Full Opinion

In the United States Court of Appeals For the Seventh Circuit ____________________ No. 24-1231 CHRISTOPHER RODGERS, Petitioner-Appellant, v. RON NEAL, Respondent-Appellee. ____________________ Appeal from the United States District Court for the Southern District of Indiana, Indianapolis Division. No. 22-cv-0198 — James R. Sweeney II, Chief Judge. ____________________ ARGUED SEPTEMBER 29, 2026 — DECIDED NOVEMBER 20, 2026 ____________________ Before BRENNAN, Chief Judge, and ST. EVE and TAIBLESON, Circuit Judges. ST. EVE, Circuit Judge. Indiana law permits discretionary al- ternative sentencing for juveniles tried in criminal court. Christopher Rodgers pled guilty to felony murder at age sev- enteen and was thus eligible to be sentenced pursuant to the scheme. His trial attorney, however, never informed him of the possibility, and Rodgers agreed to a fixed forty-five-year sentence. 2 No. 24-1231 Rodgers later filed for postconviction relief, arguing that his attorney’s failure to advise him of alternative sentencing denied him the effective assistance of counsel. State courts and a federal district court disagreed, however, and Rodgers appealed. Because he provides no contemporaneous evidence that any of counsel’s errors prejudiced him, we affirm. I. Background On June 27, 2013, the State of Indiana charged then seven- teen-year-old Rodgers with felony murder, attempted rob- bery, and conspiracy to commit robbery. Under Indiana Code § 31-30-1-4, juvenile courts lacked jurisdiction over him and his case because of the nature of the offense. Rodgers was therefore a “direct file” into adult criminal court, where his murder charge alone carried a sentence of between forty-five and sixty-five years, with an advisory sentence of fifty-five years. Ind. Code § 35-50-2-3. 1 His conspiracy to commit rob- bery charge, meanwhile, carried an additional six to twenty years—for a total of eighty-five years. Id. § 35-50-2-5. He agreed to plead guilty to felony murder in March 2014, ac- cepting a fixed forty-five-year sentence—the minimum, per the murder statute—and, in exchange, the State dropped the additional charges and forewent anything higher. Counsel represented Rodgers during plea negotiations, and she ap- peared at both his plea and sentencing hearings on March 24 and May 23, 2014, respectively. At the plea hearing, Rodgers’s attorney and the State’s at- torney informed the court that—notwithstanding the agreed- 1 Section 35-50-2-3(b)(2) permits a sentence of life imprisonment with- out parole for seventeen-year-olds, but the Indiana trial court never ap- peared to consider that possibility for Rodgers. No. 24-1231 3 upon forty-five years—new, late-breaking information might yield more favorable terms than those tendered to the court. But if that information did not pan out, they would proceed with the fixed forty-five-year sentence. The court then recom- mended that they move forward with the guilty plea and, bar- ring any changes, proceed with sentencing pursuant to the plea agreement. Both attorneys agreed. Rodgers did not raise any concerns or otherwise demur during the court’s plea colloquy, including throughout the following exchange: THE COURT: While we’ve had discussions here that indi- cate[] that the State might give you a better offer if certain information turns out to be true and valuable, there is also the possibility that the State will think either it’s not relia- ble information or it’s not valuable enough to get off this plea agreement. Do you understand that? MR. RODGERS: Yes. THE COURT: So as you sit here right now you can plan on 45, do you understand? MR. RODGERS: Yes, sir. THE COURT: Things might get better, but there is no guarantee. And even though you might think that the in- formation is very very valuable, [the] State might disagree. Is that clear to you? MR. RODGERS: Yes, sir. THE COURT: And even if it is valuable the State might not be able to prove it, it might not be reliable. Do you under- stand that? MR. RODGERS: Yes, sir. 4 No. 24-1231 THE COURT: So as we sit here right now, you’re planning on getting 45, right? MR. RODGERS: Yes, sir. THE COURT: You hope it’s better, but you don’t know for sure. And you understand you’re looking at 45 right now? MR. RODGERS: Yes, sir. Rodgers also confirmed that he discussed the plea with his attorney; that he was, in fact, culpably involved in the felony murder; that he understood the agreement and sentence; and that he understood the rights he was forfeiting, including his right to a trial. The court found Rodgers’s plea knowingly, in- telligently, and voluntarily made, and that a factual basis for it existed. It scheduled his sentencing for April, assuming— correctly so, it turned out—that nothing changed. Rodgers returned to court two months later for sentenc- ing. There, for the first time, he sought to formally withdraw his plea. He explained that he pled guilty because he knew nothing about the law and believed that doing so was best for him, even though he was “not guilty of what [he] said [he] was.” According to Rodgers, he was told by someone in prison—not his attorney—that he was guilty simply by “be- ing there,” which he believed to be inaccurate, though he stopped short of saying he lied under oath. Acknowledging that the request to set aside the plea was untimely, the trial court then asked the attorneys for input. The State’s lawyer pressed that, because of the strong factual basis for felony murder, the plea was in his best interest—and if Rodgers wanted something better, they were going to trial. Rodgers’s counsel agreed: in her evaluation of the evidence, the agreement was “absolutely” in his best interest. So, the No. 24-1231 5 court denied Rodgers’s untimely request, emphasizing that he would receive the minimum sentence for the lead offense. At no point did any attorney—or the court—tell Rodgers about Indiana Code § 31-30-4-2, which went into effect on July 1, 2013. The law established an alternative sentencing scheme for individuals like Rodgers, i.e., juveniles “charged with a felony over which a juvenile court does not have jurisdiction under IC 31-30-1-4.” Ind. Code § 31-30-4-2(a)(2). As the Indi- ana Supreme Court explained, the law permits juvenile of- fenders to “receive a suspended sentence or be placed in a ju- venile facility. … [B]ut this decision is left to the trial court’s discretion.” Harris v. State, 165 N.E.3d 91, 99 (Ind. 2021); see also Ind. Code § 31-30-4-2(b)(2)–(3). Rodgers, by agreeing to a fixed forty-five-year sentence, relinquished the possibility of receiving anything less. By August 2016, Rodgers learned of his eligibility for alter- native sentencing and filed for postconviction relief in state court. Proceeding on a pro se basis, he argued that he was “misled by prosecution and attorney which resulted in the de- fendant entering an involuntary plea of guilt.” When ques- tioned during the postconviction evidentiary hearing, his trial counsel maintained that Rodgers was never eligible for alter- native sentencing, and that the “bare minimum sentence” for felony murder was forty-five years. Rodgers therefore argued, among other things, that he did not receive the effective assis- tance of counsel as guaranteed by Article 1 of the Indiana Constitution and the Sixth Amendment of the United States Constitution. The postconviction court rejected the argument, conclud- ing that Rodgers’s petition provided no facts or details sup- porting his claim. Rodgers appealed, and the Indiana Court 6 No. 24-1231 of Appeals affirmed on different grounds: that § 31-30-4-2 did not become effective until July 1, 2014, and so counsel “could not have acted deficiently by failing to advise Rodgers of a sentencing scheme which did not yet exist.” This, of course, was incorrect; the law went into effect on July 1, 2013—well before Rodgers’s sentencing. Even so, the Indiana Supreme Court declined to hear the case. Rodgers filed for federal habeas corpus relief in the South- ern District of Indiana under 28 U.S.C. § 2254. The State con- ceded the Indiana appellate court’s error, but the district court nonetheless rejected the petition, concluding that nothing in the guilty plea agreement prevented Rodgers from later seek- ing alternative sentencing. But this, too, was erroneous; the fixed forty-five years stripped the trial court of its discretion to suspend Rodgers’s sentence. The district court acknowledged as much on a motion for reconsideration yet still denied relief. It explained that it was unlikely the trial court would have or- dered alternative sentencing given Rodgers’s circumstances. Rodgers now appeals. 2 II. Discussion We review a district court’s decision on a petition for writ of habeas corpus de novo, “but our inquiry is an otherwise narrow one.” Schmidt v. Foster, 911 F.3d 469, 476 (7th Cir. 2018) (en banc). Generally, we begin by looking to the “last rea- soned state-court decision on the merits,” and our review is 2 A judge granted a certificate of appealability on July 18, 2025. In do- ing so, she invited the parties to address counsel’s failure to advise Rodg- ers of both § 31-30-4-2 and then-pending amendments to Indiana’s sen- tencing modification statute, § 35-38-1-17(i). Rodgers declined to pursue any argument involving the latter. No. 24-1231 7 both governed and limited by the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”). See id. at 476–77. When, however, the state court based its decision on an “unreasona- ble factual determination,” we analyze the substantive merits “under the pre-AEDPA standard—that is, de novo—because there is no state court analysis to apply AEDPA standards to.” Carlson v. Jess, 526 F.3d 1018, 1024 (7th Cir. 2008); see also Thomas v. Clements, 789 F.3d 760, 768 (7th Cir. 2015); Conner v. McBride, 375 F.3d 643, 655 n.5 (7th Cir. 2004). Here, all agree that the Indiana appellate court’s opinion, as the last reasoned state court decision on the merits, relied on an unreasonable factual determination: that the scheme was not yet in effect at Rodgers’s sentencing. We thus approach our analysis de novo. Rodgers’s appeal turns on allegations of ineffective assis- tance of counsel, a well-established violation of the Sixth Amendment and basis for habeas relief. See Strickland v. Wash- ington, 466 U.S. 668, 686 (1984); Lee v. United States, 582 U.S. 357, 363 (2017). To establish that counsel was constitutionally ineffective, Rodgers must show both that counsel’s perfor- mance was deficient and that it prejudiced him as a result. Strickland, 466 U.S. at 687. These Strickland prongs “are at best difficult showings to make.” Lickers v. United States, 98 F.4th 847, 857 (7th Cir. 2024) (quoting Ebert v. Gaetz, 610 F.3d 404, 411 (7th Cir. 2010)). The Supreme Court has counseled that we “need not determine whether counsel’s performance was de- ficient before examining … prejudice,” and “[i]f it is easier to dispose of an ineffectiveness claim on the ground of lack of sufficient prejudice, … that course should be followed.” Strickland, 466 U.S. at 697. We do so here, addressing three sce- narios relevant to the prejudice inquiry: (1) that, notwith- standing counsel’s performance, the trial court was unlikely 8 No. 24-1231 to order alternative sentencing; (2) that, but for counsel’s per- formance, Rodgers would have negotiated a better plea deal; and (3) that, but for counsel’s performance, Rodgers would have instead gone to trial. A. Likelihood of Alternative Sentencing After initially misunderstanding the preclusive effect of a fixed plea, the district court reaffirmed its denial of habeas re- lief because there was “no reasonable probability that the trial court would have utilized the option of alternative juvenile sentencing for Petitioner even if that option had been open to it.” Rodgers says this misunderstands the prejudice standard. We agree. We have explained that “[t]o establish prejudice in the plea-bargaining context, [a petitioner] must show that ‘the outcome of the plea process would have been different with competent advice.’” Anderson v. United States, 94 F.4th 564, 581–82 (7th Cir. 2024) (quoting Lafler v. Cooper, 566 U.S. 156, 163 (2012)); Strickland, 466 U.S. at 694 (“defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different”). “One way for [Rodgers] to show this is to demonstrate that ‘but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.’” Anderson, 94 F.4th at 582 (quoting Hill v. Lockhart, 474 U.S. 52, 59 (1985)). This is a meaningfully different question than “whether, had he gone to trial, the result of that trial would have been different than the result of the plea bargain.” Lee, 582 U.S. at 364 (citation modified). The question is simply whether, “properly advised, [he] would have opted to go to trial.” Id. at 365. “That is because, while we ordinarily ‘apply a strong presumption of reliability to judicial proceedings,’ No. 24-1231 9 ‘we cannot accord’ any such presumption ‘to judicial proceed- ings that never took place.’” Id. at 364 (quoting Roe v. Flores- Ortega, 528 U.S. 470, 482–83 (2000)). The district court asked and answered the wrong ques- tion. As the Supreme Court explained in Lee, the correct in- quiry “focuses on a defendant’s decisionmaking, which may not turn solely on the likelihood of conviction after trial,” id. at 367—or, in our case, the likelihood of alternative sentenc- ing. If Rodgers would have chosen trial, the prejudice analysis ends there. He argued as much before the district court, and the court erred by insisting he do more. B. Better Plea Deal Turning, then, to the first of Rodgers’s two theories. Rodg- ers argues that he can show prejudice by (hypothetically) hav- ing negotiated a more favorable plea, regardless of whether he would have chosen trial over the deal. More specifically, he presses the difference between an open plea, which would have preserved his eligibility for alternative sentencing, and a fixed plea. He claims that, properly advised, he could have negotiated the former. Our court has not yet endorsed “a prejudice theory posit- ing that absent plea counsel’s ineffectiveness, the defendant would have negotiated a different, more favorable plea agree- ment.” Anderson, 94 F.4th at 573 n.7. We explained in Anderson that the Supreme Court’s decision in� Missouri v. Frye, 566 U.S. 134, 148 (2012), “comes the closest to establishing a prejudice standard” applicable in these circumstances. Anderson, 94 F.4th at 573 n.7. But Frye involved a lapsed plea deal that counsel never communicated to the defendant, which we re- garded as different. Id. We also observed that a previous 10 No. 24-1231 decision of ours, Brock-Miller v. United States, 887 F.3d 298, 312 (7th Cir. 2018), “may have left the door open” for the more expansive twist on the theory, but we had “yet to fully em- brace and define [its] boundaries.” Anderson, 94 F.4th at 573 n.7. We then declined to do so, instead “confin[ing] ourselves to”—and rejecting the merits of—“Anderson’s first and pri- mary theory of prejudice, which is that he would have taken his chances at trial.” Id. We need not definitively embrace nor define the bounda- ries of the theory today. Viable or not, Rodgers cannot rely on it here. The Supreme Court has explained in no uncertain terms that “[c]ourts should not upset a plea solely because of post hoc assertions from a defendant about how he would have pleaded but for his attorney’s deficiencies.” Lee, 582 U.S. at 369. Rather, “[j]udges should instead look to contemporane- ous evidence to substantiate a defendant’s expressed prefer- ences.” Id; see also Nissenbaum v. Jennings, 148 F.4th 548, 556 (7th Cir. 2025) (applying Lee standard in situation involving lapsed plea). Moreover, when a petitioner insists he would have hypothetically accepted an offer, we have explained that he “must ‘show—at a minimum—that the prosecutor would have actually offered him a deal had his attorney been com- petent.’” Nissenbaum, 148 F.4th at 555 (quoting Delatorre v. United States, 847 F.3d 837, 846 (7th Cir. 2017)); see also Frye, 566 U.S. at 148–49 (holding that defendant must show “a rea- sonable probability neither the prosecution nor the trial court would have prevented the offer from being accepted or im- plemented”); Lafler, 566 U.S. at 164 (“defendant must show that but for the ineffective advice of counsel there is a reason- able probability that the plea offer would have been presented No. 24-1231 11 to the court”). It follows that this latter showing also requires contemporaneous evidence. See, e.g., Delatorre, 847 F.3d at 847 (“Nothing in the record indicates that the prosecutor would have provided Delatorre, an uncooperative defendant, with a plea deal had Morelli performed any differently.”) Even assuming that Frye’s more expansive approach to prejudice also encompasses Rodgers’s better-plea-deal the- ory, Rodgers has not made the necessary showings. He says that knowledge of the alternative sentencing scheme would have increased his bargaining leverage and thus led to a better deal. Perhaps, but this is pure speculation. Rodgers “ha[d] no right to be offered a plea,” Frye, 566 U.S. at 148, so there is no way of knowing whether, or at what point, the State would have walked away from the negotiating table. Rodgers also asks us to weigh the State’s argument on ap- peal—that Rodgers was unlikely to ever obtain alternative sentencing. Specifically, he asks us to infer from this position that the State was unlikely to make something so improbable a sticking point in negotiations. But this, too, is post hoc spec- ulation. Even Rodgers, elsewhere in his briefing, acknowl- edges that a party’s strategic incentives can change over time. We think this succinctly explains why his emphasis on a years-later litigating position is without merit. The circuits that have applied the better-plea-deal theory have done so only in the immigration context, and only in nar- row circumstances, where the record showed the government was willing to accept a plea that did not result in the defend- ant's removal or where the defendant negotiated the plea spe- cifically to avoid removal. See, e.g., Kovacs v. United States, 744 F.3d 44, 52–53 (2d Cir. 2014); United States v. Rodriguez-Vega, 797 F.3d 781, 788–89 (9th Cir. 2015); United States v. Swaby, 855 12 No. 24-1231 F.3d 233, 241–42 (4th Cir. 2017); Rodriguez-Penton v. United States, 905 F.3d 481, 488 (6th Cir. 2018); see also United States v. Dominguez, 998 F.3d 1094, 1117 n.14 (10th Cir. 2021) (rejecting theory and considering it an open question whether other cir- cuits intended their decisions to apply beyond immigration). Rodgers presents no evidence that the State offered or would have accepted a concrete alternative plea, and we see no basis to extend the theory to the circumstances presented here. So, even if we were to embrace this theory of prejudice, Rodgers supports it with speculation—and only speculation. We require contemporaneous evidence, including that the State would have actually offered Rodgers’s preferred terms. He presents none. C. Opt for Trial Rodgers can satisfy the prejudice prong if he can establish that, “but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.” Hill, 474 U.S. at 59. He has pursued this theory since state court and now does so in the alternative. Again, however, Rodgers fails to overcome Lee’s requirement of contemporaneous evidence. See 582 U.S. at 368. Rodgers offers three proofs in support of his argument that he might have chosen trial over the fixed forty-five-year sentence. He first points us to Pidgeon v. Smith, where we cred- ited post hoc testimony of a defendant’s preference for trial “because he face[d] the exact same decision today as he did when he initially accepted the plea deal—to either spend ten years in prison pursuant to the plea or to proceed to trial.” 785 F.3d 1165, 1174 (7th Cir. 2015). We added: “unlike a defendant who has rejected a plea deal and then loses at trial, Pidgeon No. 24-1231 13 has no incentive to lie about his earlier choice.” Id. (emphasis in original). Pidgeon, however, issued two years before Lee. We can no longer rely solely on post hoc assertions, even if the in- centives theoretically align. See Nissenbaum, 148 F.4th at 556; United States v. Barr, 960 F.3d 906, 918 (7th Cir. 2020) (rejecting assertion in defendant’s post hoc affidavit filed before he was committed to custody). Rodgers next, and more fulsomely, emphasizes his at- tempt to withdraw his plea at sentencing, describing it as con- temporaneous evidence that he would have made a different decision if fully informed. 3 He likens his situation to that in Anderson v. United States, where an attorney failed to investi- gate and dispute facts relevant to a sentencing enhancement, and where the defendant’s statements at his change-of-plea hearing showed he was dubious of the government’s presen- tation of those same facts. 981 F.3d 565, 577–78 (7th Cir. 2020). We see things differently. Regarding his attempted withdrawal, Rodgers asks us to infer some generalized concern over longer-than-necessary sentences rather than a point-in-time willingness to go to trial. This makes sense; there is an inherent contradiction in Rodg- ers citing an attempt to set aside his plea at sentencing—long before he learned of counsel’s errors—as contemporaneous evidence that, “but for counsel's errors, he would not have pleaded guilty and would have insisted on going to trial.” 3 Rodgers does not challenge the court’s denial of his request to set aside the guilty plea—only his attorney’s performance. Nor does he argue that, if counsel were effective, the court would have granted the request. He says only that he would have made a different decision if properly ad- vised. 14 No. 24-1231 Hill, 474 U.S. at 59 (emphasis added). We must therefore con- sider the actual substance of his attempt to withdraw the plea, specifically to the extent it shows “what [was] important to the defendant asserting the ineffective assistance claim.” United States v. Murillo, 927 F.3d 808, 818 (4th Cir. 2019); see also Lee, 582 U.S. at 361, 370 (emphasizing “the paramount im- portance Lee placed on avoiding deportation,” which his at- torney erroneously advised him was not a consequence to pleading guilty). Rodgers, however, sought withdrawal from his plea because he suddenly claimed that he was actually in- nocent—a theory separate from what he now argues was an error on counsel’s part. By way of comparison, the Anderson defendant expressed reservations about the very facts we said his attorney should have investigated. 981 F.3d at 577–78. Rodgers, to his credit, does not shy away from the incon- gruity; rather, he argues that his actions at sentencing show that, for him, the length of his sentence was of paramount im- portance—and that he was always willing to risk a guarantee if the possibility of something better existed. We cannot agree. True, Rodgers said at sentencing that he “lack[ed] knowledge of what [he] was getting [him]self into.” In context, though, that statement—and his more general request to set aside the plea—concerned the question of guilt, not apprehension over his sentence. Indeed, if true that by attempting to withdraw his plea at sentencing, he was simply searching for anything better than forty-five years, nothing prevented him from refusing the deal in the first place. Rodgers always had the option to pro- ceed to trial. He declined to do so. In sum, we cannot credit Rodgers’s attempt to set aside the plea at his sentencing as contemporaneous evidence at the No. 24-1231 15 time of his plea that he might have elected to go to trial. His statements suggest only that he did so because he suddenly claimed he was innocent, not because he was always willing to gamble to minimize his sentence. Finally, Rodgers emphasizes the State’s and his own attor- ney’s statements at the two hearings. Specifically, he says both attorneys framed the agreement as “the absolute best deal for Rodgers,” and that this characterization suggests that Rodg- ers expressed in negotiations that he was searching for “any viable path to a lesser sentence.” Once more, we cannot agree. Neither attorney described the agreement as the “absolute best deal,” but rather as being in Rodgers’s best interest. There is daylight between the two, especially since forty-five years was the minimum the State could itself guarantee. Meanwhile, regardless of the words used, neither attorney referenced Rodgers’s backroom bargaining priorities. They simply of- fered their own characterizations of the deal. Anything else is pure speculation. And even if Rodgers had demanded the absolute best deal, we cannot assume that the State would have offered the “better” deal. So, would Rodgers have chosen trial—and risked eighty-five years—simply because there was a chance that the court would go lower than forty-five, even though all agreed that the evidence against him was strong? Indeed, “[t]he decision whether to go to trial or plead guilty involves ‘assessing the respective consequences of a conviction after trial and by plea.’” Anderson, 981 F.3d at 577 (quoting Lee, 582 U.S. at 358). But again, nothing from the plea hearing sug- gests that Rodgers was a gambler, or that he viewed forty-five years as so “similarly dire” to the worst-case scenario that it warranted a “Hail Mary.” See Lee, 582 U.S. at 367–68. To the 16 No. 24-1231 contrary, Rodgers showed then that he was comfortable with forty-five years even though, at his hearing, there arose the contingent possibility of something better. Rodgers offers no relevant contemporaneous evidence, so his alternative theory of prejudice fails. His ineffective assis- tance of counsel claim therefore fails, too. III. Conclusion Rodgers may well have rejected the fixed forty-five-year sentence if properly informed of the alternative sentencing scheme. Perhaps he would have gone to trial, or perhaps the State would have offered something better. But the evidence he offers amounts only to post hoc assertions, unsupported as- sumptions, and bare speculation. Without more, he cannot show prejudice—no matter the theory. The district court’s judgment is therefore affirmed.