concurring:
I join the majority opinion as to Parts II.B and II.C, on Irwin’s ineffective assistance of counsel claims and his request for an evidentiary hearing. I respectfully concur in the judgment as to Part II.A on Irwin’s insufficiency of the evidence claim. We heard oral argument in this case on May 17, 2012. On May 29, 2012, the Supreme Court decided Coleman v. Johnson, —. U.S.-, 132 S.Ct. 2060, 182 L.Ed.2d 978 (2012), a case that, in my view, deprived Irwin of what had previously been a meritorious insufficiency of the evidence claim. It is clear that New York law places substantive limitations on what inferences a jury may draw in concluding that a defendant caused a substantial risk of death within the meaning of N.Y. Penal Law §§ 120.10(1) and 10.00(10), and that the jury in this case exceeded them. It is also clear that after Johnson, such limitations are irrelevant in habeas.
The majority concedes that the only fact in this case from which the jury could possibly conclude that Irwin was guilty beyond a reasonable doubt of assault in the first degree was Smalls’s loss of two liters of blood. I agree that such a fact could theoretically allow a jury to conclude that Irwin caused a substantial risk of death. But in my view, New York law— properly applied — required the prosecution to introduce evidence in addition to the mere fact of that blood loss to allow the jury to draw the conclusion it created a substantial risk of death. As it was, the prosecution provided the jury no evidence, outside of the blood loss itself, from which they could reasonably conclude that the risk of death was substantial, rather than merely extant. Perhaps in implicit recognition of this lack of trial evidence, the government includes on appeal an extra-record footnote explaining the significance of two liters of blood in relationship to the *157total amount of blood in the human body— a simple notation, but a powerful one. This is the sort of evidence which, had it been provided to a jury, would ground in reason the conclusion that Smalls’s blood loss constituted proof beyond a reasonable doubt that Irwin caused a substantial risk of death.
Without such evidence, I cannot conclude that, as a matter of New York law at least, the state proved Irwin caused a substantial risk of death beyond a reasonable doubt. Other than the Appellate Division’s decision here, I have found no New York case in which the bare and unsubstantiated possibility that a person might bleed to death absent medical intervention has constituted a substantial risk of death. New York’s jurisprudence compels the opposite conclusion. In People v. Sleasman, 24 A.D.3d 1041, 805 N.Y.S.2d 736 (3d Dep’t 2005), the Third Department reduced a first-degree assault conviction where the victim “had sustained, among other injuries, a knife wound to her neck.” Id. at 1042, 805 N.Y.S.2d 736. The victim awoke twelve hours after she was stabbed to find herself “covered with dried blood.” Id. After proceeding by ambulance to the hospital, her “wound was closed with butterfly sutures, [and] she was kept overnight for observation and released the following day.” Id. Despite the severity of the injuries the victim sustained, the Appellate Division still found the evidence insufficient to show that the risk of death was substantial. The court wrote that
[w]hile [the victim’s] injury was characterized as having the potential to be life threatening, tests conducted at the hospital determined the wound to be a cut in the platysma muscle and disclosed no damage to her trachea, esophagus or the major vessels in her neck. She was given two units of blood, which stabilized her heart rate and blood pressure. Although testing revealed that her CPK enzyme level was elevated and a cause for concern, cross-examination revealed that the risk of death was not substantial. After reviewing and weighing all of the evidence in the record, we cannot say that it supports a finding of a substantial risk of death
Id. at 1042-43, 805 N.Y.S.2d 736. The victim in Sleasman sustained blood loss at least as significant, if not more so, than that sustained by Smalls, since she, too, was “covered in blood” and, unlike Smalls, actually required a blood transfusion. The Sleasman victim was also stabbed in the neck, a far more vulnerable part of the body than the hand or arm, where Smalls was injured. And yet the Appellate Division found this evidence insufficient for assault in the first degree. This case suggests that, at a minimum, a New York jury’s ability to infer a substantial risk of death based only on blood loss — or even, as in Sleasman, significant blood loss in combination with injury to a vulnerable part of the body — is limited. And while Sleasman is most relevant to our analysis, it is far from the only New York case suggesting such a limitation. See, e.g., People v. Nimmons, 95 A.D.3d 1360, 1360, 945 N.Y.S.2d 358 (2d Dep’t 2012) (reducing second degree assault conviction where victim suffered a gunshot wound to the chest, despite emergency medical technician’s testimony as to the potential life threatening consequences because “the EMT’s testimony, along with the victim’s medical records, which were not explained or amplified by the testimony of a health care provider, were legally insufficient to establish that the injury to this victim ‘create[d] a substantial risk of death’ ”); People v. Tucker, 91 A.D.3d 1030, 1031-32, 936 N.Y.S.2d 386 (3d Dep’t 2012) (reducing a first degree assault conviction where victim was stabbed eight times, including a wound “approximately four inches long *158and ¿Hi inches deep [which] transected the victim’s rectus abdominis muscle,” where bleeding was “not massive” and was “stopped with a few sutures,” despite testimony from treating emergency room physician “that the wounds collectively ‘could[] have caused substantial risk of death’ ”); People v. Matos, 107 A.D.2d 823, 824, 484 N.Y.S.2d 844 (2d Dep’t 1985) (finding evidence to be insufficient to support a conviction for first degree assault where victim of a gunshot wound to the leg required a three-day hospitalization).
The cases on which the majority and the government rely do not change this understanding of New York law. Unlike this case they all contain crucial other facts which would support the jury’s finding of assault in the first degree. In People v. Jeanty, for example, on which the Appellate Division relied on in denying Irwin’s claim, the victim “sustained seven deep head or facial lacerations which were 1 to 3 inches in length and penetrated either the skull membrane or bone, requiring two layers of 50 sutures and resulting in permanent scarring.” 268 A.D.2d 675, 678, 702 N.Y.S.2d 194 (3d Dep’t 2000). Given the disparity between the Jeanty victim’s injury and Smalls’s, the Appellate Division’s determination there that a substantial risk of death had been shown because had the victim’s “injuries [ ] been left untreated he could have bled to death,” id., cannot fairly be used to support Irwin’s conviction. In People v. Gordon, also cited by the majority, sufficient evidence was found where the victim was stabbed “just below the rib cage” with a seven-inch knife. 257 A.D.2d 533, 533-34, 685 N.Y.S.2d 28 (1st Dep’t 1999). Though the court noted that the victim “bled copiously from the resulting wound, to the extent that her shirt was soaked with blood,” there was also evidence that she “was subjected to a CAT scan, an X-ray, and monitoring ‘for a couple of days’ to determine whether the heart, intestines or other vital organs had been damaged, which was a possibility from a knife wound in this area of the body.” Id. In People v. House, the Appellate Division specifically indicated that “medical proof established that the stab wound to the victim’s chest was potentially lethal.” 278 A.D.2d 244, 245, 717 N.Y.S.2d 273 (2d Dep’t 2001). No such medical proof established the potential lethality of the wounds in this case. People v. Riccardi is similarly inapposite. 199 A.D.2d 432, 432, 605 N.Y.S.2d 112 (2d Dep’t 1993) (“The victim’s own testimony as well as that of the Emergency Medical Technician and the uncontroverted evidence of the People’s medical expert that the wounds, if left untreated, were ‘life-threatening’, is sufficient to support the jury’s verdict.”). Likewise, the victim in People v. Thompson, 224 A.D.2d 646, 647, 639 N.Y.S.2d 52 (2d Dep’t 1996), sustained not only “substantial bleeding” as a result of the stab wound to his chest, but also suffered a collapsed lung requiring treatment and hospitalization. None of these cases involve a conclusion of substantial risk of death based only on bleeding. They all contain other evidence, whether it be testimony from medical experts, as in Riccardi, or evidence of extreme wounds, as in Jeanty, from which a jury could reasonably draw the conclusion that a substantial risk of death was attendant. In the absence of some evidence in addition to the mere fact of blood loss, New York law simply does not permit the inference the jury made here.
I recognize, of course, that “[a] federal court may not issue the writ on the basis of a perceived error of state law.” Pulley v. Harris, 465 U.S. 37, 41, 104 S.Ct. 871, 79 L.Ed.2d 29 (1984). But the error in this case as to what inferences a jury may draw, seems part and parcel of a federal question: what is sufficient proof beyond a reasonable doubt? Or at least that was so before Johnson, which compels the conclu*159sion that the limitations which inhere in New York law are irrelevant to our analysis of this case. Johnson tells us that it would be error “to look to [state] law in determining what distinguishes a reasoned inference from ‘mere speculation.’” 132 S.Ct. at 2064. And while we continue to “look to state law for the substantive elements of the criminal offense ... the minimum amount of evidence that the Due Process Clause requires to prove the offense is purely a matter of federal law.” Id. (internal quotation marks and citation omitted). Johnson makes explicit that this minimum amount of evidence and what inferences can be drawn from it is guided solely by Jackson v. Virginia, 443 U.S. 307, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979). Under Jackson, juries have “broad discretion in deciding what inferences to draw from the evidence presented at trial, requiring only that jurors ‘draw reasonable inferences from basic facts to ultimate facts.’ ” Johnson, 132 S.Ct. at 2064 (quoting Jackson, 443 U.S. at 319, 99 S.Ct. 2781). That discretion is so broad that “the only question under Jackson is whether [the jury’s] finding was so insupportable as to fall below the threshold of bare rationality.” Id. at 2065. And the only question under AEDPA is whether the state court’s conclusion that that finding was not so insupportable was unreasonable. Though I have no doubt the state court was wrong, its determination that the jury’s leap from blood loss to substantial risk of death did not fall below the threshold of bare rationality was not “manifestly contrary to common sense,” Anderson v. Miller, 346 F.3d 315, 324 (2d Cir.2003). Accordingly, I must reluctantly agree with the analysis of my colleagues and conclude that Irwin is not entitled to habeas relief.