dissenting.
No right is more fundamental to the notion of a fair trial than the right to an impartial judge. Johnson v. Mississippi, 403 U.S. 212, 216, 91 S.Ct. 1778, 1780, 29 L.Ed.2d 423 (1971) (per curiam); In re Murchison, 349 U.S. 133, 136, 75 S.Ct. 623, 625, 99 L.Ed. 942 (1955); see also Aetna Life Ins. Co. v. Lavoie, 475 U.S. 813, 106 S.Ct. 1580, 89 L.Ed.2d 823 (1986); Ward v. Village of Monroeville, Ohio, 409 U.S. 57, 93 S.Ct. 80, 34 L.Ed.2d 267 (1972); Tumey v. Ohio, 273 U.S. 510, 47 S.Ct. 437, 71 L.Ed. 749 (1927). “The truth pronounced by Justinian more than a thousand years ago, that ‘[ijmpartiality is the life of justice,’ is just as valid today as it was then.” United States v. Brown, 539 F.2d 467, 469 (5th Cir.1976) (per curiam). The constitutions of our nation and of our states, the rules of evidence and of procedure, and 200 years of case law promise a full panoply of rights to the accused. But ultimately the guarantee of these rights is no stronger than the integrity and fairness of the judge to whom the trial is entrusted.
The State of Illinois placed the fate of William Bracy and Roger Collins in the hands of a racketeer. Thomas Maloney is presently serving a prison term of nearly sixteen years for racketeering, conspiracy to commit racketeering, extortion under color of official right, and obstruction of justice. A jury determined that Maloney had accepted $10,000 in 1986 to acquit two El Rukn leaders of a double murder, an undetermined portion of a $100,000 payment in 1981 to acquit three New York gang members of murdering a rival in Chicago’s Chinatown, and between $4,000 and $5,000 in 1982 to convict another individual of voluntary manslaughter rather than felony murder. These were but three of the bribes that witnesses attributed to Maloney at his trial. See United States v. Maloney, 71 F.3d 645 (7th Cir.1995); Matt O’Connor, Judge Maloney found guilty in corruption case, CHICAGO Tribune, April 17, 1993, News, at 1. At sentencing, Judge Leinenweber also found that Maloney had, while still a practicing lawyer, cooperated in procuring the notorious acquittal of reputed mob. hitman Harry Aleman by Judge Frank Wilson in 1977. Mob enforcer Michael Bertucci testified that Maloney had helped him to evade a series of criminal charges by bribing judges as far back as the late 1960s, although Judge Leinenweber discounted this testimony. It would seem, in any event, that by the time Maloney ascended to the bench in 1977, he was well groomed in the art of judicial corruption, an art that he would practice at least until 1986, when he correctly perceived that he was under the watchful eye of the FBI and returned the $10,000 bribe he had accepted in the El Rukn prosecution.
Bracy and Collins were tried before Malo-ney in 1981, in the midst of Maloney’s bribe taking. Maloney did not solicit a bribe from either defendant, nor did the defendants offer him one. Nor did the prosecution bribe Maloney. But given the abundant proof (and a federal jury’s finding) that justice was for sale in Maloney’s courtroom, we are compelled to consider whether Maloney may be deemed the impartial judge to which due process entitled these defendants.
1.
The petitioners argue that in cases that were not fixed, Maloney had an incentive to be particularly tough on defendants, in order to divert suspicion that might otherwise be aroused by the acquittals he was paid to render and to strengthen the incentive for defendants to bribe him. Without conceding that further evidence of Maloney’s partiality was required to establish Maloney’s constitutional inadequacy as a judge, the petitioners sought leave from the district court to engage in discovery, with the aim of establishing a pattern of corruption that affected Ma-loney’s conduct in not only those cases in which he had accepted a bribe but also those *697in which he had not. Judge Hart denied their request, deeming what the petitioners sought to prove to be a matter of speculation and, in any event, insufficient to establish a constitutional deprivation. United States ex rel. Collins v. Welbom, 868 F.Supp. 950, 991 (N.D.Ill.1994). Like my colleagues in the majority, the district judge concluded that the most the circumstances permitted Bracy and Collins to argue was the mere possibility of bias, which Del Vecchio v. Illinois Dep’t of Corrections, 31 F.3d 1363 (7th Cir.1994) (en banc), cert. denied, — U.S.-, 115 S.Ct. 1404, 131 L.Ed.2d 290 (1995), indicates is not enough to establish a deprivation of due process. 868 F.Supp. at 991.
In fact, the notion that Maloney was deliberately tough on defendants who did not bribe him finds support in the testimony presented at Maloney’s trial. Defense attorney William Swano arranged several of the bribes for which Maloney was prosecuted and was a key government witness against him. In 1985, .Swano represented James Davis, whom the state had charged with armed robbery. The case was assigned to Maloney for trial. By this time, Swano had already bribed Maloney on a number of occasions. But after investigating the prosecution’s ease against Davis, Swano concluded that it would be unnecessary to bribe Malo-ney in order to obtain an acquittal in this case: three witnesses to the robbery knew the two perpetrators and said that Davis was not one of them; Davis had an alibi; and the victim, of the crime, who had initially identified Davis as one of the perpetrators, had confessed uncertainty about the identification. Swano was confident that “[t]he case was a not guilty in any courtroom in the building.” United States v. Thomas J. Maloney and Robert McGee, No. 91 CR 477, 1994 WL 96673 Tr. 2528 (N.D.Ill. March 24,1993). To Swano’s surprise, however, Maloney convicted his client after a bench trial. Swano took this as a lesson that “to practice in front of Judge Maloney ... we had to pay.” Tr. 2530. That Swano had correctly interpreted this conviction as a lesson was arguably confirmed by Maloney’s bagman, Robert McGee. Swano met with McGee soon after Davis was convicted to discuss a fix in the double murder trial of the two El Rukns. Swano had persuaded his client that bribing Maloney was the prudent course, explaining that he had only lost one case before Maloney, “and that was the one that we didn’t work,” ie., fix. Tr. 2544; see also Tr. 2559. In arranging the meeting with McGee, Swano had told him that he wanted to discuss “a hot case in front of Judge Maloney”. Tr. 2566. McGee said that he would first have to obtain Malo-ney’s permission. Tr. 2567. When Swano and McGee subsequently met at Le Bour-deux, a local watering hole, McGee told Swa-no that he had gotten the okay from Maloney to speak about the matter. Swano recalled: “He told me that the judge had said we could talk, especially in view of the way the judge had screwed me on the last case,” which Swano understood to be a reference to the Davis case. Tr. 2567-68. Swano voiced agreement with the assessment “that the judge had screwed me on the case and had screwed my client.” Tr. 2568. He and McGee then got down to details about the El Rukn fix. That bribe was one of four that the jury later found Maloney guilty of accepting.
One may infer from Swano’s testimony that Maloney saw the Davis prosecution, in which no bribe was tendered, as an opportunity to teach Swano a lesson that would ensure bribes in future cases. That, at least, was the moral of the story for Swano. If Swano was right (a matter for the factfinder, not us, to determine), then it would seem that Maloney’s approach to case fixing was indeed the global view that Bracy and Collins posit: fixed cases were a source of illicit profit, whereas unfixed cases were an opportunity, as Bracy puts it, to “advertise” in the defense bar (Bracy Reply at 1) while at the same time protecting his franchise by currying favor with law and order minded voters and avoiding the ire of the law enforcement community. Like my colleagues, I think that the petitioners face an exceedingly difficult task in attempting to unearth evidence that will lend further support to their theory (see ante at 691), but Swano’s testimony suggests that the search may not be futile.
At bottom, my colleagues believe that Bra-cy and Collins are not entitled to discovery because the most they can hope, to prove is *698that Maloney made it a practice to lean over backwards in favor of the prosecution in cases in which he was not bribed; without proof that he followed that practice in this case, they reason, Bracy and Collins have no claim. Ante at 691. But when the trial judge is tainted by a pervasive conflict of interest — in other words, one not limited to a particular litigant or type of case-evidence that the taint had a discernible effect on a given case is unnecessary. Here the showing that my colleagues require would be all but impossible to make, absent either an extraordinary admission from Maloney, which is not forthcoming (Maloney continues to proclaim his innocence) or the kind of over-the-top courtroom behavior that makes a judge’s partiality plain (see, e.g., United States v. Dellinger, 472 F.2d 340, 386-88 (7th Cir.1972), cert. denied, 410 U.S. 970, 93 S.Ct. 1443, 35 L.Ed.2d 706 (1973)), a rare phenomenon not evident from the record here. In any event, the Supreme Court has expressly rejected the idea that such proof is mandated, finding that it “requires too much and protects too little.” Ward, 409 U.S. at 61, 98 S.Ct. at 83; see also Aetna, 475 U.S. at 830-31, 106 S.Ct. at 1590 (Brennan, J., concurring); id. at 831-33, 106 S.Ct. at 1590-91 (Blackmun, J., concurring). To establish a deprivation of due process, the petitioners need only show that the circumstances “would offer a possible temptation to the average man as a judge to forget the burden of proof required to convict the defendant, or which might lead him not to hold the balance nice, clear, and true between the state and the accused.” Tumey, 273 U.S. at 532, 47 S.Ct. at 444; accord Aetna, 475 U.S. at 822, 106 S.Ct. at 1585; Ward, 409 U.S. at 60, 93 S.Ct. at 83; Del Vecchio, 31 F.3d at 1372, 1373. Proof that Maloney was motivated by virtue of his bribe taking to favor the prosecution and disfavor the defense in unfixed cases would permit, if not compel, the inference that Maloney’s adverse rulings against Bracy and Collins were animated by a pernicious intent on Maloney’s part (see Fed. R.Evid. 404(b)) and would more than satisfy the standard the Supreme Court has enunciated.
Habeas Rule 6(a) requires only that the petitioner demonstrate “good cause” to engage in discovery, and Bracy and Collins have certainly done so here. In view of the jury’s verdict against Maloney, it is undisputed that he was accepting bribes at the very time that Bracy and Collins came to trial. Although there is no suggestion that money oiled the wheels of justice in this case, my colleagues concede the plausibility of the notion that “a judge’s corruption is likely to permeate his judicial conduct rather than be encapsulated in the particular cases in which he takes bribes.” Ante at 689. Putting aside for the moment contamination of the judge’s philosophy, it seems to me likely that any judge who accepts bribes and wishes to remain on the bench will think in strategic terms about his other eases. As Chief Judge Posner points out, one might wonder whether it would really have been in Maloney’s interest to assume a pro-prosecution mantle in unfixed cases, lest the occasional acquittal purchased from him look out of character. Ante at 689-90. But the time for such deliberation is after an evidentiary hearing on the matter, when the petitioners have had the opportunity to find and present whatever evidence there may be to establish any practice Maloney may have followed. Swano’s testimony provides some evidence in that regard, and there may be more. Without giving Bracy and Collins the opportunity to present that evidence to Judge Hart, their claim of conflict' can only be resolved on the basis of speculation, as my colleagues agree. Ante at 690. In view of the grave and structural nature of the petitioners’ claim, not to mention the fact that this is a capital case, which “magnifies the appearance of impropriety,” (ante at 689), the petitioners are entitled to more from us.
My colleagues note that Bracy and Collins have long had access to some of the information that they profess an interest in exploring — the records of other trials over which Maloney presided, and the record of Malo-ney’s own trial, for example — but have not pointed to anything that bolsters their claim of partiality. Ante at 691. But if petitioners can be,faulted for not making the most of the available material, we can be faulted for being naive about what the cold page of a trial record will reveal. A judge who wishes to be *699tough on the defendant need not adopt the manner of the Tasmanian Devil to do it. Maloney was by no account stupid. When he sold an acquittal, he wanted facts that he could hang his hat on (e.g., Maloney Tr. 2571, 2669-70, 2682); and we have no reason to doubt that if he wanted to cultivate a pro-prosecution record to protect his interests as a bribe taker, he had the ability to do so discretely, without appearing to have abused his discretion as a trial judge. Cases are fixed not in the courtroom but in bars, bathrooms, and back hallways. If there is evidence of the kind Bracy and Collins hope to find, it is in the hands of persons familiar with these venues of injustice. Both Malo-ney and his bagman McGee are continuing their version of “stand[ing] tall” (see United States v. Maloney, 71 F.3d at 651-52), but Swano, Robert Cooley, and others may have something material to say, and the U.S. Attorney may be of some help in identifying whom the petitioners should approach. But it is likely that no one is going to talk without a subpoena, and petitioners should not be deprived of that instrument.
We are venturing into a realm noir with which, I may say with confidence, none of us is on intimate terms. We cannot simply assume that “the probability is slight” that discovery will yield Bracy and Collins anything. Ante at 691. Let them try. If their discovery proves fruitless, we can at least take comfort in the knowledge that we have given them every opportunity to prove that Maloney’s corruption deprived them of a fair trial. We cannot, after all, have it both ways: we cannot criticize Bracy and Collins for speculation and at the same time deprive them of the chance to render their theory anything more. I understand that Illinois has an interest in the finality of its judgments, and allowing the discovery that the petitioners seek would, if nothing else, portend a significant delay in the implementation of their death sentences. But having left Bracy and Collins to the mercies of a corrupt judge, the State should not be heard to complain in this matter. (In fact, its brief is utterly silent on this point.) The people of Illinois have as great an interest in the integrity of capital trials as Bracy and Collins do.
2.
My disagreement with the majority goes deeper than the question of discovery, however. In the end, I agree with Collins and Bracy that proof of the impact Maloney’s corruption had, or probably had, on the petitioner’s trial is unnecessary. We do not know, and we likely will never know, what Maloney thought about Bracy and Collins. But we have a pretty clear picture of how he viewed justice. The price tag may have varied, but as Maloney’s conviction proves, justice was for sale in Maloney’s courtroom to the defendants who could afford to pay, even when they were charged with the most heinous of crimes. That fact carries far more significance than the majority is willing to recognize. As this court acknowledged in Del Vecchio, in considering whether a biasing influence requires the disqualification of a judge, “we begin ... by presuming ‘the honesty and integrity of those serving as adjudicators.’ ” 31 F.3d at 1375 (quoting Withrow v. Larkin, 421 U.S. 35, 47, 95 S.Ct. 1456, 1464, 43 L.Ed.2d 712 (1975)). That presumption, as my colleagues acknowledge, is “obviously inapplicable here.” Ante at 688. Ma-loney was not Louis Garippo, an esteemed and honest judge whose impartiality was argued to have been potentially compromised by his prior involvement with the defendant as a prosecutor. See Del Vecchio, 31 F.3d at 1375-80 (majority); id. at 1398 (Cummings, J., dissenting); id. at 1399 (Cudahy, J., dissenting); id. (Ripple, J., dissenting). Nor was he even Otto Kerner, a judge whose crimes pre-dated his service on the bench. See United States v. Isaacs, 493 F.2d 1124 (7th Cir.), cert. denied, 417 U.S. 976, 94 S.Ct. 3183, 41 L.Ed.2d 1146 and cert. denied, 417 U.S. 976, 94 S.Ct. 3184, 41 L.Ed.2d 1146 (1974). Maloney was a criminal who, as a judge, transformed his very office into a racketeering enterprise.
Faced with the possibility of vacating numerous convictions obtained in Maloney’s courtroom, we would like to believe that he was capable of the impartiality that the Fourteenth Amendment requires when no money changed hands. But how can we? Once he embarked on the path of bribe taking, Maloney had forsaken his judicial oath. *700Justice was a mere commodity to him, defendants nothing more than a profit center. His deviation from the path of righteousness was not, moreover, momentary and uncharacteristic; it was cold, calculated, and spanned a period of years, if not the entirety of his tenure on the bench. Thus, Maloney’s “bias,” if we can call it that, cannot be conveniently compartmentalized. Cf. Diversified Numismatics, Inc. v. City of Orlando, Florida, 949 F.2d 382, 385 (11th Cir.1991) (per curiam). We may no more' treat Maloney as an impartial arbiter for constitutional purposes than a delusional megalomaniac who locks a judge in the closet, dons a black robe, and hoodwinks everyone with a credible impersonation of Oliver Wendell Holmes. Ma-loney’s willingness to exchange money for the freedom of those charged with the most abhorrent crimes displayed his scorn for the very concept of justice.
By demanding proof that Maloney’s corruption either had or likely had an identifiable impact on the petitioners’ trial, we fail to come to grips both with the gravity of Malo-ney’s offense and with the constitutional imperative that the accused be tried before a judge of integrity and impartiality. “[T]o perform its high function in the best way, ‘justice must satisfy the appearance of justice.’ ” Aetna, 475 U.S. at 825, 106 S. Ct at 1587 (quoting Murchison, 349 U.S. at 136, 75 S.Ct. at 625). We said in Del Vecchio that “bad appearances alone [do not] require disqualification” (31 F.3d at 1372), but that an external influence requires the judge’s disqualification when “the influenee[ ] involved str[ikes] at the heart of human motivation, that an average man would find it difficult, if not impossible, to set the influence aside” (id. at 1373). My colleagues and I have been discussing Maloney’s bribetaking as just another “bias” or “influence,” something external to his personality, or at lease some sever-able part of it, that at most “might” have given him the “incentive” to behave in a particular fashion on occasions when he was not bribed. But there is a more malevolent side to Maloney’s offense that cannot be ignored. Maloney’s bribetaking removes him from the category of the “average” man we addressed in Del Vecchio. We are speaking in this case about what motivates a criminal, and this implicates a far darker set of impulses than we confront in the usual bias case. The question we should be asking ourselves is not what impact the lack of a bribe had on Maloney’s decisionmaking in a particular case, but what his willingness to accept a bribe tells us about his view of judging. Maloney proved himself willing to acquit defendants charged with capital offenses for a few thousand dollars. The victims of those crimes, their families, the people of Illinois, the concept of justice, were apparently worth no more to him. Why should we assume that defendants were worth anything more to Ma-loney? How, in particular, can we trust Ma-loney to have treated a defendant fairly when that defendant had not offered him any money? 1 If due process means anything, I think we must assume that Maloney’s corruption pervaded his work as a judge. The Supreme Court could not have put it more clearly: “[W]hen the trial judge is discovered to have had some basis for rendering a biased judgment, his actual motivations are hidden from review, and we must presume that the process was impaired.” Vasquez v. Hillery, 474 U.S. 254, 263, 106 S.Ct. 617, 623, 88 L.Ed.2d 598 (1986).2
*701Although Maloney’s crimes reveal no fealty to his oath as a judge, my coEeagues nonetheless refuse to relinquish the presumption that he acted fairly when not bribed. The notion that “a judge who accepts bribes in some eases is corrupt in all” is not “a sufficiently compelling empirical proposition,” they say, to treat this case as if the government had bribed Maloney to convict the petitioners. Ante at 690. But this is not an empirical matter. We cannot assign a value of a; to a judge’s abiEty to be fair, divide it by y (y representing Maloney’s bribe taking), and determine whether the result is less than the constitutionaEy minimal level of impar-tiaHty, z. Like so many other elements of our democracy, justice requires a leap of faith: faith that the defendant is in fact presumed innocent until proven guilty beyond a reasonable doubt; faith that the prosecutor and defense counsel alike wiE act as zealous advocates for their principals within the confínes of law and ethics; faith that the trial judge wiE favor no party but wiE strive “to hold the balance nice, clear, and true between the state and the accused.” Turney, 273 U.S. at 532, 47 S.Ct. at 444. Maloney’s crimes shatter that faith. We have no reason to beEeve that had Bracy and Collins possessed sufficient funds and willing attorneys, they could not have bought acquittals from Maloney. That alone suggests that Maloney’s was not the court of “law” to which Bracy and Collins were entitled as the forum for their trial. We do have reason, based on Swano’s testimony, to beEeve that Maloney’s corruption extended beyond the cases in which he accepted a bribe, and that Maloney saw unfixed eases as an opportunity to “screw” the defendant and thereby further his own ends as a bribe taker. No “empirical” proof is necessary to demonstrate that the petitioners did not stand equal before the law in Maloney’s courtroom.
I realize, of course, that Bracy and Collins were convicted by a jury, not by Maloney. Jurors have minds of their own; they can and do defy the expectations of the judge. That is but one reason that the jury has been viewed by some as “the very paEadium of free government.” The Federalist No. 83, at 499 (Alexander Hamilton) (Clinton Rossiter ed. 1961); see Duncan v. Louisiana, 391 U.S. 145, 155-58, 88 S.Ct. 1444, 1451-52, 20 L.Ed.2d 491 (1968). But we cannot ignore the influence that the judge retains even in a jury trial. See Walker v. Lockhart, 726 F.2d 1238, 1259 (8th Cir.1984) (en banc) (Bright, J., dissenting), cert. dismissed, 468 U.S. 1222, 105 S.Ct. 17, 82 L.Ed.2d 912 (1984), and cert. denied, 478 U.S. 1020, 106 S.Ct. 3332, 92 L.Ed.2d 738 (1986). I do not refer so much to the abiEty of the judge to communicate his opinions to the jury through raised eyebrows, choice bits of sarcasm, and questioning of the witnesses that strays into advocacy, although this happens. E.g., Dellinger, 472 F.2d at 386-88; see also United States v. Filani, 74 F.3d 378, 387 (2d Cir.1996); Bufford v. Rowan Cos., 994 F.2d 155, 159 (5th Cir.1993). I mean the extraordinary abiEty of the trial judge to shape the trial itself. It is she who decides what evidence the jury may hear, how counsel may behave in front of the jury, what arguments may be made, how they may be made, what legal principles the jury must apply, and even, to a significant degree, who wül sit on the jury. Thus, Oven when the verdict is not entrusted to her, a partial judge retains great influence, if not directly upon the jury, then upon the myriad events that culminate in the jury’s decision. See Tyson v. Trigg, 50 F.3d 436, 439 (7th Cir. 1995), cert. denied, — U.S.-, 116 S.Ct. 697, 133 L.Ed.2d 655 (1996).3
Our own abiEty to monitor the influence of the trial judge, and to discern the taint of *702partiality, is narrowly circumscribed. Appellate review of the mundane decisions that can make or break a party’s case (the exclusion or admission of a key piece of evidence, for example) typically is quite limited. E.g., United States v. Marshall, 75 F.3d 1097, 1109 (7th Cir.1996) (evidentiary issues); United States v. Pulido, 69 F.3d 192, 204 (7th Cir.1995) (limitations on cross-examination); United States v. Fish, 34 F.3d 488, 495 (7th Cir.1994) (request for continuance); United States v. $94,000.00 in U.S. Currency, 2 F.3d 778, 788 (7th Cir.1993) (questions posed on voir dire). In the rare instance that we find error, it is more often than not deemed harmless. E.g., Jones v. Page, 76 F.3d 831, 855-56 (7th Cir.1996). Even errors of constitutional dimension may be labelled benign on review. Tyson, 50 F.3d at 446-47. Theoretically, we could require a more searching review of the record when there is evidence that the trial judge was corrupt, but we would be fooling ourselves to think that even our best efforts would suffice to expose the ways in which the judge’s criminal behavior may have tainted the trial. See United States v. Guglielmini, 384 F.2d 602, 605 (2d Cir.1967) (“Few claims are more difficult to resolve than the claim that the trial judge, presiding over a jury trial, has thrown his weight in favor of one side to such an extent that it cannot be said that the trial has been a fair one.”), cert. denied, 400 U.S. 820, 91 S.Ct. 38, 27 L.Ed.2d 48 (1970). So much goes on in the courtroom that the written record can never reveal. Why else do we routinely grant so much deference to the trial judge, who sees and hears the witnesses first hand, who supervises the trial from start to finish, who can best gauge the impact of any development upon the jury before him? Our acquiescence in the decisions of the trial court is dictated as much by pragmatism as by principle.
It is no answer to the charge of corruption that Maloney’s discretionary rulings on their face appear to fall within the realm of reason. See ante at 690. All that means is that a reasonable judge might have rendered the same rulings. But we assume that the reasonable judge does not act for malignant ends, that she exercises
a sound judicial discretion, enlightened by intelligence and learning, controlled by sound principles of law, of firm courage combined with the calmness of a cool mind, free from partiality, not swayed by sympathy nor warped by prejudice nor moved by any kind of influence save alone the overwhelming passion to do that which is just.
Davis v. Boston Elevated Ry. Co., 235 Mass. 482, 126 N.E. 841, 844 (1920); see also Henry J. Friendly, Indiscretion About Discretion, 31 EMORY L. J. 747, 784 (1982) (“discretionary choices are not left to a court’s ‘inclination, but to its judgment; and its judgment is to be guided by sound legal principles” ’ ”) (quoting United States v. Burr, 25 F. Cas. 30, 35 (C.C.Va.1807) (No. 14,692d) (Marshall, C.J.)). If, on the other hand, a judge exercises her discretion for invidious reasons, she has exceeded her authority. See 1 Steven. Alan Childress and Martha S. Davis, FEDERAL Standards Of Review, § 4.01 at 4-2 through 4 — 3 (2d ed. 1991); Wong Wing Hang v. INS, 360 F.2d 715, 719 (2d Cir.1966) (Friendly, J.). But, except in the rare case in which the judge’s agenda is obvious, we cannot expect to autopsy a trial and find evidence that the cancer of the judge’s corruption has invaded her decisionmaking. When the majority finds it “unlikely” that the discretionary trial rulings of which Bracy and Collins complain were the product of Maloney’s corruption (ante at 690), it is doing exactly what the petitioners are faulted for doing: speculating.
We cannot, therefore, hide behind the jury’s verdict. We simply cannot know what impact Maloney’s corruption may have had on the trial and on the jury’s decision. Cf. Vasquez, 474 U.S. at 260-64, 106 S.Ct. at 622-24 (discrimination in grand jury not rendered harmless by subsequent trial). The evidence against Bracy and Collins may seem *703to ,us overwhelming, but that does not strip them of their right to due process. “No matter what the evidence was against [them], [they] had the right to have an impartial judge.” Tumey, 273 U.S. at 535, 47 S.Ct. at 445. Nor, as a matter of principle, can the process of presumably unbiased appellate review cure the taint of Maloney’s corruption;-Bracy and Collins were “entitled to a neutral and detached judge in the first instance.” Ward, 409 U.S. at 61-62, 93 S.Ct. at 84.
Ultimately, although my colleagues concede the plausibility of the notion that a judge’s corruption cannot be confined to the cases in which he accepts a bribe, they decline to embrace it, finding the consequences “unacceptable.” Ante at 689. Maloney alone presided over some 6,000 cases. See United States v. Thomas J. Maloney and Robert McGee, No. 91 CR 477, 1994 WL 96673, Sentencing Tr. 571 (N.D.Ill. July 21, 1994). My colleagues believe, and perhaps rightly so, that we cannot vacate the convictions of Bracy and Collins without calling into doubt the many other judgments entered not only by Maloney (the only Illinois judge thus far convicted of corruption in murder prosecutions), but also the seventeen other Cook County judges found guilty of taking bribes. Ante at 689. Given the expanse of time that has passed since these judgments were entered, were retrials ordered across the board, there are doubtless many guilty individuals, murderers even, who would go free. It is an appalling prospect. But we must not allow ourselves to become paralyzed by the possibilities. We decide today not the validity of every judgment ever rendered by a corrupt judge, but the fate of two individuals who are about to pay the ultimate criminal penalty without having been afforded the most basic rudiment of due process. What are we to say to Bracy and Collins, that they had the right to an honest, impartial judge but that the breadth of past corruption in the Illinois judiciary makes it too costly for us to enforce that right? Are they to become the latest victims of Maloney’s bribe taking, and we his accomplices after the fact? The Constitution was not written for easy cases and likeable defendants, and we are sworn to uphold it no matter what the result. Knowing full well the perils that may confront us if we insist that defendants be given trials before honest judges, I believe we have no choice but to take the first step down that path here. We cannot turn our backs on the Constitution, especially when the petitioners’ very lives are at stake. If nothing else, “[d]eath is factually different. Death is final. Death is irremediable. Death is unknowable.” Anthony G. Amsterdam, Tr. of oral argument, Gregg v. Georgia, 428 U.S. 153, 96 S.Ct. 2909, 49 L.Ed.2d 859 (1976) (No. 74-6257), quoted in May It Please The Court ... 233 (Peter Irons & Stephanie Guitton eds. 1993); see also Furman v. Georgia, 408 U.S. 238, 306, 92 S.Ct. 2726, 2760, 33 L.Ed.2d 346 (1972) (Stewart, J., concurring).
3.
The quality of justice we can claim to have achieved in this nation is not measured by what our best judges can.do but by what the worst, of our judges have done. Today we say that Thomas Maloney’s handiwork is good enough.. An Illinois defendant, it appears, is entitled to appear before a judge who is not under indictment for bribery, Ill. S.Ct. R. 56(a)(1), but today’s opinion deprives him of the right to appear before a judge who is not engaged in bribery. Bracy and Collins will thus have to be content with the judgment of a criminal. I do not know which I find more shocking: the base quality of justice that Bracy and Collins received in the Illinois courts, or our holding today that the Constitution requires no more.
4.
■ I must, finally, say a word about the majority’s invocation of Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989). Ante at 689. Teague generally bars the application of “new rules” of federal law on habeas review. But Teague’s proscription is not jurisdictional, Collins v. Youngblood, 497 U.S. 37, 41, 110 S.Ct. 2715, 2718, 111 L.Ed.2d 30 (1990), and if the state fails to invoke Teague, we may, but are not compelled, to do so ourselves, Goeke v. Branch, — U.S.-,-, 115 S.Ct. 1275, 1276, 131 L.Ed.2d 152 (1995) (per curiam). See Stewart v. Lane, 60 F.3d 296, 304-05 (7th Cir.*7041995) (Ripple, J., concurring), supplemented, on reh’g, 70 F.3d 955, petition for cert. filed (Jan. 16, 1996) (No. 95-7444). The state has never cited Teague as a defense to the petitioners’ claim of judicial corruption, and I am not convinced that the circumstances of this case warrant our sua sponte reliance upon it as an alternate basis for denying the petitioners the relief they seek. See Stewart at 304 (concurrence) (“Because, in a capital case, invocation of Teague can often mean the difference between life and death for the petitioner, we need to be particularly circumspect as to when we shall invoke Teague sua sponte.”). The majority offers no reason why Teague is particularly apposite here, and I discern none. Since when is it news that the accused has the right to be tried before an honest, impartial judge? I had rather thought that to be a cornerstone of our system of justice. And indeed, Supreme Court precedent reveals the notion to be anything but novel. The Court’s 1927 decision in Tu-rney, for example, holds that when a judge has a financial incentive to see the defendant convicted, he is not sufficiently impartial for constitutional purposes. 273 U.S. at 531-35, 47 S.Ct. at 444-45. This case is, in a real sense, but a factual variant of Turney. I grant that no court has yet found it necessary to hold that a judge engaged in serial bribetaking is not the impartial adjudicator that the Constitution requires, and in that respect one might argue that the rule the petitioners posit “was not dictated by precedent existing at the time [their] convietion[s] became final.” Caspari v. Bohlen, —— U.S. -, -, 114 S.Ct. 948, 953, 127 L.Ed.2d 236 (1994) (quoting Teague, 489 U.S. at 301, 109 S.Ct. at 1070) (emphasis in Teague). But surely common sense counts for something in the Teague analysis. The Greylord prosecutions had not yet taken place in 1981 when Bracy and Collins were tried, but the State of Illinois cannot claim to have been ignorant of the notion that bribery is illegal and that judges who accept bribes belong in prison, not on the bench. There is, in short, nothing surprising in the petitioners’ claim. The prospect of retrying Bracy and Collins (not to mention other defendants convicted by or before Maloney) is an onerous one for the State, but that burden has nothing to do with the novelty of the principle that a defendant is entitled to a judge who is not on the take. Our invocation of Teague in this circumstance makes that precedent look less like a shield protecting the State from the retroactive application of new rules than a sword depriving habeas petitioners of constitutional rights that have long been recognized.
5.
Eighteen judges of the Cook County Circuit Court have been convicted of corruption in the last decade. We would like to think that rampant corruption on the Cook County bench is a relic of the past. But it will not be, it cannot be, so long as we refuse to recognize just how fundamentally at odds this corruption is with the constitutional guarantee of due process. Like Terrence Hake, who risked his own career to expose the criminals clothed in the robes of judges, we too have a role to play in restoring integrity to the bench. We cannot embrace the judicial services of outlaws without deepening the stain their crimes have already left on our courts.
I respectfully dissent.
. Consider the words of a government attorney who prosecuted Maloney:
As a judge [Maloney] was tough and hard-nosed. Many prosecutors liked working in his courtroom because he was tough and hard-nosed. But one of the things that I have heard over and over again from lawyers in the community is that he took it far too far; that he was ruthless; that he heartlessly meted out sentences without any compassion. |T]he only time there was compassion that we can see has to do with the times in which money was being passed.
United States v. Thomas J. Maloney and Robert McGee, No. 91 CR 477, Sentencing Tr. 559-60 (N.D.Ill. July 21, 1994) (remarks of Assistant United States Attorney Scott Mendeloff).
. I recognize, of course, that there are prosecution-minded judges and defense-minded judges and that although these predispositions can have a very real impact on the kind of trial the parties receive, we ordinarily do not recognize this bias as a constitutional deprivation. See ante at 688; see also Del Vecchio, 31 F.3d at 1390-91 (Easter-brook, J., concurring). Judges are, after all, human beings, and we dare not fool ourselves into thinking that any judge can completely divorce herself from her own experiences and predilections. Id. at 1372 (majority); see also Benjamin *701N. Cardozo, The Nature of the Judicial Process 168-69 (Yale Univ. Press 1921). But the distinction between the honest judge, who labors to varying degrees of success to rise above his prejudices, and the dishonest judge, who willingly abandons his oath and yields to the coarsest of proclivities, cannot be overstated. We simply have no business assuming that a judge who is willing to acquit an accused murderer for a few thousand dollars will make any effort to protect the rights of a defendant who has not greased his palm.
. In this case there were plenty of issues that implicated Judge Maloney's discretion and thus his ability to influence the case against Bracy and Collins: the credibility questions presented by the petitioners' motion to suppress key evidence; the bolstering of prosecution witnesses; the collateral impeachment of defense witnesses; *702improper prosecution argument to the jury; the denial of a continuance prior to the sentencing hearing; and the refusal to sever the sentencing hearings. Viewed singly, none of Maloney’s rulings on these issues may seem critical to the outcome of the case; but as my colleagues concede, the cumulative effect of these rulings may be “greater than we imagine.” Ante at 690 (citing Tyson, 50 F.3d at 439).