¶1 Leroy Jones was convicted of second degree assault for his role in a street fight involving five people. In a motion for a new trial made shortly after the verdict and before appeal, he asserted that his trial lawyer failed to interview and call certain eyewitnesses who were clearly identified in discovery that the State provided. Jones argued that these failures constituted ineffective assistance.
¶2 To prevail on a claim of ineffective assistance of counsel, Jones must establish both deficient performance and prejudice. Strickland v. Washington, 466 U.S. 668, 687, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984); State v. Hendrickson, 129 Wn.2d 61, 77-78, 917 P.2d 563 (1996). He *331has done so. It is clear that defense trial counsel’s failure to interview three previously identified and easily accessible eyewitnesses before trial constituted deficient performance. This deficiency also caused prejudice: it deprived Jones of the opportunity to develop a theory of the case that Jones was the victim rather than the aggressor, and it deprived him of neutral bystander eyewitness testimony in support of that theory. When considering the case as a whole, defense counsel failed to provide the meaningful adversarial role that the Sixth Amendment to the United States Constitution guarantees. Following Strickland, we must reverse.
¶3 Jones also argues that two prior Florida assault convictions should not have been used to increase his sentence because they are not “comparable” (per RCW 9.94A.525(3)) to second degree assault in Washington for “persistent offender” sentencing purposes. The difference is that Florida courts reject the defense of “diminished capacity” that Washington courts accept as a defense to this specific-intent crime. Because this issue might arise again if there is a retrial, we address it here also. Recently, in State v. Sublett, 176 Wn.2d 58, 88-89, 292 P.3d 715 (2012) (plurality opinion), we held that a difference in the availability of this particular defense — diminished capacity — is not relevant to whether the out-of-state conviction is “comparable” to an otherwise nearly identical Washington crime. Thus, the trial court did not err in its sentencing calculations.
FACTS AND PROCEDURAL HISTORY
1. Proceedings in the Trial Court
¶4 Jones and Taurian Alford had a fight on a public street in downtown Seattle on September 10, 2007. Three of Alford’s friends joined in the fray. There were several witnesses, including the other men in the fight. They generally identified Jones as the aggressor and testified that he held a knife. A jury convicted Jones of second degree assault.
*332f 5 Right after the jury returned that verdict, however, appointed defense counsel withdrew due to concerns about his own ineffectiveness. He realized that he had failed to interview witness Lori Brown, who was clearly identified in police reports. Brown was not called to his attention until a detective testified at trial about his interview with Brown.1 Shortly after withdrawal, new defense counsel discovered a second witness, Michael Hamilton, who was also clearly identified in pretrial discovery and whom defense trial counsel also failed to interview. It appears that the new lawyer found Hamilton while simply reviewing discovery that was already in defense trial counsel’s possession.
¶6 The new defense lawyer therefore moved for a new trial on the ground of ineffective assistance of counsel. He argued, and presented written documents showing, what Hamilton would have said. Clerk’s Papers (CP) at 92. He presented no evidence about why the original defense lawyer did not previously interview Brown or Hamilton. The trial court entered findings of fact based on the written materials, without an evidentiary hearing, and denied the motion.
¶7 Jones had two prior Florida convictions for crimes that the trial court deemed comparable to second degree assault. The court therefore sentenced Jones to life without parole under the Persistent Offender Accountability Act (POAA), RCW 9.94A.570.
2. Trial Testimony
¶8 Neither Jones nor Alford — the men who fought initially — testified at trial. According to the testimony of other witnesses, at some point one of the men chased the other and tackled him at a bus stop where several people were *333standing. Three of Alford’s friends eventually joined the fight; all of them were much younger than Jones. Jones held a knife sometime during the fight — the exact time is in dispute.2 The younger men held Jones down and punched him while restraining Jones’s hand holding the knife. Several people called 911, and when police arrived the younger men were restraining Jones, who still held the knife. The King County prosecutor charged Jones with second degree assault with a deadly weapon.
¶9 According to four bystanders who did not witness the fight, Alford approached them and said that he was being chased and that someone was trying to stab him. Somewhat inconsistently, all four bystanders testified that Jones was the aggressor and had a knife either while he was chasing Alford or while he and Alford were on the ground.
¶10 Alford’s cousin testified, similar to some of the bystanders, that Jones was the aggressor and was attacking Alford when he and two other friends came upon Jones and Alford. He also stated that Alford’s three friends kicked and punched Jones to protect Alford and to restrain Jones’s hand holding the knife.
¶11 On the other hand, defense witness Mark Forbes— another neutral bystander — testified that he stood about 15 feet away from Jones and Alford when one of them “swung [his sweater] at the other gentleman” before they started punching each other and wrestling on the ground. Verbatim Report of Proceedings (VRP) (Apr. 14, 2008) at 67. Forbes also testified that he saw three other men “running down and they started kicking the gentleman on the ground and punching him, and they were very violent about it, too.” Id. at 69. Critically, Forbes further testified that Jones drew the knife to protect himself after the other three men joined the fight. Id. at 70.
*334¶12 There was another witness listed in discovery: Brown. CP at 215. Defense counsel did not notice that, though, until a detective mentioned her name during testimony in the middle of trial. Defense counsel moved for a mistrial on the ground that the prosecutor did not provide the detective’s interview report; instead, the trial court gave him a three-day recess.
¶13 Brown then testified for the State. But, notably, she said that Alford chased Jones. VRP (Apr. 14, 2008) at 23. She also testified that she never saw a weapon, that she didn’t hear a reference to a knife until Alford’s three friends joined the fight, and that “I wasn’t clear who had a knife.” Id. at 18-20.
¶14 The jury convicted Jones of the assault charge.
3. Motion for a New Trial
¶15 As discussed above, Jones’s attorney withdrew after trial over concerns that he had been ineffective. CP at 87, 131. Jones’s replacement counsel reviewed the discovery and found the 911 dispatch report that named not just Brown, whom trial counsel already realized he had overlooked, but also a second witness whom Jones’s original attorney failed to interview: Hamilton.
¶16 New counsel then interviewed Hamilton. CP at 218-36. According to the transcript of the defense interview filed with the trial court in support of the motion for a new trial, Hamilton said that he was at the bus stop when the fight occurred and that he was standing very close to Alford and Jones. He was certain that the younger man (Alford) tackled the older man (Jones) and started beating him before the young man’s friends joined in. Hamilton said, “I did not see the actual extraction of the knife. I did see it in his hand after he had been tackled and after [Alford] started hitting him.” CP at 223. Hamilton also stated, “[W]hat I saw was guy number two [Alford] tackled guy number one [Jones], then the knife coming out, subdued the knife [in the hand of guy number one]. . . . Guy number *335three came up, clocked him, a beating ensued, and I called 911 and went away on the bus.” CP at 226. Hamilton believed that the police were bound to get the wrong idea when they arrived and would think that the young men were restraining an armed attacker. CP at 225. Hamilton believed that the older man (Jones) was acting in self-defense. Hamilton also said that he was with another man who would have testified to the same thing, but because no one contacted him sooner, he could no longer remember the other man’s name. According to police reports filed in support of the new trial motion, Hamilton’s name and phone number were recorded on a 911 dispatch report provided to the prosecutor and to the defense through discovery, but neither party contacted Hamilton before the trial.
¶17 Finally, Jones’s new lawyer filed a declaration stating that the original defense lawyer failed to interview Brown and Hamilton. CP at 131-35.
¶18 Jones argued that trial counsel’s failure to interview Brown and Hamilton and to call Hamilton to testify constituted ineffective assistance. The trial court concluded that the failure to interview Brown before trial was not prejudicial because Brown ultimately testified at trial. CP at 888. The trial court also concluded that the failure to interview Hamilton was not prejudicial because “Hamilton’s proposed testimony is not exculpatory because it contradicts the defense position at trial. At trial defendant testified he drew the knife in self-defense after he was assaulted by Alford and his two friends.” CP at 889. The trial court clearly erred on this point: Jones did not testify at trial.
4. Court of Appeals Decision
¶19 The Court of Appeals affirmed and used the same reasoning as the trial court. State v. Jones, noted at 157 Wn. App. 1052, 2010 WL 3490255, 2010 Wash. App. LEXIS 2017. It did not address whether the failure to interview these *336identified witnesses constituted deficient performance. 2010 WL 3490255, at *3, 2010 Wash. App. LEXIS 2017, at *10. Instead, the Court of Appeals found that the failure to contact Hamilton did not cause prejudice, noting that his testimony “would not likely have changed the outcome of the trial because it contradicted four other eyewitnesses.” 2010 WL 3490255, at *3-4, 2010 Wash. App. LEXIS 2017, at *10. Notably, neither the trial court nor the Court of Appeals made an adverse credibility finding about Hamilton.
¶20 Contradictorily, the Court of Appeals ruled that the failure to contact Brown was not prejudicial — she actually testified and her testimony did not affect the outcome because it “was similar to that of the other eyewitnesses.” 2010 WL 3490255, at *4, 2010 Wash. App. LEXIS 2017, at *10. Thus, in the appellate court’s view, new evidence will not affect the outcome if it is cumulative and will not affect the outcome if it is different.
5. Remand for a RAP 9.11 Hearing
¶21 Jones sought review in this court of the denial of his claim of ineffective assistance of counsel regarding the two witnesses, the affirmance of the trial court’s ruling that his prior Florida felonies are comparable to Washington assaults, and the rejection of his claim that the prior convictions should have been proved to the jury beyond a reasonable doubt rather than to the judge. We granted review on the first two issues. State v. Jones, 177 Wn.2d 1007, 300 P.3d 416 (2013).
¶22 On April 10, 2014, we ordered a Rules of Appellate Procedure (RAP) 9.11 hearing and directed the trial court “to take additional evidence and to make factual findings based on that evidence, to enable this court to determine whether defense counsel provided ineffective assistance ... including but not limited to: (1) whether defense counsel’s performance was deficient for failure to interview witnesses; (2) why defense counsel did not interview all the *337witnesses listed in the discovery; and (3) why defense counsel did not call one of the witnesses listed in the discovery, Michael Hamilton, to testify.”
¶23 At the remand hearing, Jones’s original defense counsel testified about his failure to interview the two witnesses already discussed previously — Brown and Hamilton — and his failure to call Hamilton as a witness. VRP (Aug. 21, 2014) at 6-66.
¶24 Defense counsel at the remand hearing then identified yet another witness listed in discovery whom trial counsel had failed to interview: Sulva Ooveda. An incident report provided to Jones during discovery listed Ooveda’s name. CP at 216. Notably, the prosecutor interviewed her at the beginning of trial and actually informed defense counsel that she might have favorable evidence. Remand CP at 33-34; VRP (Aug. 21, 2014) at 27-28. Despite this notification from the prosecutor, defense counsel still failed to contact Ooveda. During the remand hearing, Jones’s original defense attorney noted that he had asked his investigator to interview Ooveda before trial but that she had failed to do so and he had not followed up. VRP (Aug. 21, 2014) at 26-28. Defense attorney expert Richard Hansen testified that trial counsel’s performance was deficient and that it likely affected the outcome of the trial. Id. at 70-108.
¶25 The trial court also admitted other evidence, without objection, confirming that trial counsel failed to interview Ooveda, Brown, and Hamilton.3 Def.’s Ex. 2, at 5-7; Def’s Ex. 4, at 2; Remand CP at 34-37; VRP (Aug, 21, 2014) at 26-28, 33-36, 53, 111.
¶26 With regard to witnesses Brown and Ooveda, the trial court found, “The communication from [Deputy Prosecuting Attorney] Richey to [defense counsel] piques curiosity and raises the inference that [Ooveda’s] testimony *338may have been helpful, and that an interview, albeit at the start of trial [,] would occur.” Remand CP at 34. And the trial court found that defense counsel “testified that he was at a disadvantage not having [Brown’s] witness statement prior to trial,” even though the incident report listed her name. Id. Defense counsel offered no reasons for failing to interview these witnesses. The trial court accordingly concluded that “[t]his failure to interview Brown and 0[o]veda, witnesses listed on the incident report, clearly is not objectively reasonable” and that counsel’s performance was therefore deficient. Id. at 35. The trial court concluded, however, that this did not cause prejudice, “given the testimony of the other State’s witnesses who testified that the Defendant Jones first introduced the knife.” Id.
¶27 With regard to witness Hamilton, the trial court found that transcripts of Hamilton’s posttrial defense interview and his 911 call were “unclear” about when Jones wielded the knife and that “[b]oth transcripts show Hamilton mixed up the parties, having the Defendant chased by the younger man, rather than as the majority of witnesses testified.” Id. at 36. The trial court concluded that defense counsel’s failure to call Hamilton to testify “is not objectively unreasonable. This decision appears strategic in nature and hence not deficient performance.” Id. at 37.4
ANALYSIS
Ineffective Assistance
I. Standard of Review
¶28 A claim that counsel was ineffective is a mixed question of law and fact that we review de novo. Strickland, *339466 U.S. at 698; In re Pers. Restraint of Brett, 142 Wn.2d 868, 873, 16 P.3d 601 (2001). “A defendant is denied effective assistance of counsel if the complained-of attorney conduct (1) falls below a minimum objective standard of reasonable attorney conduct, and (2) there is a probability .that the outcome would be different but for the attorney’s conduct.” State v. Benn, 120 Wn.2d 631, 663, 845 P.2d 289 (1993) (emphasis omitted) (citing Strickland, 466 U.S. at 687-88). Thus, to prevail on a claim of ineffective assistance of trial counsel, an appellant must show both deficient performance and prejudice. Strickland, 466 U.S. at 687; Hendrickson, 129 Wn.2d at 77-78. To show prejudice, the appellant need not prove that the outcome would have been different but must show only a “reasonable probability” — by less than a more likely than not standard — that, but for counsel’s unprofessional errors, the result of the proceedings would have been different. Strickland, 466 U.S. at 694; Hendrickson, 129 Wn.2d at 78.
II. Deficient Performance
¶29 As discussed above, following the remand hearing, the trial court ruled that trial counsel’s failure to interview the witnesses identified in the police reports “demonstrated a deficient performance using the Strickland standard.” Remand CP at 39; see also id. at 34-35.
¶30 The facts certainly supported the trial judge’s conclusion on this point. A criminal defendant has a state and federal constitutional right to effective assistance of counsel. Strickland, 466 U.S. at 686; State v. Tinkham, 74 Wn. App. 102, 109, 871 P.2d 1127 (1994). To discharge this duty, trial counsel must investigate the case, and investigation includes witness interviews. State v. Ray, 116 Wn.2d 531, 548, 806 P.2d 1220 (1991) (“Failure to investigate or interview witnesses, or to properly inform the court of the substance of their testimony, is a recognized basis upon which a claim of ineffective assistance of counsel may rest.” (citing State v. Visitacion, 55 Wn. App. 166, 173-74, 776 P.2d 986 (1989))).
*340¶31 Thus, failure to interview a particular witness can certainly constitute deficient performance. Id. (“Failure to investigate or interview witnesses ... is a recognized basis upon which a claim of ineffective assistance of counsel may rest.”); Jones v. Wood, 114 F.3d 1002 (9th Cir. 1997) (failure to investigate witnesses called to attention of trial counsel as important constitutes ineffectiveness). It depends on the reason for the trial lawyer’s failure to interview.
¶32 In this case, trial counsel offered absolutely no reason for failing to interview these three witnesses. Remand CP at 35. With regard to Hamilton in particular, the trial court ruled that the defense lawyer “does not recall” why he failed to interview Hamilton and “does not provide any reason either because it is clear from the incident report there was a 9-1-1 call from him.” Id. at 36. The trial court then concluded that the failure to interview all witnesses so identified was “deficient performance.” Id. at 39.
¶33 We agree. We can certainly defer to a trial lawyer’s decision against calling witnesses if that lawyer investigated the case and made an informed and reasonable decision against conducting a particular interview or calling a particular witness. See, e.g., State v. Hess, 12 Wn. App. 787, 788-90, 532 P.2d 1173 (1975) (decision not to subpoena potentially harmful witness was justified); State v. Floyd, 11 Wn. App. 1, 2, 521 P.2d 1187 (1974) (decision not to call alibi witness legitimate part of trial strategy). But courts will not defer to trial counsel’s uninformed or unreasonable failure to interview a witness. See Ray, 116 Wn.2d at 548. As the United States Supreme Court has explained, “[Strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.” Strickland, 466 U.S. at 690-91.
¶34 On the other hand, we disagree with the trial court’s conclusion on remand that the failure to call Hamilton to testify was “not objectively unreasonable.” Remand CP at *34137. Specifically, we disagree with the trial court’s conclusion that the failure to call Hamilton to testify, after failing to interview him, was “strategic in nature.” Id. This is because defense trial counsel testified that when he prepared for trial and failed to interview Hamilton, he “did not have any idea what Mr. Hamilton would have said about this case.” VRP (Aug. 21, 2014) at 41. That is not strategic decision-making.
¶35 The trial court, however, relied on trial counsel’s posttrial remand hearing testimony that when he eventually read the transcript of Hamilton’s 911 call, after trial, it made him think that Hamilton probably would not have offered any helpful testimony. But trial counsel made this conclusion after trial, in hindsight. VRP (Aug. 21, 2014) at 42, 45, 48-49, 50-52. Strategic decisions are those made before, not after, taking the challenged action. Avila v. Galaza, 297 F.3d 911, 920 (9th Cir. 2002) (“ ‘[C]ounsel can hardly be said to have made a strategic choice when s/he has not yet obtained the facts on which a decision could be made.’ ” (alteration in original) (quoting Sanders v. Ratelle, 21 F.3d 1446, 1457 (9th Cir. 1994))); see Strickland, 466 U.S. at 690-91 (“[Strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.”). Trial counsel did not make an informed decision against interviewing Hamilton, and he therefore could not have made a strategic — i.e., an informed — decision against calling Hamilton as a witness.
III. Prejudice
¶36 We therefore come to the question of prejudice. We disagree with the trial court’s conclusion on this point. In our view, there is certainly a reasonable probability that the failure to interview or call witnesses affected the trial’s outcome.
¶37 We start with defense counsel’s failure to interview Hamilton. On the one hand, Hamilton would have testified *342that the young man — Alford—chased and tackled Jones, not the other way around. This testimony would have corroborated Brown’s testimony to the same effect. VRP (Apr. 14, 2008) at 23. This is important. See Howard v. Clark, 608 F.3d 563, 573 (9th Cir. 2010) (“Although Hernandez positively identified Howard as the shooter, if Ragland had testified otherwise, thereby buttressing Fontaine’s trial testimony, some jurors might well have had a reasonable doubt as to Howard’s guilt.”). And Hamilton would have provided the very defense-favorable testimony that Jones was the victim and that Hamilton called 911 to report what he saw because he knew that another bystander might think that because Jones had the knife, he was the aggressor. CP at 225,233; Remand CP at 36. Further, although the trial court thought that Hamilton was confused because his recollection differed from the testimony of other witnesses, there was no finding that Hamilton was lying or unbelievable. The difference is critical. State v. West, 139 Wn.2d 37, 43-44, 983 P.2d 617 (1999); see Howard, 608 F.3d at 573 (“Whatever the challenges to Ragland’s credibility, his testimony might well have tipped the balance in Howard’s favor. At the very minimum, if Ragland was ready and willing to testify as to Howard’s innocence, and Howard was deprived of such testimony because of his attorney’s shoddy investigation, our confidence in the jury’s verdict would be significantly undermined.”). In fact, unlike many of the State’s witnesses, Hamilton was a neutral observer with no relationship to Jones or Alford.
¶38 On the other hand, Hamilton’s testimony contradicted Jones’s chosen theory of self-defense by placing a knife in Jones’s hand during his fight with the younger man, and before rather than after the others joined the fight, which was the defense’s theory of the case. CP at 225-26, 888-89; Remand CP at 36. But defense counsel adopted and used that theory without knowing that Hamilton’s testimony existed. VRP (Aug. 21, 2014) at 62. If trial counsel had known before trial about Brown’s statement that she heard *343a reference to a knife and saw jabbing motions after Alford’s friends joined the fight, and had had the chance to consider it along with Hamilton’s testimony that the knife appeared before the friends joined the fight, he might not have been boxed into that theory. See Johnson v. Baldwin, 114 F.3d 835, 839-40 (9th Cir. 1997) (deficient performance caused prejudice when trial counsel failed to interview petitioner’s girlfriend or grandmother because counsel would have learned that petitioner’s alibi was false and pursued a different trial strategy); see also Rios v. Rocha, 299 F.3d 796, 808, 812 (9th Cir. 2002) (defense counsel’s “failure, in a first-degree murder trial, to interview more than one witness, when there were dozens of potential eyewitnesses available, before deciding to abandon a potentially meritorious defense constituted constitutionally deficient performance”; defense counsel’s decision to present “unconsciousness” defense as opposed to a misidentification defense was prejudicial because counsel failed to interview and call five eyewitnesses to testify who would have each stated that Rios was not the shooter).
¶39 One final consideration in the prejudice inquiry regarding Hamilton is whether calling him to testify might have resulted in any other adverse consequences to Jones. The State argues that if Hamilton had testified, then it could have introduced Jones’s pretrial statement that it now calls inculpatory.5 But before trial, the defense characterized this statement as favorable to the defense and the State moved — successfully—to exclude it. Given the State’s pretrial efforts to exclude this statement, we are skeptical about its new, postremand position that the statement was more helpful than hurtful to the State.
¶40 Then there is witness Brown. Although the jury had an opportunity to consider Brown’s testimony, Jones’s trial *344counsel explained that if he had known about her testimony before trial, he would have made it the centerpiece of his case and the focal point of cross-examination of other witnesses. VRP (Apr. 10, 2008) at 15.
¶41 Finally, we consider witness Ooveda. The prosecutor specifically told trial counsel on the first day of trial, after interviewing Ooveda, that she may have exculpatory information. VRP (Aug. 21, 2014) at 27-28. Defense counsel still failed to find out what information she might have provided.
¶42 We cannot avoid the conclusion that there is a reasonable probability that the failure to interview and to call Hamilton affected the outcome of the trial. This case involves a credibility contest between the State’s witnesses and Jones’s witness. Although the State’s witnesses would still have outnumbered Jones’s witnesses, the jury would have had the opportunity to weigh the credibility of two witnesses — rather than just one — claiming that Alford chased after Jones against five witnesses who testified for the State that Jones was the aggressor. There is a reasonable probability that this would have affected the outcome. See Avila, 297 F.3d at 918-23 (counsel’s failure to interview eight additional eyewitnesses who would have testified in an attempted murder trial that the defendant was not the shooter was prejudicial even though counsel presented three eyewitnesses who corroborated the defendant’s testimony that he was not the shooter). Further, Hamilton’s testimony tends to bolster Forbes’s credibility and, concomitantly, diminish the credibility of the State’s witnesses who testified to the contrary. There is a reasonable probability that this would have affected the outcome. See Nealy v. Cabana, 764 F.2d 1173, 1179 (5th Cir. 1985). And although Hamilton’s account about the time that the knife appeared seems to conflict with Forbes’s and Brown’s accounts, Hamilton’s testimony would have corroborated Forbes’s testimony that Jones acted in self-defense. There is a reasonable probability that this would have affected the *345outcome. See Howard, 608 F.3d at 573 (even though State’s witness identified petitioner as the shooter, testimony from surviving victim that he could not identify petitioner as the shooter would have buttressed another witness’s trial testimony, possibly creating reasonable doubt about petitioner’s guilt). The failure to interview Brown and Ooveda compounds the prejudice.
¶43 Thus, counsel’s unexplained failure to interview clearly identified and accessible witnesses undermines our confidence in the jury verdict rejecting Jones’s self-defense claim. We therefore reverse the appellate court’s decision that Jones failed to prove ineffective assistance of counsel.
IV. The Availability of a Diminished Capacity Defense in Washington, but Not in Florida, Does Not Affect Our Comparability Analysis
¶44 To determine whether a prior out-of-state conviction counts as a strike under Washington’s POAA, the court must determine if there is a Washington offense to which the out-of-state conviction is “comparable.” RCW 9.94A.525(3); State v. Ford, 137 Wn.2d 472, 479-80, 973 P.2d 452 (1999). The State bears the burden of establishing the comparability of the out-of-state convictions. Ford, 137 Wn.2d at 479-80. The court compares the elements of the foreign crime with the elements of the purportedly comparable Washington crimes. Id.
¶45 If the elements differ, the sentencing court can, in some cases, look at portions of the record of the prior proceeding to see if the conduct of which the defendant was convicted was identical to what is required for a comparable Washington conviction; but the portion of the foreign record that the Washington court can consider is very limited. Id. The sentencing court can look at the charging instrument from the foreign proceeding, but it cannot consider “facts and allegations contained in [the] record of prior proceedings, if not directly related to the elements.” Id. at 480 (citing State v. Morley, 134 Wn.2d 588, 606, 952 P.2d 167 *346(1998)). This limitation is compelled by not just statutory interpretation but also constitutional concerns. See Shepard v. United States, 544 U.S. 13, 26, 125 S. Ct. 1254, 161 L. Ed. 2d 205 (2005). As this court explained in In re Personal Restraint of Lavery, 154 Wn.2d 249, 258, 111 P.3d 837 (2005):
Any attempt to examine the underlying facts of a foreign conviction, facts that were neither admitted or stipulated to, nor proved to the finder of fact beyond a reasonable doubt in the foreign conviction, proves problematic. Where the statutory elements of a foreign conviction are broader than those under a similar Washington statute, the foreign conviction cannot truly be said to be comparable.[6]
The defenses, however, differ. Assault is a specific intent crime. Diminished capacity is a defense to a specific intent crime in Washington. Id. at 255-56. Diminished capacity is not a defense in Florida. See, e.g., Evans v. State, 946 So. 2d 1, 11 (Fla. 2006); Chestnut v. State, 538 So. 2d 820, 820 (Fla. 1989).
¶46 In Sublett, however, this court held that the availability of the defense of diminished capacity in Washington, but not in the foreign jurisdiction, does not prevent two crimes from being “comparable.” 176 Wn.2d at 88-89. Sublett did not discuss the role of other defenses in making this determination — but as to the defense of diminished capacity, the one at issue here, it stands as controlling precedent.
*347CONCLUSION
¶47 Defense trial counsel’s failure to investigate and to interview easily identified, available eyewitnesses, without a legitimate tactical reason, constitutes deficient performance and caused prejudice in this case. With regard to sentencing, Sublett controls. It held that if the elements of a Washington crime and a foreign prior conviction are the same, then the crimes are comparable, even if the defense of diminished capacity is unavailable in the foreign jurisdiction. Jones’s prior Florida assault convictions are comparable to second degree assault convictions in Washington. We therefore reverse Jones’s conviction and remand for a new trial.
Madsen, C.J., and C. Johnson, Wiggins, and González, JJ., concur.This detective’s interview notes were never transmitted to the defense (or the prosecutor). Verbatim Report of Proceedings (VRP) (Apr. 9, 2008) at 58-63.
The defense’s theory was that Jones held the knife only after Alford’s friends joined the fight. VRP (Aug. 21, 2014) at 48, 72. The State’s witnesses were inconsistent about when they saw the knife in Jones’s hand. Id. at 47-48, 62.
The State attached to its prehearing memorandum a transcript of Hamilton’s 911 call, in which Hamilton stated, “The knife is in the hand of the man being held down.” Remand CP at 14.
Additionally, on a separate issue, the trial court found that defense counsel, “in private discussions with his client he told Jones that a conviction on the pending charges would be [a] ‘third strike,’ but also later in open court he had agreed with the State that it would not be treated as such.” Remand CP at 39. The court found nothing in the record showing that these different statements confused Jones about the fact that it was a three-strikes case. Id.
Jones’s statement says, “They sold me some bullshit dope and I went fighting for my money. They jumped me when I was fighting with the young one. I bought $10.00 rock of bullshit. I was trying to stab him because three of these guys jumped me. I was defending myself.” State’s Ex. 8.
In this case, the elements of the prior Florida assault convictions are practically identical to the elements of second degree assault in Washington. Under Fla. Stat. § 784.021(l)(b), “aggravated assault” was defined as “an assault [w]ith an intent to commit a felony.” Under Fla. Stat. § 784.045(l)(a)(2), “[a] person commits aggravated battery who, in committing battery [u]ses a deadly weapon.” In Washington, under RCW 9A.36.021(1), “[a] person is guilty of assault in the second degree if he or she, under circumstances not amounting to assault in the first degree:... (c) Assaults another with a deadly weapon; or ... (e) With intent to commit a felony, assaults another.”