 (Page ID #34456). Addressing this court’s decision on mootness, the district court added, “And I’m not going to get into an argument over mootness like the Court of Appeals has done recently. I’m not going to get into that mess, as I’m sure Judge Sutton would not like to get back into that mess.” Id. Instead, he explained, “I’m trying to suggest a way in which the record can get cleaned up and where we present arguments, present with new arguments, that have anything to do with the new protocol. It just seems to me to be the better way in which to proceed in this case, but, again, it’s up to you guys how we decide this.” Id. After some discussion, attorneys for both sides agreed to withdraw their pending motions, with the understanding that Plaintiffs would file amended complaints challenging the November 30, 2009 protocol. Defendants agreed not to assert a statute-of-limitations defense to Plaintiffs’ amended complaints, and the district court granted leave to Plaintiffs to amend their complaints. Id. at 43, 46 (Page ID #34471, 34474).
Litigation proceeded, and so did executions. After the execution of Kenneth Biros on December 8, 2009, Ohio executed an additional twenty people until the State halted executions after the Dennis McGuire execution in 2014. However, prior to the McGuire execution, the State replaced the November 30, 2009 protocol with a protocol providing for a single-injection of midazolam and hydromorphone. See R. 323 (10/10/2013 Ohio DRC Execution Protocol, 01-COM-11 at 1-19) (Page ID *906#9568-86). McGuire’s execution using the October 10, 2013 protocol prompted .questions about midazolam and caused Ohio again to change its protocol, this time to the midazolam protocol at issue in this case.
The State decided that it would switch to the current midazolam three-drug protocol months before it revealed this switch to Plaintiffs, the district court, or the public. See R. 941 (Hr’g Tr. at 800-01) (Page ID #31862-63). Director Mohr admitted that it was a strategic decision to conceal the switch. Id. at 803 (Page ID #31865).
The “rule[ ] known as judicial estoppel” provides that “[w]here a party assumes a certain position in a legal proceeding, and succeeds in maintaining that position, he may not thereafter, simply because his interests have changed, assume a contrary position, especially if it be to the prejudice of the party who has acquiesced in the position formerly taken by him.” New Hampshire v. Maine, 532 U.S. 742, 749, 121 S.Ct. 1808, 149 L.Ed.2d 968 (2001). “We review de novo a district court’s decision regarding the application of judicial estoppel.” Javery v. Lucent Techs., Inc., 741 F.3d 686, 697 (6th Cir. 2014).1 Three factors “typically inform the decision whether to apply the [judicial estoppel] doctrine.” New Hampshire v. Maine, 532 U.S. at 750, 121 S.Ct. 1808. “First, a party’s later position must be ‘clearly inconsistent’ with its earlier position.” Id. (quoting United States v. Hook, 195 F.3d 299, 306 (7th Cir. 1999)). “Second, courts regularly inquire whether the party has succeeded in persuading a court to accept that party’s earlier position, so that judicial acceptance of an inconsistent position in a later proceeding would create ‘the perception that either the first or the second court was misled.’ ” Id. (quoting Edwards v. Aetna Life Ins. Co., 690 F.2d 595, 599 (6th Cir. 1982)). “A third consideration is whether the party seeking to assert an inconsistent position would derive an unfair advantage or impose an unfair detriment on the opposing party if not estopped.” Id. at 751, 121 S.Ct. 1808.
The first factor for judicial estoppel is satisfied. The State’s earlier position is “clearly inconsistent” with its current position. Id. at 750, 121 S.Ct. 1808. The State represented to the district court and this court that it would no longer use pancuro-nium bromide or potassium chloride for executions. The Director of the ODRC swore that “going forward, pancuronium bromide no longer will be used as part of the lethal injection process” and that “potassium chloride no longer will be used as part of that process.” R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335). In its motion for summary judgment, the State represented to the district court not only that it had “eliminate[ed] ... the second and third drugs” but that “there is no possibility here that the allegedly unconstitutional conduct will reoccur.” R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329). The State’s motion explicitly stated that “[t]here is absolutely no reason to believe that defendants will reinstate the previous ‘three-drug protocol.’ ” Id. The State’s motion also argued that its decision to stop using pancuronium bromide and potassium *907chloride mooted Plaintiffs’ claims. At the December 9, 2009 hearing, the State reasserted its promise that it would stop using pancuronium bromide and potassium chloride, and said that because of this promise, “[t]o the extent that the other motions are based on the old protocol, we think it’s appropriate that they be dismissed as moot or withdrawn.” R. 966-10 (12/9/2009 Hr’g Tr. at 43) (Page ID #34471).
The State’s representations that there was “no possibility” of reverting to a three-drug protocol using pancuronium bromide or potassium chloride and Director Collins’s sworn statement that the State would not use these two drugs “going forward” are inconsistent with the State’s current position. R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335); R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329). The State’s current execution protocol, which it is seeking to use in executing Otte, Phillips, and Tibbetts, includes pancuronium bromide and potassium chloride. R. 667-1 (Ohio DRC Execution Protocol, 01-COM-ll at 2) (Page ID #19813). By repeatedly representing that it would no longer use pancuronium bromide or potassium chloride in executions but now attempting to execute condemned inmates with these very drugs, the State has taken directly contradictory positions.
The second factor, whether the State succeeded in persuading a court to accept its earlier position that it would not use pancuronium bromide or potassium chloride in executions, is the most difficult. See New Hampshire v. Maine, 532 U.S. at 750-51, 121 S.Ct. 1808. As noted above, this court did not make clear whether its November 25, 2009 decision held that the claims of all Plaintiffs were moot or only Biros’s claims were moot. At the December 9, 2009 hearing, the district court expressed its view that all of the motions pending as of December 9, 2009 were moot, but the district court also stated that it was not going to get into the “mess” over mootness, and urged the parties to withdraw their pending motions as a way “the record can get cleaned up.” R. 966-10 (12/9/2009 Hr’g Tr. at 26) (Page ID #34454). The State withdrew the November 13, 2009 motion for summary judgment that argued mootness, but it is not clear whether the State withdrew that motion because the motion’s argument had already been successful or because the district court was not going to entertain the motion’s argument.
If this court held that the claims of all Plaintiffs were moot, then the State’s mootness argument was successful, regardless of whether the State withdrew the motion. Similarly, if the district court held that all Plaintiffs’ claims were moot, then the State’s mootness argument was successful, notwithstanding the fact that as a procedural matter the State withdrew the motion. On the other hand, if this court’s holding applied only to Biros and the district court prompted the parties to withdraw their motions because of practical concerns rather than a determination that the claims of all Plaintiffs were moot, then the State’s mootness argument was not successful.
Ultimately, it appears that the State succeeded in persuading at least the district court, if not also this court, that the claims of all the Plaintiffs were moot. At the December 9 hearing, the district court repeatedly emphasized its view that all the motions pertaining to the old protocol were moot, and encouraged the parties to withdraw their motions for precisely that reason, even if it offered practical reasons as well. R. 966-10 (12/9/2009 Hr’g Tr. at 25-26) (Page ID #34453-54). The State also expressed its view that Plaintiffs should withdraw their motions because they were moot. Id. at 43 (Page ID #34471). Based *908on the statements of the district court and the State, and after some hesitation, Plaintiffs withdrew their challenge to the three-drag protocol. Id. at 42, 46 (Page ID #34470, 34474).
Significantly, the Plaintiffs’ withdrawal of their challenge to the old protocol cleared the way for the State to proceed with executions. After Biros’s execution, the State executed twenty other individuals until it halted executions in the wake of the McGuire execution. The fact that Ohio no longer had to litigate the constitutional-; ity of its three-drag protocol and was able to proceed with executions beginning in December 2009 using other protocols indicates that its mootness argument succeeded. Resuming executions was the State’s ultimate goal in the litigation, and it achieved that goal by affirmatively stating that it was no longer going to use pancuro-nium bromide or potassium chloride “going forward.” R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335). If the State were now allowed to revert to using pancuronium bromide or potassium chloride, it would create the perception that the district court, and perhaps this court, had been misled about the abandonment of pancuro-nium bromide and potassium chloride. Accordingly, the second factor is satisfied.
The third factor, “whether the party seeking to assert an inconsistent position would derive an unfair advantage or impose an unfair detriment on the opposing party if not estopped,” is also satisfied. New Hampshire v. Maine, 532 U.S. at 751, 121 S.Ct. 1808. Earlier in this litigation, by representing that there was “no possibility” that it would use pancuronium bromide or potassium chloride “going forward,” the State avoided having to litigate the constitutionality of an execution protocol that relied on those drags. R. 966-3 (Collins Aff: at ¶ 6) (Page ID #34335); R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329). Given the possibility that the State would revert to an execution protocol that relies on pancuronium bromide and potassium chloride—as State officials and attorneys represented that the State would not do, but as the State has now done— Plaintiffs were entitled to continue litigating the constitutionality of those drags.
In 2009, by making the unnecessarily broad and, we now know, -false representations that there was “no possibility” that the State would use pancuronium bromide and potassium chloride “going forward,” the State prevented Plaintiffs from having a trial on their claim that an execution protocol including those drags is unconstitutional. R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335); R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329). In 2016, the State concealed for months its intention to switch from a one-drug to a three-drug protocol, and the State has admitted that this was a strategic decision. Since revealing its current midazolam three-drag protocol, the State has pushed for this litigation to proceed quickly. Now, the State argues that Plaintiffs are not entitled to trial on their constitutional claims, including their claim that a two-drug protocol that does not include a paralytic is an available alternative that will significantly reduce a substantial risk of severe pain. The upshot of the State’s behavior—in making unequivocal representations that it would not use pancuronium bromide or potassium chloride in executions, seeking to moot Plaintiffs’ case based on those representations, acting in contravention of those representations, and now seeking to prevent Plaintiffs from having a trial—has been to thwart Plaintiffs’ efforts to litigate the constitutionality of Ohio’s use of a three-drug protocol or the question whether a two-drug protocol is an available alternative that significantly reduces a substantial risk of severe pain. *909This court’s acquiescence to the State’s behavior—by refusing to estop the State from using pancuronium bromide or potassium chloride—means that the State has succeeded in thwarting the Plaintiffs’ efforts try their constitutional claim before being executed.
The majority’s defense of the State’s behavior as based on changed circumstances is unconvincing. Even if Ohio changed its position because of changed circumstances, the change in circumstances had no bearing on Ohio’s decision to represent to two federal courts that the State would no longer use pancuronium bromide or potassium chloride. Nothing required Ohio to make the sweeping assertions that there was “no possibility” it would use potassium chloride “going forward,” rather than making a narrower representation to which it could adhere. R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335); R. 966-2 (Defs.’ Second Mot. Summ. J. with Req. Expedited Briefing Sched. at 5) (Page ID #34329).
Additionally, even if changed circumstances mean that Ohio can no longer obtain barbiturates, the unavailability of barbiturates does not require Ohio to revert to pancuronium bromide and potassium chloride, as opposed to using a different drug or combination of drugs. Ohio’s previous representations—including representations made under oath, see R. 966-3 (Collins Aff. at ¶ 6) (Page ID #34335)— prevent it from reverting to pancuronium bromide and potassium chloride. Given these representations, if barbiturates are not available to Ohio, Ohio should rely on an execution protocol not involving pancu-ronium bromide or potassium chloride.
Allowing the State to reverse course and use pancuronium bromide and potassium chloride in executions not only unfairly advantages the State, it also undermines the integrity of this litigation. The majority asserts that “if any gamesmanship led us to this pass, it was not gamesmanship by the State.” Maj. Op. at 892. Whether or not characterized as gamesmanship, there no question that the State has publicly taken inconsistent positions, concealed facts from Plaintiffs to gain strategic advantage, and attempted at every turn to deny Plaintiffs an opportunity to try their constitutional claims. The purpose of judicial estoppel is to ensure that litigants will not be rewarded for such behavior. The majority has ensured that the State will be rewarded. Therefore, reviewing the issue de novo, I come to the same conclusion as the district court. I would hold that the State of Ohio is judicially estopped from using pancuronium bromide or potassium chloride for executions.
[[Image here]]
Plaintiffs should not be executed before a trial on the constitutionality of Ohio’s execution method. The district court did not err by finding that Plaintiffs satisfied the requirements for a preliminary injunction or that the State of Ohio should be judicially estopped from using execution drugs it swore that it would no longer use. I respectfully dissent.
. "In several recent cases, this Court has 'questioned the continuing viability of the de novo standard for judicial estoppel, noting the Supreme Court’s characterization of the doctrine as an equitable remedy 'invoked by the court at its discretion’ and recognizing that the 'majority of federal courts’ review for abuse of discretion.’ ” Javery, 741 F.3d at 697 (quoting Kimberlin v. Dollar General Corp., 520 Fed.Appx. 312, 313 n.1 (6th Cir. 2013)). Because the en banc court is not bound by prior panel decisions stating that the appropriate standard of review is de novo, the en banc court should consider the continuing validity of the de novo standard in light of the considerations pointed out in Javery.