OPINION AND ORDER
KATHERINE POLK FAILLA, District Judge:On December 10, 2007, Petitioner Brian Carmichael was convicted in the Supreme Court of New York, New York County, of three counts of second-degree sale of a controlled substance, for which he is now serving a seventeen-year sentence. On December 19, 2014, Petitioner filed a petition for a writ of habeas corpus, pursuant to 28 U.S.C. § 2254, in this Court (the “Petition”). In sum, the Petition argues that: (i) Petitioner’s conviction was obtained in violation of Batson v. Kentucky, 476 U.S. 79, 106 S.Ct. 1712, 90 L.Ed.2d 69 (1986); and (ii) Petitioner received ineffective assistance of counsel. On July 17, 2015, Magistrate Judge Andrew J. Peck issued a Report and Recommendation (the “Report”), in which he recommended that this Court deny Petitioner’s request for relief. Petitioner timely objected to Magistrate Judge Peck’s conclusions. For the reasons set forth in the remainder of this Opinion, this Court respectfully declines to adopt the Report, and grants the Petition.
BACKGROUND1
The Report provides a thorough factual and procedural history of this case. (Report 2-18). As a result, this Court will only recount the portion of this history that is most relevant here: the conduct of the parties during jury selection.
Petitioner was charged in the Supreme Court of New York, New York County, with multiple drug crimes.' (Petition 2-3). On September 17, 2007, Justice Robert Straus began selecting a jury to hear Petitioner’s case. (See T. 1-2). The Court gave each side 20 peremptories that could be used to strike prospective members of the twelve-person jury. (T. 169). The Court also gave each party six peremptories that could be used to strike prospective alternates. (See T. 416).
The jury was selected from three separate panels, each of which contained 26 prospective jurors. (See T. 1-423). After the trial court and the parties questioned the prospective jurors on the first panel, the court asked whether the prosecutor wished to exercise any peremptory challenges against the first twelve individuals on that panel. (T. 171, 173). The prosecutor exercised five challenges. (Id. at 173). Then, the court asked defense counsel whether he wished to strike any of the remaining individuals in seats 1 through 12. (See id.). Defense counsel struck three people. (Id.).
*79The trial court next considered the individuals in seats 13 through 24 of the panel. (T. 174-77). The court excused three for cause, and the parties agreed to excuse a fourth because she had a high-risk pregnancy. (See id.). Then, the court asked whether the prosecutor wished to strike anyone else in seats 13 through 24. (Id. at 180). The prosecutor challenged four people. (Id.). At this point, defense counsel raised a Batson objection. (Id.). Defense counsel explained that the panel was “sparse of minorities,” and the prosecutor had challenged the only two African-American jurors wh'o had been considered for service, Ms. Boiken (in seat number 7) and Ms. Hamilton (in seat number 21). (Id. at 181). The court rejected the Batson' claim because, in its view, Petitioner had not established a prima facie case of discrimination. (Id.). The court went on to ask defense counsel whether he wished to exercise any peremptory challenges, and defense counsel struek the remaining four individuals in seats' 13 through 24. (Id. at 182).
When the court turned to the individuals in seats 25 and 26 of the first panel—Ms. Velarde and Mr. Sweeny—the prosecutor declined to exercise a peremptory challenge. (T. 183). Defense counsel, however, struck both prospective jurors. (Id.). The court later suggested that one of these individuals (Ms. Velarde) might be African-American, but defense counsel insisted that she was Hispanic. (Id. at 321-23).
Once the parties finished discussing the jurors in the first panel, the court asked whether they wished to withdraw any of their peremptory challenges. (T. 184). The prosecutor withdrew his challenge against the individual in seat 6 and defense counsel withdrew his challenge against the individual in seat 20, on the condition that these jurors would serve as alternates; neither individual was African-American. (Id.).
The court ■ proceeded to fill a second panel of 26 individuals. (T. 233-34). After these prospective jurors were questioned, the court struck the person in the first seat for cause, and asked , the parties whether they wished, to exercise peremptory challenges against anyone in seats 2 through 9. (T. 266, 311-12). The prosecutor struck one prospective juror and defense counsel struck six more. (Id. at 312).
Then, the court asked whether the parties wished to challenge any of the individuals in seats 10 through 16. (T. 312). Both the prosecutor and defense counsel exercised two peremptory challenges' against members of this group. (Id. at 314-15). Notably, however, neither attorney struck Ms. Bode (in seat number 10), who was African-American. (See id. at 314-15, 320).
Next, the court considered the prospective jurors in seats 17 through 21. (T. 315). The court struck one of these individuals for cause, and then asked whether the prosecutor wished to exercise any peremptory challenges. (Id. at 315, 320). The prosecutor stuck two people, including Ms. Grant (in seat number 21), who was African-American. (Id. at 320-21). In response, defense counsel renewed his Batson objection. (Id.). Defense counsel observed that
[although [the prosecutor] ... allowed Ms. Bode to remain on the jury, the lone black juror selected so far, he has challenged Ms. Grant[,] who is an African American juror.
So I believe out of the four African American jurors we have considered on the panel[,] [the prosecutor] has challenged three of them.
(Id.). Once again, the court concluded that Petitioner had not stated a prima facie case of Batson discrimination, and asked whether defense counsel wished to 'Strike any of the prospective jurors under consideration. (See id. at 323-23a). Defense counsel struck one individual. (Id. at 323a).
*80The court went on to discuss the prospective jurors in seats 22 through 24. (T. 323a). It excused juror 22 for cause, and then asked the prosecutor whether he wished to strike juror 23 or 24. (Id.). The prosecutor struck both, and defense counsel raised a third Batson objection. (Id. at 323b). Defense counsel noted that Ms. Simmons, in seat- number 24, was African-American; consequently, the prosecutor had stricken four out of the five African-Americans who had been considered for jury service. (Id.). Counsel suggested that this statistic was particularly troubling because the parties “had probably 140 people that [they had] considered in two days [of jury selection],” only five of whom were African-American. (Id.). Still, the court maintained that defense counsel had not articulated a 'prima facie case of Batson discrimination. (Id. at 323c). Consequently, the court moved on to consider the prospective jurors in seats 25 and 26. (Id.). The individual in seat 25 was excused on consent of the parties, and the prosecutor used a peremptory strike to remove juror 26. (Id. at 266-67, 323c).
The trial court filled a third panel of prospective jurors, and then asked the parties to consider the individuals in seats 1 through 3. (T. 413). The prosecutor struck the juror in the first seat, but the parties accepted the individuals in the second and third seats. (Id.). The juror in seat 3, Ms. Duggins, was African-American. (Id. at 420).
At this point, the court asked the parties to consider prospective jurors one at a time. (See T. 413-21). The prosecutor struck the person in seat 4, and the person in seat. 5 was excused on consent of the parties; the person in seat 6 became the last juror. (Id. at 413-16).
The court then turned to the task of selecting alternates. (T. 416). The prosecutor struck three potential alternates, two of whom were African-American. (Id. at 416-20). As a result, defense counsel raised a fourth Batson challenge, explaining:
It seems again that [the prosecutor] is exercising his challenges to exclude African Americans. I do note that as we proceeded with selection he did not challenge Ms. Duggins who was the sixth in my view African American that we have considered ..., but when we got to the alternates he challenged Ms. Sanders a black female and now he’s also challenging Mr. Pratt who is a male black, so I see a clear pattern of challenging African Americans, your Honor. I make my Batson challenge on that basis. Four of the six we have considered have been challenged. We have been through three panels so approximately 210 have come into this courtroom. Only considered—-we have considered in total eight African Americans and six of those have been challenged by [the prosecutor] in my view.
(Id. at 420). For a final time, the court rejected counsel’s Batson claim, without requiring the prosecutor to state the reasons for any of his peremptory strikes on the record. (Id).
After the court determined that there had been no Batson violation, the parties selected the man in seat number 16 as the fifth and final alternate. (T. 421). Thus, the court did not consider any of the individuals in seats 17 through 26. (See id. at 421-23).
DISCUSSION
A. Applicable Law
1. The Standard of Review
a. Reviewing the Report and Recommendations of a Magistrate Judge
A court may accept, reject, or modify, in whole or in part, the findings or *81recommendations made by a magistrate judge. See 28 U.S.C. § 636(b)(1); Fed. R. Civ. P. 72(b)(3); Grassia v. Scully, 892 F.2d 16, 19 (2d Cir.1989). A court may accept those portions of a report to which no specific, written objection is made, as long as the factual and legal bases supporting the findings are not clearly erroneous. See Ramirez v. United States, 898 F.Supp.2d 659, 663 (S.D.N.Y.2012). By contrast, when a petitioner makes specific objections to a magistrate judge’s findings, the reviewing court must undertake a de novo review of those findings. See 28 U.S.C. § 636(b)(1); Fed. R. Civ. P. 72(b)(3); United States v. Male Juvenile, 121 F.3d 34, 38 (2d Cir.1997).
b. Reviewing State Court Decisions Under the Antiterrorism and Effective Death Penalty Act
Under the Antiterrorism and Effective Death Penalty Act of 1996 (“AED-PA”), a federal court cannot grant a petition for a writ of habeas corpus based on a claim that was “adjudicated on the .merits in State court proceedings” unless the state court’s decision: (i) “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States”; or (ii) “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d). This is a “highly deferential standard for evaluating state-court rulings, which demands that state-court decisions be given the benefit of the doubt.” Cullen v. Pinholster, 563 U.S. 170, 181, 131 S.Ct. 1388, 179 L.Ed.2d 557 (2011) (quoting Woodford v. Visciotti, 537 U.S. 19, 24, 123 S.Ct. 357, 154 L.Ed.2d 279 (2002) (per curiam)).
Federal law is “clearly established” when it is expressed in “the holdings, as opposed to the dicta, of [the Supreme Court’s] decisions.” Howes v. Fields, - U.S. -, 132 S.Ct. 1181, 1187, 182 L.Ed.2d 17 (2012) (internal quotation marks omitted). A state court’s decision is “contrary” to clearly established federal law when the state court “applies a rule that contradicts the governing law set forth in” a Supreme Court opinion or “confronts a set of facts that are materially indistinguishable from a decision of [the' Supreme] Court and nevertheless arrives at a ... different [result].” Williams v. Taylor, 529 U.S. 362, 405-06, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000). And a state court’s decision can only be considered “unreasonable” if “there is no possibility fairminded jurists could disagree that the state court’s decision conflicts with [the Supreme] Court’s precedents.” Harrington v. Richter, 562 U.S. 86, 102, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011); see also Woods v. Donald, - U.S. -, 135 S.Ct. 1372, 1376, 191 L.Ed.2d 464 (2015) (per curiam) (explaining that AEDPA only allows federal habeas courts to overturn state court decisions “when there could be no reasonable dispute that they were wrong”); Vega v. Walsh, 669 F.3d 123, 126 (2d Cir.2012) (same).
When a federal court reviews a state court’s factual determinations, those decisions “shall be presumed to be correct,” and that presumption can only be rebutted by “clear and convincing evidence.” 28 U.S.C. § 2254(e)(1); see also McKinney v. Artuz, 326 F.3d 87, 101 (2d Cir.2003).
2. The Batson Framework
In Batson, the Supreme Court held that, under the Equal Protection Clause, it is impermissible for a prosecutor to exercise peremptory challenges against potential jurors “solely on account of their race.” 476 U.S. at 89, 106 S.Ct. 1712. Indeed, it is a violation for a prosecutor to strike even a single juror with a discriminatory purpose. *82See Walker v. Girdich, 410 F.3d 120, 123 (2d Cir.2005) (“[UJnder Batson and its progeny, striking even a single juror for a discriminatory purpose is unconstitutional,”); see also United States v. Vasquez-Lopez, 22 F.3d 900, 902 (9th Cir.1994) (“[T]he Constitution forbids striking even a single prospective juror for a discriminatory purpose.”); Jones v. Ryan, 987 F.2d 960, 972 (3d Cir.1993) (“[T]he exclusion of even one juror on the basis of race may be sufficient to establish a prima facie case [of discrimination].” (internal citation omitted)).
Courts follow a three-step procedure to determine whether a prosecutor has used a peremptory strike to exclude jurors with a particular racial or ethnic background. First, courts consider whether a defendant has established a “prima facie case” of discrimination “by showing that the totality of the relevant facts” supports an inference that the prosecutor acted with a “discriminatory purpose.” Id. at 93-94, 106 S.Ct. 1712. “Second, once the defendant has made out a prima facie case, the ‘burden shifts to the [prosecutor] to explain adequately the racial exclusion’ by offering permissible race-neutral justifications for the strikes.” Johnson v. California, 545 U.S. 162, 168, 125 S.Ct. 2410, 162 L.Ed.2d 129 (2005) (quoting Batson, 476 U.S. at 94, 106 S.Ct. 1712). Third, “[i]f a race-neutral explanation is tendered, ... court[s] must then decide ... whether the opponent of the strike has proved purposeful racial discrimination.” Purkett v. Elem, 514 U.S. 765, 767, 115 S.Ct. 1769, 131 L.Ed.2d 834 (1995) (per curiam).
B. Analysis
Because Petitioner timely objected to the. legal conclusions reflected in the Report, this Court will review those conclusions de novo. See 28 U.S.C. § 636(b)(1); Fed. R. Civ. P. 72(b)(3); Male Juvenile, 121 F.3d at 38.
1. The Trial Court Acted Contrary to Clearly Established Federal Law When It Determined That Petitioner Had Not Established a Prima Facie Case of Batson Discrimination
The only case that the trial court cited in the course of its Batson analysis was People v. Brown, 97 N.Y.2d 500, 743 N.Y.S.2d 374, 769 N.E.2d 1266 (2002) (“Brown I”). (See T. 181-82). Thus, in order to assess the constitutionality of the trial court’s Batson rulings, it is necessary to examine the Brown I case in some detail.
The prosecutor in Brown I used seven of his first eight peremptory challenges to strike African-American members of the venire. Brown I, 97 N.Y.2d at 508, 743 N.Y.S.2d 374, 769 N.E.2d 1266. Relying on this statistic, the defendant raised a Bat-son challenge, but the trial court found that she had not established a prima facie case of discrimination. Id. The New York Court of Appeals upheld this decision, noting that, when & prosecutor uses a “disproportionate number of strikes ... against members of a particular racial or ethnic group,” that fact is “rarely” sufficient to establish a prima facie case of discrimination “in the absence of other facts or circumstances.” Id. at 507, 743 N.Y.S.2d 374, 769 N.E.2d 1266. Applying this rule, the court found:
[T]he People’s removal of seven African-Americans through the exercise of eight peremptory challenges was inadequate, without more, to require the trial court to find a prima facie showing of discrimination. After defendant raised her Batson challenge during the second round of voir dire, the Judge stated that, by his count, nine *83potential jurors in the first panel and six in the second panel appeared to be African-American and as such the People had challenged 7 of the 15 African-Americans in the venire. Further, four of the seven sworn jurors were African-American.
Defendant was explicitly invited by the trial court to articulate any facts and circumstances that would support a prima facie showing of discrimination. Instead of making “a record comparing Caucasians accepted with similarly situated African-Americans challenged, or by establishing objective facts indicating that the prosecutor has challenged members of a particular racial, group who might be expected to favor the prosecution because of their backgrounds” ([People v.] Bolling, 79 N.Y.2d [317] at 324, 582 N.Y.S.2d 950, 591 N.E.2d 1136 [(1992)]), defense counsel responded that certain persons excused by prosecution per-emptories had no prior jury service or had attended college and, thus, gave no indication that they could not be “fair.”
Based on the numbers and arguments presented, the trial court ruled that it did not find a discriminatory pattern. No further Batson objection was raised during the remainder of voir dire proceedings. Upon this record, we conclude that defendant’s numerical argument was unsupported by factual assertions or comparisons that would serve as a basis for a prima facie case of impermissible discrimination (see [People v.] Jenkins, 84 NY.2d [1001] at 1003, 622 N.Y.S.2d 509, 646 N.E.2d 811 [ (1994) ]; Steele, 79 N.Y.2d at 325, 582 N.Y.S.2d 950, 591 N.E.2d 1136).
Brown I, 97 N.Y.2d at 508, 743 N.Y.S.2d 374, 769 N.E.2d 1266.
There are two ways to interpret the Brown I decision, one of which is consistent with federal constitutional law and one of which is not. First, the decision might be read to observe that one particular statistic—the percentage of strikes used against African-American members of the venire (or any other racial or ethnic group)—is not particularly persuasive evidence of discrimination when it is considered in isolation. Brown I, 97 N.Y.2d at 507, 743 N.Y.S.2d 374, 769 N.E.2d 1266. However, that statistic may be more persuasive when it is considered together with other numerical information, such as the number of African-Americans in the venire and number of African-Americans who have already been sworn in or seated as jurors. Id. at 508, 743 N.Y.S.2d 374, 769 N.E.2d 1266. For example, if a prosecutor uses 60% of his peremptory strikes against African-Americans, but only 10% of the individuals on the venire and 10% of sworn jurors are African-American, then the prosecutor’s use of strikes may be highly suspicious; by contrast, if the prosecutor uses 60% of his peremptory strikes against African-Americans, but 60% of the individuals on the venire and 60% of sworn jurors are African-American, then the prosecutor’s use of strikes may seem perfectly benign. Similarly, the percentage of strikes used against African-Americans may be more or less persuasive evidence of discrimination when it is considered together with non-numerieal information, such as statements made by various individuals during voir dire. Id. If the prosecutor uses a high percentage of his peremptory challenges against African-Americans, even though these African-Americans provided roughly the same answers to voir dire questions as other jurors, that may raise a red -flag. But a prosecutor’s decision to strike individuals who seem unusually sus*84picious of law enforcement or sympathetic toward the defendant may be less problematic. If Brown I is read in this manner, it is perfectly consistent with federal constitutional law.
There is, however, a second, more troubling interpretation of Brown I. According to this second interpretation, Brown I does not simply make an observation about the persuasive power of a particular statistic; it establishes a presumption that “numerical” information is insufficient to satisfy the defendant’s burden at step one of the Batson inquiry unless it is supplemented with non-numerical information. Brown I, 97 N.Y.2d at 508, 743 N.Y.S.2d 374, 769 N.E.2d 1266. Thus, even when a defendant can identify a robust set of statistics suggesting that the prosecutor has used a “disproportionate number of strikes ... against members of a particular racial or ethnic group,” those statistics will “rarely” give rise to an inference that the prosecutor’s strikes were racially motivated. Id. at 507, 743 N.Y.S.2d 374, 769 N.E.2d 1266.
If Brown I creates a presumption that numerical evidence—by itself—is insufficient to support a prima facie case of discrimination, then it is “contrary to ... clearly established federal law,” as articulated by the Supreme Court. 28 U.S.C. § 2254(d); cf. Truesdale v. Sabourin, 427 F.Supp.2d 451, 459-60 (S.D.N.Y.2006) (finding that a New York court acted contrary to clearly established federal law when it adopted a presumption that numerical arguments, “presented alone, [must] be sufficiently ‘compelling as to be conclusive’ of discrimination”). Batson clearly contemplated that numerical evidence, taken alone, could support an inference of discrimination. See Batson, 476 U.S. at 93, 106 S.Ct. 1712 (“[S]eriously disproportionate exclusion of Negroes from jury venires is itself such an unequal application of the law [ ] as to show intentional discrimination.” (internal quotation marks and citations omitted)); id. at 97, 106 S.Ct. 1712 (“[A] ‘pattern’ of strikes against black jurors included in the particular venire might give rise to an inference of discrimination.”). Thus, to the extent Brown I creates a presumption that numerical evidence must be corroborated before it can support an inference of discrimination, the opinion makes it more difficult for New York defendants to move past the first step of the Batson inquiry. Such impediments plainly contravene the Constitution: In Johnson v. California, 545 U.S. 162, 125 S.Ct. 2410, 162 L.Ed.2d 129 (2005), the Supreme Court confirmed that it was impermissible for a state to heighten the bar that defendants must clear before courts will move to the second Bat-son step. See id. at 169-70, 125 S.Ct. 2410. It is impossible to square the second reading of Brown with the holding of Johnson.
The Report suggests that the constitutional problems with the second reading of Broum I are of less concern because the Second Circuit has not adopted this reading of the case. (See Report 30). After the New York Court of Appeals decided Brown I, the defendant filed a petition for a writ of habeas corpus in federal court, and the petition was denied. Brown v. Alexander, 543 F.3d 94, 96 (2d Cir.2008) (“Brown II''). On appeal, the Second Circuit adopted the first construction of the Court of Appeals’ decision, and concluded that this construction was not contrary to clearly established federal law. See generally id.
Crucially, however, the fact that the Second Circuit adopted a constitutional reading of the Broum I decision does not mean that New York state courts have adopted the same reading. New York courts are not bound by Second Circuit interpretations of state case law. See, e.g., People v. Kin Kan, 78 N.Y.2d 54, 60, 571 *85N.Y.S.2d 436, 574 N.E.2d 1042 (1991) (explaining that, the New York Court of Appeals is not bound by the decisions of lower federal courts); People v. Brown, 235 A.D.2d 344, 653 N.Y.S.2d 544, 544 (1st Dep’t 1997) (explaining that, when an intermediate state court is faced with “a conflict between the decisional law of the [New York] Court of Appeals and that of an intermediate Federal appellate court, the ruling of the state Court of Appeals should be followed”); People v. Battease, 74 A.D.3d 1571, 904 N.Y.S.2d 241, 247 (3d Dep’t 2010) (same). This is true even if the Second Circuit has placed a gloss on a state court decision in order to rescue that decision from federal constitutional shoals. See Kin Kan, 78 N.Y.2d at 60, 571 N.Y.S.2d 436, 574 N.E.2d 1042, (“[T]he interpretation of a Federal constitutional question by the lower Federal courts ... [is] not binding [on the New York Court of Appeals].”); cf. People v. Joyner, 303 A.D.2d 421, 755 N.Y.S.2d 866, 867 (2d Dep’t 2003) (rejecting a lower federal court’s conclusion that state law violated the federal constitution). Precisely for this reason, this Court may not presume that, whenever a state court cites Brown I, it is reading that case in a constitutional manner.2
Such a presumption would be particularly inappropriate in this case, as the trial court applied an unconstitutional construction of the Brown I opinion. In the course of making his Batson challenges, Petitioner’s attorney did not simply rely on the fact that the prosecutor used six of his 21 peremptories to strike African-American jurors. Cf. Brown I, 97 N.Y.2d at 508, 743 N.Y.S.2d 374, 769 N.E.2d 1266 (defense counsel relied on the fact that the prosecutor used 7 of his 8 peremptories to strike African-Americans). Rather, defense counsel explained that it was suspicious for the prosecutor to use so many of his per-emptories against African-American individuals because African-Americans comprised such a small fraction of the venire. (See T. 323b, 420). Defense counsel also suggested that the prosecutor’s use of per-emptories was problematic because the prosecutor had removed most of the African-Americans (six of eight or six of nine) who had been' considered for jury service. (See T. 321-23, 420). In the face of all this statistical evidence—and despite an expanding, consistent record of prosecutorial exercise of peremptory strikes against African-American jurors—the trial court rejected Petitioner’s Batson claims because it believed that, under “New York law,” a “percentage” argument is “generally not sufficient to raise ... an inference or create a prima facie case of discrimination.” (T. 323b-c; see also id. at 421 (explaining that Petitioner’s “statistical analysis by itself1’ did not support a prima facie case of discrimination)). Thus, the trial court applied a presumption that statistical evidence is generally insufficient to support an inference that a prosecutor is using peremptory challenges in a discriminatory manner. In so doing, the trial court acted contrary to clearly established federal law.
2. The Appellate Division Did Not Cure the Trial Court’s Error
Having determined that the trial court acted contrary to clearly established federal law, this Court must determine whether *86the Appellate Division cured the trial court’s error. In other words, the Court must ask whether the Appellate Division: (i) applied the correct legal standard to evaluate Petitioner’s Batson claims; and (ii) applied the legal standard in a reasonable way. If the answer to both of these questions is yes, then- Petitioner is not entitled to relief. See Johnson v. Williams, - U.S. -, 133 S.Ct. 1088, 1094 n. 1, 185 L.Ed.2d 105 (2013) (noting that the Ninth Circuit had correctly considered whether the “last reasoned state-court decision” in the case was consistent with federal law),
a. This Court Must Presume That the Appellate Division Applied the Correct Legal Standard When It Rejected Petitioner’s Batson Arguments
The Appellate Division’s Bat-son analysis was limited to the following paragraph:
The court properly denied defendant’s applications made pursuant to Batson v. Kentucky, 476 U.S. 79, 106 S.Ct. 1712, 90 L.Ed.2d 69 [1986]. Viewing jury selection as a whole, we conclude that defendant did not meet his burden at step one of the inquiry. Defendant did not produce “evidence sufficient to permit the trial judge to draw an inference that discrimination ha[d] occurred” in the exercise of peremptory challenges (Johnson v. California, 545 U.S. 162, 170, 125 S.Ct. 2410, 162 L.Ed.2d 129 [2005]). While numerical evidence may suffice, in this case it did not warrant an inference of discrimination.
People v. Carmichael, 73 A.D.3d 622, 901 N.Y.S.2d 48, 49 (1st Dep’t 2010). From these brief remarks, it is impossible to tell whether the Appellate Division factored Broum I into its decision, and, if so, whether it read Broum 7 in a constitutional or unconstitutional manner. Under AEDPA, however, “state-court decisions [must] be given the benefit of the doubt.” Cullen, 563 U.S. at 181, 131 S.Ct. 1388 (quoting Woodford, 537 U.S. at 24, 123 S.Ct. 357). As a result, this Court must presume that, if the Appellate Division applied Broum I, it placed a constitutional gloss on the case. In other words, this Court must presume that the Appellate Division applied the correct legal standard as it evaluated Petitioner’s Batson claim.
b. The Appellate Division Applied Federal Law in an Unreasonable Manner When It Rejected Petitioner’s Final Batson Argument
Assuming, as the Court must, that the Appellate Division applied the .correct legal standard to resolve the Batson issue, the remaining issue is whether it was reasonable for the Appellate Division to conclude that Petitioner was not entitled to relief. Petitioner does not dispute that, for each of his four Batson challenges, his trial counsel relied solely on numerical information to support his allegation that the prosecutor was using peremptory challenges in a discriminatory manner. More specifically, counsel relied on the number of African-American individuals in the venire, the number of non-African-American individuals in the venire, and the number of African-American individuals challenged by the prosecutor. (See T. 181, 320, 323b, 420). Consequently, this numerical information was the only evidence that was properly before the trial court and the Appellate Division as they decided whether Petitioner had established a prima facie case of discrimination. See People v. Calas, 134 A.D.3d 1043, 22 N.Y.S.3d 217, 219 (2d Dep’t 2015) (explaining that, under New York law, “[i]t is incumbent upon a party making a Batson challenge to articulate and develop all of the grounds supporting the claim, both factual and legal, during the colloquy in which the objection is raised and discussed” (quoting People v. *87Cuesta, 103 A.D.3d 913, 959 N.Y.S.2d 744, 746 (2d Dep’t 2013))).
Considering only the relevant numerical information, it was reasonable for the Appellate Division to uphold the trial court’s rulings on Petitioner’s first three Batson objections. All of these objections were lodged during the middle of voir dire, and state courts are given wide latitude to wait until the end of jury selection in order to determine whether a defendant has established a prima facie case of discrimination. See Brown II, 543 F.3d at 102 (“It is not ordinarily unreasonable for a state court to conclude that a petitioner has not made out a prima facie case when she raises a Batson challenge before jury selection is completed and before the facts are even fully established on the record.” (internal quotation marks, brackets, and ellipsis omitted)); Sorto v. Herbert, 497 F.3d 163, 170 (2d Cir.2007) (“The need to examine statistical disparities [at step one of the Batson inquiry] may commend a wait-and-see approach.”); Overton v. Newton, 295 F.3d 270, 280 (2d Cir.2002) (finding that it was reasonable for a state court to deny a Batson challenge raised “midway” through jury selection).
By contrast, it was unreasonable for the Appellate Division to uphold the trial court’s ruling on Petitioner’s fourth Batson objection, which was raised at the very end of jury selection. At that point, the trial court had the following data to consider: Over the course of the jury selection process, the trial court had questioned 68 potential jurors.3. Eleven had been excused for cause or on consent of the parties, leaving 57 individuals who were qualified to serve. Of these 57 individuals, 8 or 9 (14 to 16 percent) were African-American.4 Thus, one might have expected the prosecutor to use approximately 3 challenges (14 percent) against African-American individuals. Yet the prosecutor used 6 of his challenges (29 percent) to strike African-American jurors. In addition, the prosecutor eliminated 67 to 75 percent of the black jurors on the panel. Taken together, these data points are highly suspicious. See United States v. Alvarado, 923 F.2d 253, 256 (2d Cir.1991) (explaining that, when a prosecutor strikes black jurors at “nearly twice” the rate that one would expect given the composition of the venire, that fact “strongly supports a prima facie case under Batson”).
Moreover, there is nothing apparent from the record that could explain the prosecutor’s strikes against African-American individuals. Cf. United States v. Stephens, 421 F.3d 503, 515 (7th Cir.2005) (“[C]ourts considering Batson claims at the prima facie stage may consider apparent reasons for the, challenges discernible on the record, regardless of whether those reasons were the actual reasons for the challenge.”). For example, there is no evidence suggesting that á disproportionate number of African-American jurors had negative interactions with the police, or had friends or relatives who had been convicted of a crime. Nor is there evidence suggesting that the family and professional backgrounds of the African-American jurors were significantly different from the family and professional backgrounds of other jurors. Thus, the record does noth*88ing to allay concerns that the prosecutor may have exercised his peremptory challenges in a racially discriminatory manner.5 ■
It is true that the statistical evidence in this case is similar to the statistical evidence in Overton, where the Second Circuit held that it was reasonable for the state court to deny the petitioner’s Batson challenge without requiring the prosecutor to place race-neutral reasons for his strike on the record. See generally. Overton, 295 F.3d 270. In Overton, 34 percent of venire-people whose races were known were black, but the prosecutor used 70 percent of his first ten peremptory challenges against black jurors. See id, at 274-75, 279. Furthermore, in the first two rounds of jury selection, the prosecutor struck 70 percent of the black individuals who were qualified to serve. Id. at 274.
However, the Overton Court’s decision turned on the fact that defense counsel raised his Batson objection in the middle of jury selection, and did not renew it after all the jurors and alternates had been chosen. 295 F.3d at 279. As the Court explained:
Because [the Batson challenge was not renewed], the trial judge never confronted, and the trial record does not reveal, what the statistics would have shown at the conclusion of jury selection. If those statistics sufficiently established the inference that challenges were based on race, the court could then have implemented the Bat-son process to ensure that impermissible challenges would not be allowed. If, on the other hand, the statistics at the conclusion failed to support a sufficient inference, there would be no need to engage in the process. We cannot say, on this record, that the trial judge’s refusal to implement Bat-son’s process for testing each questioned challenge midway in the process was an unreasonable application of the Batson requirements.
Id. at 279-80. In this case, however, the record contains ample statistical information regarding the use of peremptory challenges over the entire jury selection process. Consequently, this Court is in a position to say what the Overton Court could not: statistical information from the second half of jury selection supports— rather than dispels—an inference that the prosecutor’s peremptory strikes were racially motivated. (See, e.g., T. 416-20 (at the tail end of jury selection, the prosecutor used two of his three peremptory strikes against prospective alternates to remove African-Americans)). Based on the complete set of statistical information presented to the trial court, it was unreasonable for the Appellate Division to conclude that Petitioner had not established a pri-ma facie case of discrimination.
The Report articulates four arguments in support of its contrary conclusion. Each of them finds some support in the record and the relevant case law, but, individually and collectively, they are ultimately unavailing. First, the Report suggests that Petitioner’s Batson challenge turns on the *89percentage of African-American venirepeo-ple who were excluded from the jury (the “exclusion rate”). (Report 33). More specifically, Petitioner’s challenge turns on the fact that the trial court excluded 66 to 75 percent of the African-American individuals who were considered for jury service. (Id.). “Courts that have found [the] exclusion rate alone sufficient to make out a 'prima facie case [of discrimination], however, have done so where [the exclusion rate] was greater than sixty-six percent[.]” (Id. at 33-34).
This Court agrees that, in many cases where courts have found that a prosecutor’s exclusion rate—by itself—-supported a prim a facie case of discrimination, the exclusion rate was very high. See Jones v. West, 555 F.3d 90, 98 (2d Cir.2009) (collecting cases). Crucially, however, Petitioner has not relied solely on the exclusion rate to support his Batson argument. While Petitioner has placed nearly all of his emphasis on the exclusion rate, he has also suggested—from the time of trial up through this proceeding—that the prosecutor’s repeated use of strikes against African-American jurors was troubling because African-Americans comprised a small fraction of the venire. (See T. 181, 320, 323b, 420; see also Pet. Br. 1). Thus, this Court can consider whether the number of peremptory challenges exercised against African-American jurors was suspicious in light of the additional fact that there were so few in the jury pool. As explained above, the answer to this question is a resounding yes.
Second, the Report contends that, while “ ‘statistical disparities are to be examined,’ courts must also consider ‘any other relevant circumstances’” that could support or dispel an inference of discrimination. (Report 34 (quoting Butler v. Fischer, 345 Fed.Appx. 642, 644 (2d Cir.2009) (summary order)). In this case, the Report observes, defense counsel alerted the trial court to statistical disparities, but said he saw no other circumstances that were relevant to Petitioner’s Batson claims. (Id. at 34-35 (citing T. 323b-23c)). To be sure, it may have been unwise for counsel to say that there were no “other circumstances” indicating that the prosecutor was acting with a discriminatory motive. See supra at 21, n.5. But this statement did not undermine the argument that the statistical disparities, taken alone, were more than sufficient to support a prima facie case of discrimination,
Third, the Report notes, as one relevant consideration, that Petitioner is white. (Report 35). The Supreme Court has recognized that “[r]acial identity” between the defendant and the individuals who are excused from jury service can be one consideration that supports an inference of discrimination. Powers v. Ohio, 499 U.S. 400, 416, 111 S.Ct. 1364, 113 L.Ed.2d 411 (1991). But “to say that the race of the defendant may be relevant to discerning bias in some cases does not mean that it will be a factor in others, for race prejudice stems from various causes and may manifest itself in different forms.” Id. Consequently, even in cases where the defendant is white, the prosecutor’s severely disproportionate use of peremptory challenges against African-Americans can give rise to an inference of discrimination. See id. That is precisely what happened in this ease.
Finally, the Report suggests that it was reasonable for the Appellate Division to conclude that Petitioner had not established a prima facie case of discrimination because “two African-American jurors were seated on the jury.” (Report 35). Because two African-American individuals were seated, the racial composition of the jury reflected the racial composition of the venire: 17 percent of sworn jurors (and 12 *90percent of sworn jurors plus alternates) were black, while' 14 to 16 percent of qualified venire members were black.
To be clear, this is a compelling point. Ultimately, however, neither the'presence of African-American individuals on the jury nor the symmetry between the racial composition of the jury and the venire is dispositive. As noted earlier, a single improper strike would have violated Petitioner’s constitutional rights. What is more, the Second Circuit has recognized that “[a] prosecutor may not avoid the Batson obligation to provide race-neutral explanations for what appears to be a statistically significant pattern of racial peremptory challenges simply by forgoing the opportunity to use all of his challenges against minorities.” Alvarado, 923 F.2d at 266; see also Batson, 476 U.S. at 93, 106 S.Ct. 1712 (“[Tjotal or seriously disproportionate exclusion of Negroes from jury venires ... is itself such an unequal application of the law ... as to show intentional discrimination.” (emphasis added; internal citation and quotation marks omitted)); Miller-El v. Dretke, 545 U.S. 231, 240-41, 125 S.Ct. 2317, 162 L.Ed.2d 196 (2005) (finding that the prosecutor violated Batson by excusing a disproportionate number of black venire members, despite the fact that one of the seated jurors was black),
Similarly, under Batson and its progeny, a prosecutor may not mask or to cure the discriminatory use of peremptory challenges by seating just enough members of a cognizable group to achieve a racial balance between the jury and the venire. See Batson, 476 U.S. at 95, 106 S.Ct. 1712 (“A single invidiously discriminatory governmental act is not immunized by the absence of such discrimination in the making of other comparable decisions.” (internal "quotation marks omitted)); see also, e.g,, Rice v. White, 660 F.3d 242, 256 (6th Cir.2011) (“[Tjhe trial judge erroneously believed ... that Batson violations may be ‘cured’ by reference to the ultimate racial composition of the [petit jury.]”); Strickland v. State, 980 So.2d 908, 915 (Miss.2008) (“The Batson doctrine is not concerned with racial, gender, or ethnic balance on petit juries. ... Rather, it is concerned exclusively with discriminatory intent on the part of the lawyer against whose use of his peremptory strikes the objection is interposed.”); cf. United States v. Nelson, 277 F.3d 164, 207-08 (2d Cir.2002) (holding that it was impermissible for the trial court to balance the number of African-American and Jewish jurors in a racially-charged case). As a result, courts cannot bless any jury that happens to reflect the racial composition of the venire, even if other evidence suggests that the prosecutor struck one or more venirepeo-ple on account of them race.
Here, such other evidence was abundant: the prosecutor struck twice the number of black jurors than one would expect, and two-thirds to three-quarters of the black jurors under consideration. There is absolutely nothing in the record to explain this phenomenon. Confronted with this information, any reasonable court would have felt compelled, at the very least, to guess why the prosecutor behaved as he did. But the Batson framework was designed to avoid such “needless and imperfect speculation,” by asking prosecutors to perform the “simple” task of placing race-neutral reasons for peremptory strikes on the record. Johnson, 545 U.S. at 172, 125 S.Ct. 2410. As a result, this Court must conclude that the Appellate Division applied Batson and its progeny in an unreasonable manner.6
*91The Court appreciates the gravity of this holding, particularly in light of the Supreme Court’s admonition that AEDPA creates a “highly deferential standard for evaluating state-court rulings.” Cullen, 563 U.S. at 181, 131 S.Ct. 1388. Even in the Batson context, however, AEDPA deference has its limits. See, e.g., Jones, 555 F.3d at 101 (holding that New York courts unreasonably applied Batson). Enforcing those limits is the only way to preserve the few remaining teeth ■ in the Supreme Court’s Batson’s jurisprudence.
c. The Error in This Case Warrants Reversal
Ordinarily, when a federal habeas court determines that a state court applied clearly established federal law in an unreasonable way, the federal court must determine whether the state court’s error was harmless. See Nappi v. Yelich, 793 F.3d 246, 253 (2d Cir.2015); Fry v. Pliler, 551 U.S. 112, 121-22, 127 S.Ct. 2321, 168 L.Ed.2d 16, (2007). There is, however, a small class of state-court errors that are considered “structural defects.” United States v. Gonzalez-Lopez, 548 U.S. 140, 148-49, 126 S.Ct. 2557, 165 L.Ed.2d 409 (2006). These defects “defy analysis by harmless-error standards because they affect the framework within which the trial proceeds, and are not simply an error in the trial process itself.” Id. (internal quotation marks and brackets omitted).
This Court concludes, contrary to the Report (see Report 26), that Batson errors still qualify as structural defects. See Galarza v. Keane, 252 F.3d 630, 638 n. 8 (2d Cir.2001); Tankleff v. Senkowski, 135 F.3d 235, 248 (2d Cir.1998). In Vasquez v. Hillery, 474 U.S. 254, 263-64, 106 S.Ct. 617, 88 L.Ed.2d 598 (1986), the Supreme Court recognized that racial discrimination in the selection of grand jurors “undermines the structural integrity of the criminal tribunal itself, and is not amenable to harmless-error review.” Accord Arizona v. Fulminante, 499 U.S. 279, 294, 111 S.Ct. 1246, 113 L.Ed.2d 302 (1991); Gonzalez-Lopez, 548 U.S. at 149 n. 4, 126 S.Ct. 2557. And in Batson itself, the Supreme Court observed that “the basic principles prohibiting exclusion of persons from participation in jury service on account of their race are essentially the same for grand juries and for petit juries.” 476 U.S. at 84 n. 3, 106 S.Ct. 1712 (internal quotation marks and citation omitted). Thus, the logic of Vasquez applies with equal force in cases where defendants are alleging that the prosecutor used race-based criteria to select petit jurors.
Further, the Supreme Court’s recent decision in Davis v. Ayala, - U.S. -, 135 S.Ct. 2187, 192 L.Ed.2d 323 (2015), in no way detracts from the position that Batson errors are structural. In Davis, the prosecutor “used seven peremptories to strike all of the African-Americans and *92Hispanics who were available for service.” Davis, - U.S. -, 135 S.Ct. at 2193-94. At three different points during the jury selection process, the defendant raised Batson objections. Id. After each objection, the trial court allowed the prosecutor to provide race-neutral explanations for the relevant peremptory challenges outside the presence of defense counsel. Id. Ultimately, the trial court concluded that the prosecutor’s race-neutral explanations were credible. Id.
On appeal, the defendant argued that “the trial court committed reversible error by excluding the defense from part of the Batson hearing.” Davis, - U.S. -, 135 S.Ct. at 2195. The state supreme court acknowledged that “excluding the defense from a [Batson]-type hearing may amount to a denial of due process.” Id. However, the court also held that, if the trial court had committed “federal error, ... [it] was harmless beyond a reasonable doubt.... On the record before us, we are confident that the challenged jurors were excluded for proper, race-neutral reasons.” Id.
After the state supreme court denied his appeal, the defendant filed a petition for a writ of habeas corpus, and the Ninth Circuit granted relief. Davis, - U.S. -, 135 S.Ct. at 2196. The Supreme Court reversed, holding that the state supreme court had reasonably applied federal law when it determined that any federal constitutional error was harmless beyond a reasonable doubt. Id. at 2208. The Court explained that, surveying the record, a reasonable jurist could conclude that there was no significant chance that defense counsel could have said or done something to persuade the trial court that the prosecutor committed Batson error. See id. at 2201 (“Ayala contends that the presence of defense counsel might have made a difference because defense counsel might have been able to identify white jurors who were not stricken by the prosecution even though they had ‘expressed similar or greater hesitancy’ about the death penalty. We see no basis for this argument.”); id. at 2204 (“[N]either Ayala nor the Ninth Circuit has identified anything that defense counsel might have done at the ex parte hearing to show that the prosecutor’s concern about [a juror’s] limited English proficiency was pretextual.”); id. at 2205 (“Nor is there a basis for finding that the absence of defense counsel affected the trial judge’s evaluation of the sincerity of [the prosecutor’s] proffered ground for the strike.”); id. (“[Tjhere is no support for the suggestion that Ayala’s attorney, if allowed to attend the ex parte hearing, would have been able to convince the judge that this reason was pretextual.”); id. at 2206 (“That Ayala’s attorney did not have the opportunity to repeat [an] argument once more at the in camera proceeding does not create grave doubt about whether the trial court would have decided the [Batson] issue differently.”). As a result, the Court concluded, a reasonable jurist could find that excluding defense counsel from part of the Batson hearing was harmless. See id. at 2208.
Significantly, however, Davis was fundamentally a case about the presence of counsel. The Supreme Court held that excluding defense counsel from a portion of a Batson hearing could reasonably be considered harmless because there was no real chance that defense counsel could have identified any Batson error. The Supreme Court never suggested that, had any Batson error occurred, the error would have been harmless.
Here, both the trial court and the Appellate Division contravened Batson by holding that, when Petitioner raised his fourth and final Batson challenge at the end of jury selection, there was no need to proceed to step two of the Batson inquiry. *93Because Batson error is structural, this Court must remand the case to allow the trial court to: (i) “hold a reconstruction hearing and take evidence regarding the circumstances surrounding the prosecutor’s use of the peremptory challenges” so that the court can determine whether, at the third step of the Batson inquiry, it would have found Batson error; or (ii) hold a new trial. Jones, 555 F.3d at 102 (internal quotation marks omitted).7
CONCLUSION
For the foregoing reasons, the petition for a writ of habeas corpus is granted, and this case is remanded to the Supreme Court of New York, New York County, for proceedings consistent with this Order. The Clerk of Court is directed to terminate all pending motions, adjourn all remaining dates, and close this case.
SO ORDERED.
. In fact, after the trial court and the Appellate Division rendered their decisions in this case, the New York Court of Appeals effectively demurred when given an opportunity to place the Second Circuit's saving gloss on the Brown I opinion. See People v. Hecker, 15 N.Y.3d 625, 651, 917 N.Y.S.2d 39, 942 N.E.2d 248 (2010) (citing Brown I for the broad proposition that "purely numerical or statistical arguments are ‘rarely conclusive in the absence of other facts or circumstances’ to give rise to an inference of discrimination”).
. These individuals included the 26 prospective jurors on the first panel, the 26 prospective jurors on the second panel, and the first 16 prospective jurors on the third panel.
. The parties dispute whether one of the jurors removed by defense counsel—Ms. Ve-larde—was African-American. (Compare Pet, Br. 6, with Resp. Br. 29). Ultimately, this Court believes that it is not necessary to determine Ms. Velarde's race in order to resolve this Petition.
. To the contrary, many of the stricken African-American jurors had ties to the law enforcement community. For example, the prosecutor struck: (i) Ms. Boiken, whose aunt was a lieutenant in the New .York City Police Department (T. 107-08, 180); (ii) Ms. Hamilton, who said her "best friend” and her best friend’s parents worked as corrections officers (id. at 101-02, 180); (iii) Ms. Grant, who noted that three friends and a brother-in-law worked as corrections officers (id. at 248; 320); and (iv) Ms. Simmons, whose uncle was a retired FBI agent (id. at 250, 323b). However, because defense counsel did not bring these facts to the attention of the trial judge, this Court has not considered them in its analysis.
. Alternatively, Petitioner contends that he has rebutted the Appellate Division’s “factual determination that no prima facie case of discrimination had been established under Bat-*91son.” (Pet. Br. 22). Contrary to Petitioner's suggestion, the presence or absence of a pri-ma facie case of discrimination is not a purely factual question; rather, it is a "mixed question of law and fact.” Sorto v. Herbert, 497 F.3d 163, 171 (2d Cir.2007); accord United States v. Martinez, 621 F.3d 101, 109-10 (2d Cir.2010). Thus, when a federal habeas court reviews a state court’s decision that a defendant failed to establish a prima facie case of Batson discrimination, the habéas court should only ask whether the state court discrimination was "contrary to, or involved an unreasonable application of, clearly established federal law.” Sorto, 497 F.3d at 171.
However, if a court progresses to step three of the Batson inquiry, the court must make a purely factual finding regarding the motive of the attorney who struck a disproportionate number of minority jurors. See Miller-El v. Cockrell, 537 U.S. 322, 340, 123 S.Ct 1029, 154 L.Ed.2d 931 (2003). On habeas review, such a factual finding may be rebutted by clear and convincing evidence. See id.
. Given the Court’s resolution of Petitioner’s first argument, it declines to address, as moot, Petitioner’s claims of ineffective representation by trial counsel.