John Wayne Gacy v. George Welborn, Warden, Menard Correctional Center, and Roland W. Burris, Attorney General of Illinois

EASTERBROOK, Circuit Judge. John Wayne Gacy is a serial killer. Between 1972 and 1978 he enticed many young men to his home near Chicago for homosexual liaisons. At least 33 never left. Gacy tied up or handcuffed his partners, then strangled or choked them. Twenty-eight of the bodies were dumped into the crawl space under the Gacy residence; one was entombed under the driveway; the rest were thrown into the Des Plaines River. Gacy, who operated a construction business, had his workers dig trenches and throw lime into the crawl space. Gacy’s wife complained about an “awful stench.” But the slaughter continued until the disappearance of 15 year old Robert Pi-est on December 11, 1978. Piest vanished after telling his mother that he was going to see a building contractor about a summer job. The presence of Gacy’s truck outside the place where Piest was to meet his potential employer led to Gacy’s arrest within two days. The discovery of so many skeletons, several with rags stuffed in the victims’ mouths, created a national sensation. Gacy regaled the police with stories about his exploits, which he attributed to “Jack,” an alternative personality. A jury convicted Gacy in March 1980 of 33 counts of murder, rejecting his defense of insanity. The same jury sentenced Gacy to death for 12 of these killings, the only 12 that the prosecution could prove had been committed after Illinois enacted its post-F^rrocm death penalty statute. The Supreme Court of Illinois affirmed. People v. Gacy, 103 Ill.2d 1, 82 Ill.Dec. 391, 468 N.E.2d 1171 (1984), cert. denied, 470 U.S. 1037, 105 S.Ct. 1410, 84 L.Ed.2d 799 (1985). That court also rejected a collateral attack, 125 Ill.2d 117, 125 Ill.Dec. 770, 530 N.E.2d 1340 (1988), cert. denied, 490 U.S. 1085, 109 S.Ct. 2111, 104 L.Ed.2d 671 (1989). A United States district court has decided that Gacy is not entitled to collateral relief. 1992 WL 211018, 1992 U.S.Dist. LEXIS 12498 (N.D.Ill.), motion to suspend judgment denied, 1992 WL 358851,1992 U.S.Dist. LEXIS 18073. Opinions in this case already exceed 200 pages. We spare readers further recapitulation and turn directly to the four arguments Gacy has culled from more than 100 raised at one or another stage of this litigation, now 14/6 years old. Gacy has in this sense already escaped the 12 judgments of execution, for judge and jury cannot decide ivhether a murderer will die, but only how soon. I Illinois commits the capital sentencing decision to the jury. If the jury convicts a defendant of a capital offense, there is a sentencing proceeding. At this proceeding *307 the prosecution bears the burden of establishing the existence of defined aggravating circumstances. If the jury unanimously decides that there is at least one aggravating circumstance, the defendant becomes eligible for a death sentence. Section 9 — 1(g) of the Illinois statute (Ill.Rev.Stat. ch. 38 ¶ 9 — 1(g) (1979), now 720 Ill.Comp.Stat. § 5/9 — 1(g)) spells out what happens next: If the jury determines unanimously that there are no mitigating factors sufficient to preclude the imposition of the death sentence, the court shall sentence the defendant to death. Unless the jury unanimously finds that there are no mitigating factors sufficient to preclude the imposition of the death sentence the court shall sentence the defendant to a term of imprisonment under Chapter V of the Unified Code of Corrections. The second paragraph means that a single juror’s belief that the defendant has demonstrated the existence of a single mitigating factor precludes the death sentence. Such a rule surprises some judges, who are accustomed to telling jurors that decisions must be unanimous. For example, in Kubat v. Thieret, 867 F.2d 351 (7th Cir.1989), the judge instructed the jury: “If, after your deliberations, you unanimously conclude that there is a sufficient mitigating factor or factors to preclude imposition of the death sentence, you should sign the form which so indicates.” The court did not tell the jury that a single juror could block capital punishment, and we held the instruction violated the Constitution. 867 F.2d at 371-74. Cf. Mills v. Maryland, 486 U.S. 367, 108 S.Ct. 1860, 100 L.Ed.2d 384 (1988); McKoy v. North Carolina, 494 U.S. 433, 110 S.Ct. 1227, 108 L.Ed.2d 369 (1990). When the jury knows about its full options, however, the Illinois system comports with all constitutional requirements. Silagy v. Peters, 905 F.2d 986, 997-1001 (7th Cir.1990); Williams v. Chuns, 945 F.2d 926, 935-38 (7th Cir.1991). A At the close of Gacy’s penalty proceeding, Judge Garippo instructed the jury: If, after your deliberations, you unanimously determine that there are no mitigating factors sufficient to preclude the imposition of the death sentence on the Defendant, you should sign the verdict form directing a sentence of death. If, after your deliberations, you are not unanimous in concluding that there are no mitigating factors sufficient to preclude imposition of the death sentence, you must sign the verdict form directing a sentence of imprisonment. As Judge Grady, who denied Gacy’s petition for a writ of habeas corpus, remarked: “This written instruction is completely accurate.” The jurors had this instruction, like the others in the three-page charge, during their deliberations. Unfortunately, Judge Garippo did not read the instruction to the jury as written — or at least the court reporter did not take down the same words that appear in the written instructions. The transcript has it that the second sentence of this instruction was delivered as: “If, after your deliberations, you unanimously conclude there are mitigating factors sufficient to preclude the imposition of the death penalty, you must sign the verdict form directing a sentence of imprisonment.” This sentence, Gacy submits, carries the same defect as the instruction in Kubat and vitiates the death sentences. Gacy’s jury was told that if it is unanimous in finding a mitigating circumstance, it must return a verdict of imprisonment. The oral version of the instruction is accurate, as far as it goes. But the instruction did not give the jury the whole truth, because it did not tell the jurors what to do in the event of disagreement. Mills and McKoy, the closest decisions from the Supreme Court, dealt with instructions telling the jury that only unanimous agreement, on a particular mitigating factor, would permit the jury to return a verdict of imprisonment. Such an instruction creates a risk that even though every juror believes that there is some mitigating factor, the jurors’ inability to agree on a particular factor would lead to the defendant’s execution. No such problem infects the oral instruction to Gaey’s jury. Kubat dealt with a related question: what if the jury is ignorant of the power of a single juror to block the *308 death penalty under state law? Both oral and written instructions in Kubat’s case dealt only with unanimous verdicts, and a reasonable jury might well have thought that it was supposed to continue deliberating until it reached agreement — just as it had done at the guilt phase of the trial. Like Judge Grady, we conclude that reasonable jurors would not have been under the misapprehension that they had to reach unanimous agreement. Judge Garippo opened the sentencing phase of Gacy’s trial by describing to the jurors the findings they would need to make. The judge said, among other things: If you cannot unanimously agree that there are no sufficient mitigating factors to preclude the imposition of the death penalty, you will sign that verdict so indicating, and the Court will sentence the Defendant to imprisonment. The final instructions came close on the heels of the preliminary set, for the jury did not receive fresh evidence; instead the lawyers presented arguments based on the evidence at the five-week trial. Defense counsel emphasized during these arguments that unanimity was unnecessary: [T]he only way that you can impose the death penalty on Mr. Gacy, and His Honor will instruct you, it is a unanimous decision, all 12 of you have to agree to give Mr. Gacy the death penalty. If there is just one of you who feels that he was acting under an emotional disturbance, or if there is just one of you who feels it would not be the right thing to do, if there is one of you who feels that he should be studied for any reason at all, if there is one of you, then you must sentence him or direct the court to sentence him to a term of imprisonment. This argument, an accurate statement of Illinois law, was presented without objection from the prosecutor. For his part, the prosecutor did not urge the jury to seek unanimity on mitigating factors. What we have, in sum, is a slip of the judicial tongue. No one noticed at the time; defense counsel did not object to the misreading of the written instruction. The complete, and completely accurate, instructions were available to the jurors during their deliberations. The text was short; vital information did not drown in a sea of words. If the jurors wondered about the consequences of disagreement, they had correct answers at their elbows. Within two hours, the jurors brought back death verdicts on all 12 counts. This is too little time to reach unanimous agreement on aggravating factors and beat down even a single holdout on mitigating factors. So the question at hand probably did not arise in the jury’s deliberations; they must have been in agreement from the outset. These circumstances “rule out [any] substantial possibility that the jury may have rested its verdict on the ‘improper’ ground”. Mills, 486 U.S. at 377, 108 S.Ct. at 1867. “Jurors do not sit in solitary isolation booths parsing instructions for subtle shades of meaning in the same way lawyers might. Differences among them in interpretation of instructions may be thrashed out in the deliberative process, with commonsense understanding of the instructions in the light of all that has taken place at the trial likely to prevail over technical hairsplitting.” Boyde v. California, 494 U.S. 370, 380-81, 110 S.Ct. 1190, 1198, 108 L.Ed.2d 316 (1990). B This assessment supposes that the written instruction is satisfactory. Although we have twice (in Silagy and Williams) rejected constitutional challenges to this instruction, a district court, relying on research by the late Hans Zeisel, has concluded that jurors do not understand such instructions, which therefore violate the Constitution. Free v. Peters, 806 F.Supp. 705 (N.D.Ill.1992). The district judge, concluding that Silagy and Williams depend on mistaken beliefs, has refused to carry out the holdings of those cases. Id. at 731. If we conclude (as we have) that the validity of his sentence depends on the written instruction, Gacy submits, we should remand so that he may present to Judge Grady the same evidence that persuaded Judge Aspen in Free. Gacy filed a post-judgment motion in the district court asking for such a hearing. The district court denied this motion, and Illinois asks us to treat the subject as waived. It *309 would have been within the district court’s discretion to treat the argument as untimely. Instead the court wrote that, because Free would govern the outcome of this ease, Gacy could raise the question in a new collateral attack, which would abide the outcome of Free. After McCleskey v. Zant, — U.S. -, 111 S.Ct. 1454, 113 L.Ed.2d 517 (1991), however, prisoners cannot wage a series of collateral attacks, adding new issues each time. There will be only one federal adjudication of Gacy’s claims. Those not made now are lost. Gacy has been attacking the jury instructions from the beginning, and it is best to consider all of the contentions he presses on us. Putting Gacy’s case on ice while another panel of this court considers the appeal of Free would be inappropriate. Implementation of Gacy’s sentences has been stayed by federal litigation since 1989, and we have been told to resolve capital appeals expeditiously. In re Blodgett, — U.S. -, —, 112 S.Ct. 674, 676, 116 L.Ed.2d 669 (1992). The district judge in Free has asked a magistrate judge to inquire into claims that the defendant misled the court about the provenance of the study. All other proceedings in Free have ground to a halt, and we cannot tell how long it may be before the appeal in that case will be resolved (or even which side will turn out to be the appellant). Both Gacy and the State of Illinois are entitled to decision without indefinite delay. 1 — 1 — Professor Zeisel conducted a survey of persons reporting to a courthouse as potential jurors. He gave them the facts of a case and a set of instructions based on the 1987 Illinois Pattern Instructions for capital cases. (These instructions are similar, though not identical, to those given to Gacy’s jury.) Then he asked a series of questions designed to test the subjects’ comprehension of the instructions. The questions most pertinent to Gacy’s argument follow: Question 4. A juror decides that the fact that Mr. Woods was only 25 year’s of age when he committed the murder is a mitigating factor sufficient to preclude the death penalty. However the other eleven jurors disagree and insist that his age is not a mitigating factor. The one juror believes that she cannot consider a mitigating factor unless the entire jury agrees upon it and votes for the death penalty. She votes for the death penalty. Has that juror followed the judge’s instructions? One quarter of the subjects answered “yes”, leading Prof. Zeisel to conclude that as much as a third of the pool of jurors would not understand a critical feature of the instructions. Prof. Zeisel computed a confidence interval of 8.7%, leading to the calculation 25.0% ± 8.7% = 16.3% «=> 33.7%. Question 5. A juror decides that the fact that Mr. Woods was good to his family is a mitigating factor sufficient to preclude the death penalty. However, the other eleven jurors disagree. The other jurors insist that no juror should consider the defendant’s good relations with his family as a mitigating factor unless they all agree it is a mitigating factor. The one juror accepts this approach and votes for the death penalty. Has the juror followed the judge’s instructions? On this question, 36.5% answered “yes”, leading to the computation 36.5% ± 9.6% = 26.9% <=> 46.1% wrong answers. See 806 F.Supp. at 767. Gacy contends that these error rates demonstrate that the written instruction is confusing and thus could not have corrected the misimpression created by the oral instruction. The district court in Free concluded that the study as a whole “refut[es] the underlying judicial assumptions driving the decisions in Silagy and Williams [and shows] a reasonable likelihood that [the] jury (1) believed that only the *310 statutory mitigating factors, or factors comparable to them, could preclude the imposition of the death penalty, and (2) was confused about (i) which side, if any, had a burden of persuasion, and (ii) the nature of that burden.” 806 F.Supp. at 731. The court’s approach would require new penalty proceedings in almost all Illinois capital cases — although Free rejected the argument that the answers to Questions 4 and 6 show that there is an independent constitutional flaw in the unanimity instruction. Id. at 720-22. As it did in Free, the state contends that the reported data are inaccurate. Illinois points out that although the jury instructions were read to the panel, the questions were administered in writing. Perhaps, then, the subjects grasped the instructions but misunderstood (or could not read) the questions. The questions are not free from ambiguity. For example, might a subject understand Question 5 as inquiring whether a juror legitimately may be persuaded by other jurors’ arguments? Questions 4 and 5 also involve mitigating factors that were not on the list recited in the instruction. Some subjects may have understood these questions as asking whether a particular factor is appropriate, rather than whether the instructions require unanimity. Finally, the state reminds us that Prof. Zeisel was a zealous opponent of capital punishment and that the study was conducted under the auspices of the MacArthur Justice Center, an organization devoted to the defense of capital litigation, which may have influenced the findings no matter how careful the principal investigator sought to be. If we believed that the validity or soundness of Prof. Zeisel’s work were controlling, we would give Gacy the hearing he requests. We conclude,-however, that even taken for all it could be worth, Prof. Zeisel’s study does not assist Gacy. — 2 — Free refused to follow two opinions of this circuit, claiming the support of “empirical evidence refuting the underlying judicial assumptions driving” our cases. 806 F.Supp. at 731. Although Prof. Zeisel’s work may persuade us to reexamine Silagy and Williams, it does not permit district courts to disregard those decisions. Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477, 484, 109 S.Ct. 1917, 1921, 104 L.Ed.2d 526 (1989); Thurston Motor Lines, Inc. v. Jordan K. Rand, Ltd., 460 U.S. 533, 103 S.Ct. 1343, 75 L.Ed.2d 260 (1983); United States v. Burke, 781 F.2d 1234, 1239 n. 2 (7th Cir.1985). Ours is a hierarchical judiciary, and judges of inferior courts must carry out decisions they believe mistaken. Hutto v. Davis, 454 U.S. 370, 375, 102 S.Ct. 703, 706, 70 L.Ed.2d 556 (1982). A district judge who thinks that new evidence or better argument “refutes” one of our decisions should report his conclusions while applying the existing law of the circuit. Panels of three appellate judges establish rules for the approximately 100 judicial officers of this circuit, and for the approximately 22 million persons living within its boundaries. Some of these decisions are bound to be erroneous, or at least to depart from those the many other judges who are not on a given panel would have reached. Illinois litigated and won Silagy and Williams; it need not re-litigate them prisoner by prisoner, district judge by district judge. Cf. Steven G. Cala-bresi & Gary Lawson, Equity and Hierarchy: Reflections on the Harris Execution, 102 Yale L.J. 255, 273-79 (1992). Now that the question is here, we must decide whether Prof. Zeisel’s work justifies overruling two carefully thought out opinions. Silagy and Williams were expressed as decisions on questions of law, not as predictions about the findings of future survey research. Legal developments since 1991 do not call those opinions into question. To the contrary, the Supreme Court has reiterated that the choice between judge and jury in sentencing, and the allocation of burdens of production and persuasion after the prosecutor establishes aggravating circumstances, are questions of state rather than constitutional law. E.g., Delo v. Lashley, — U.S. -, 113 S.Ct. 1222, 122 L.Ed.2d 620 (1993). At all events these cases are immune from overruling on collateral attack. Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989), holds that federal courts *311 may not apply rules of constitutional law devised or altered after the conclusion of the prisoner’s direct appeals. Silagy and Williams show that when Gacy’s direct appeal ended in 1984 no constitutional problem was apparent in the sentencing instructions. Discarding two recent decisions of this circuit necessarily would establish a “new rule” for purposes of Teague. None of the exceptions to Teague is available to Gacy. Overruling Silagy and Williams would not place any conduct beyond the power of law to proscribe or establish a new category of fundamental constitutional norms. 489 U.S. at 307, 109 S.Ct. at 1073. In an effort to explain why he should be allowed to raise new arguments in federal court despite not having presented comparable contentions in state court, Gacy (like Free) points to the novelty of Prof. Zeisel’s research. Arguments based on his work were unavailable at the time of Gacy’s direct appeal, and hence, Gacy contends, he has “cause” for not presenting the claim to the state court. 2 This follows the path of Reed v. Ross, 468 U.S. 1, 104 S.Ct. 2901, 82 L.Ed.2d 1 (1984), which held that prisoners have “cause” for not presenting arguments so novel or so inconsistent with established law as to be “unavailable.” After Teague, however, this stratagem to avoid Wainwright v. Sykes, 433 U.S. 72, 97 S.Ct. 2497, 53 L.Ed.2d 594 (1977), does nothing except plead the petitioner out of court. Any claim sufficiently novel that it was unavailable during the state proceedings must be a “new rule” under Teague. Prihoda v. McCaughtry, 910 F.2d 1379, 1385-86 (7th Cir.1990). See also Joseph L. Hoffman, The Supreme Court’s New Vision of Federal Habeas Corpus for State Prisoners, 1989 Sup.Ct.Rev. 165, 183. Cf. Sawyer v. Whitley, — U.S. -, 112 S.Ct. 2514, 120 L.Ed.2d 269 (1992) (defendant who is “eligible” for capital punishment, as Gacy is, may not use the “actual innocence” exception to Sykes, and presumably not the “actual innocence” exception to Teague either). And calling the Zeisel study “new evidence” does nothing to break the grip of Teague and Sykes — not only because it is not new evidence about Gacy’s guilt (as opposed to the factual underpinning of a legal argument) but also because of cases such as Keeney v. Tamayo-Reyes, — U.S. -, 112 S.Ct. 1715, 118 L.Ed.2d 318 (1992), and Herrera v. Collins, — U.S. -, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993). Whatever power to revisit Silagy and Williams we may possess, we shall not exercise. The Zeisel study does those cases no damage. To see why, consider the question: Is an error rate of 25.0% ± 8.7% large or small? Such a- question has no answer. Large or small, compared to what? Presumably compared with some lesser error rate, reflecting greater comprehension achieved by changing the instructions to the jury. That is, actual levels of comprehension must be compared with achievable levels of comprehension, not with ideal levels. Yet Prof. Zeisel did not test jurors’ comprehension of instructions worded differently. Nothing in his work shows that some other set of instructions would do better. 3 To put this differently, there are many potential reasons why jurors might not grasp what a judge tells them. Think of just a few: (1) The instructions may be poorly drafted, with needlessly big and technical words and ambiguous constructions; (2) The instructions may convey rules of some complexity, which cannot be mastered on first exposure; (3) The instructions may use concepts that are inherently complex; (4) Jurors may be unable to grasp thoughts unfamiliar to them. If explanation (1) is at work, then the State of Illinois is responsible and may be called on *312 to improve things (although, as we explain below, the Constitution has but a limited role to play even here). Explanation (2), by contrast, attributes misunderstanding to a mixture of state law with the constitutional obligations announced by the Supreme Court. Cases beginning with Lockett v. Ohio, 438 U.S. 586, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978), have established a set of increasingly reticulated rules for capital sentencing, including shifting burdens, unanimity on some issues but not on others, and consideration of mitigating factors that do not appear in state statutes. Justices of the Supreme Court occasionally complain that these rules are too complex, e.g., Graham v. Collins, — U.S. -, -, 113 S.Ct. 892, 906-13, 122 L.Ed.2d 260 (1993) (Thomas, J., concurring); Walton v. Arizona, 497 U.S. 639, 656-74, 110 S.Ct. 3047, 3058-68, 111 L.Ed.2d 511 (1990) (Scalia, J., concurring in part) — and the high reversal rate in capital cases supports this perspective. Yet if the Constitution requires this convoluted set of rules, then the attendant confusion is a regrettable cost rather than a reason why the Constitution’s own norms are, in application, im-constitutional. Explanation (3), the complexity of the concepts, also plays a role. Burdens of proof and persuasion are hard to explain (one reason why this court strongly discourages efforts to define “reasonable doubt”, see United States v. Glass, 846 F.2d 386 (7th Cir.1988)). Learned Hand confessed that, after a lifetime in the law, he still did not understand the shadings among burdens of persuasion. United States v. Feinberg, 140 F.2d 592, 594 (2d Cir.1944). See also Flamm v. Eberstadt, 814 F.2d 1169, 1173-74 (7th Cir.1987). That subjects being peppered with questions about a complex concept they have encountered for the first time will not give answers satisfactory to lawyers is no surprise. It cannot be that the Constitution, which requires judges to tell juries to use these elusive concepts, is self-destructive because lay persons will experience difficulty in answering questions about what they have been told. Confusion and misunderstanding attributable to the Constitution of the United States does not yield a violation of that document. Then there is explanation (4): Human shortcomings. Difficulty in coping with abstract concepts (most jurors spend their lives in the world of the concrete) explains why we have lengthy arguments, why judges give instructions orally as well as in writing (and reinstruct juries that ask questions), why juries deliberate. Jurors who “don’t get it” on first hearing may do better as the process continues. Professor Zeisel himself concluded as much. Harry Kalven, Jr., & Hans Zeisel, The American Jury 149-62 (1966). See also Reid Hastie, Steven D. Penrod & Nancy Pennington, Inside the Jury 81 (1983) (concluding that jurors collectively remember 90% of the evidence and 80% of the instructions). In sum, it is inadmissible to use inaccurate answers attributable to explanations (2), (3), and (4) as reasons to condemn a jury instruction. The Constitution establishes a system of jury trials, which necessarily tolerates the shortcomings of that institution. Pointing to one of these shortcomings, no matter how vivid, does nothing to undercut the Constitution’s own method. 4 Professor Zeisel did not try to isolate the effect of the state’s drafting choices. For all we can tell, the best conceivable exposition would have reduced the misunderstanding rate on Question 4 from 25% to 24%, with comparably paltry changes on other questions. The study thus does not justify chastising the state, as opposed to the Supreme Court or the institution of the jury or the failings of the species. If the study enabled us to lay some responsibility at the state’s doorstep, this still would not permit a federal court to take a blue pencil to a state’s jury instructions. For as long as the United States has been a nation, judges have been using legalese in instructing juries, with an inevitable adverse effect on the jury’s comprehension. We do not think that traditional forms of jury instruction are now, and always have been, unconstitutional because of this. *313 One enduring element of the jury system, no less vital today than two centuries ago, is insulation from questions about how juries actually decide. Jurors who volunteered that they did not understand their instructions would not be permitted to address the court, and a defendant could not upset a verdict against him even if all of the jurors signed affidavits describing chaotic and uninformed deliberations. E.g., McDonald v. Pless, 238 U.S. 264, 35 S.Ct. 783, 59 L.Ed. 1300 (1915); Hyde v. United States, 225 U.S. 347, 381-84, 32 S.Ct. 793, 807-09, 56 L.Ed. 1114 (1912); United States v. Schwartz, 787 F.2d 257, 261-62 (7th Cir.1986); Fed.R.Evid. 606(b). Instead of inquiring what juries actually understood, and how they really reasoned, courts invoke a “presumption” that jurors understand and follow their instructions. As Rule 606(b) shows, this is not a bursting bubble, applicable only in the absence of better evidence. It is a rule of law — a description of the premises underlying the jury system, rather than a proposition about jurors’ abilities and states of mind. See Parker v. Randolph, 442 U.S. 62, 73, 99 S.Ct. 2132, 2139, 60 L.Ed.2d 713 (1979) (plurality opinion) (“A critical assumption underlying [the] system [of trial by jury] is that juries will follow the instructions given them by the trial judge. Were this not so, it would be pointless for a trial court to instruct a jury, and even more pointless for an appellate court to reverse a criminal conviction because the jury was improperly instructed.”); Jackson v. Denno, 378 U.S. 368, 382 n. 10, 84 S.Ct. 1774, 1783 n. 10, 12 L.Ed.2d 908 (1964); Opper v. United States, 348 U.S. 84, 95, 75 S.Ct. 158, 165, 99 L.Ed. 101 (1954). Cf. Sparf & Hansen v. United States, 156 U.S. 51, 80, 15 S.Ct. 273, 284, 39 L.Ed. 343 (1895). Cases doubting jurors’ ability to put out of their minds events vividly described in court have not expressed equivalent doubts about jurors’ ability to follow instructions on the law. Compare Bruton v. United States, 391 U.S. 123, 88 S.Ct. 1620, 20 L.Ed.2d 476 (1968), with Richardson v. Marsh, 481 U.S. 200, 211, 107 S.Ct. 1702, 1709, 95 L.Ed.2d 176 (1987). See also, e.g., Francis v. Franklin, 471 U.S. 307, 324-25 n. 9, 105 S.Ct. 1965, 1976 n. 9, 85 L.Ed.2d 344 (1985). Social science has challenged many premises of the jury system. See generally Symposium, Is the Jury Competent?, 52 L. & Contemp. Prob. (Aut.1989); Symposium, The Selection and Function of the Modem Jury, 40 American L.Rev. (Win.1991). Students of the subject believe, for example, that jurors give too much weight to eyewitness evidence and not enough weight to other kinds. See Credibility Assessment (Yuille ed. 1989); Elizabeth F. Loftus, Eyervitness Testimony: Psychological Research and Legal Thought, 3 Crime and Justice 105 (1981). Cf. Lea Brilmayer & Lewis Kornhauser, Qriantita-tive Methods and Legal Decisions, 46 U.Chi. L.Rev. 116, 135-48 (1978). Still, the ability of jurors to sift good evidence from bad is an axiom of the system, so courts not only permit juries to decide these cases but also bypass the sort of empirical findings that might help jurors reach better decisions. See Krist v. Eli Lilly & Co., 897 F.2d 293, 296-99 (7th Cir.1990); United States v. Hudson, 884 F.2d 1016, 1024 (7th Cir.1989). Juries have a hard time distinguishing “junk science” from the real thing, but aside from some tinkering with the expert testimony admitted at trial, this shortcoming has been tolerated. Compare Carroll v. Otis Elevator Co., 896 F.2d 210, 215-18 (7th Cir.1990) (concurring opinion), with Christophersen v. Allied-Signal Corp., 939 F.2d 1106 (5th Cir.1991) (in banc). Cf. Daubert v. Merrell Dow Pharmaceuticals, Inc., 951 F.2d 1128 (9th Cir.1991), cert. granted, — U.S. -, 113 S.Ct. 320, 121 L.Ed.2d 240 (1992). Jurors reach compromise verdicts, although they aren’t supposed to. Juries return inconsistent verdicts, representing irrational behavior or disobedience to their instructions. United States v. Powell, 469 U.S. 57, 105 S.Ct. 471, 83 L.Ed.2d 461 (1984). Juries act in ways no reasonable person would act. This is the standard for granting judgment notwithstanding the verdict in a civil case, or acquittal after verdict in a criminal case, or reducing an award of damages, and there are plenty of occasions for these post-verdict correctives. Yet for all of this, courts do not discard the premises of the jury system, postulates embedded in the Constitution and thus, within our legal system, unassailable. This shows up in a striking fact about the *314 Supreme Court’s treatment of social science: of the 92 cases between 1970 and 1988 addressing issues of evidence and trial procedure, not one relied on the extensive body of evidence about jurors’ conduct. J. Alexander Tanford, The Limits of a Scientific Jurisprudence: The Supreme Court and Psychology, 66 Ind.L.J. 137, 139 (1990). None of this is to suggest that judges ought to be indifferent to the way they write instructions. Polysyllabic mystification reduces the quality of justice. One of the Illinois 1987 pattern instructions is a quadruple negative: “If you do not unanimously find from your consideration of all the evidence that there are no mitigating factors sufficient to preclude imposition of a death sentence, then you should sign the verdict requiring the court to impose a sentence other than death.” Things seem to have gone downhill since Gacy’s jury was charged in 1981 (his jurors got the same information with only three negatives). Modern pattern instructions use simple words in short, concrete, declaratory sentences. E.g., Federal Judicial Center, Pattern Criminal Jury Instructions (1987) (including an appendix of suggestions for making instructions more understandable). Would it not be better for all concerned to give a charge such as: “If after full discussion any one of you believes that a mitigating factor makes death an excessive punishment, then you must return a sentence of imprisonment.”? For reasons we have discussed, even this “simplified” charge would leave many jurors dumbfounded; the underlying ideas are not at all simple, and words such as “mitigating” and “excessive” are foreign to jurors’ daily discourse. As there are no perfect trials, so there are no perfect instructions. How best to convey the law to lay persons sitting on juries is in the end a question for state legislatures and trial courts to resolve, and not for the federal kibitzers in collateral attacks many years later. The jury is a means to resolve disputes, not a waystation by which the controversy at trial is transported to a higher level of generality as a social science dispute about juries. Gacy’s jury was instructed within the wide bounds set by the Constitution. II Gacy’s remaining arguments require less discussion. Each received extended treatment by Judge Grady, whose opinion withstands all challenges. A Because comment in the media was especially intense in Chicago, the court chose a jury in Rockford. During the four days devoted to screening potential jurors, the court put many questions to candidates. The judge asked, for example, whether the pretrial publicity made it impossible to approach the subject with an open mind, and whether Gacy’s homosexuality (and the sexual nature of the crimes) would affect their judgment. The judge declined, however, to ask members of the venire exactly what they had read in the papers and exactly what they thought about homosexuals. Gacy contends that the judge’s refusal to ask the questions his lawyer proposed deprived him of a fair trial. The Constitution does not require a judge to use the defendant’s preferred screening questions. Mu’Min v. Virginia, — U.S. -, 111 S.Ct. 1899, 114 L.Ed.2d 493 (1991). Before the trial began, court and counsel compiled an index of what had appeared in the media. The judge asked the jurors what they watched and read. This information, combined with the index, enabled Gacy’s lawyers to know the allegations to which each juror had been exposed. As Judge Garippo concluded, it could have been counterproductive to require each juror to repeat the details, because this could have summoned up information that the juror otherwise would have forgotten — and in the process contaminated other members of the ve-nire. The index, the questioning, and drawing the venire from Rockford were adequate to select an unbiased jury. Invoking Morgan v. Illinois, — U.S. -, 112 S.Ct. 2222, 119 L.Ed.2d 492 (1992), Gacy contends that the judge’s questions about homosexuality were inadequate. Judge Garippo asked general questions, such as: “Now, does the fact that the Defendant is charged with homosexual conduct prevent you from being fair to either side?” and *315 “Could you put aside any feeling that you might have regarding homosexuality in rendering your verdict in this case?” These are inadequate, Gacy submits, because they invited bland “yes” or “no” answers and thus hid any subtle signs of bias. The questions his lawyer proposed, by contrast, called for potential jurors to describe their feelings about homosexuals. It may well be that the kind of questions proposed by the defense would have afforded greater entree into the jurors’ attitudes, but Morgan does not require them. That case tells judges to ask about specific areas of bias rather than asking a general question such as “can you be fair and impartial?” Judge Garippo did what Morgan requires. To do more would have extended the jury-selection process considerably and left many jurors flustered and resentful, which in the end may have worked against the defendant. Defendants’ safety lies in the size of the jury and in cautions from the court, not in extra questions posed in advance of trial. A long series of probing questions can anesthetize or offend the panel rather than enlighten judge and counsel. Experienced judges accordingly prune the list, omitting some that may look appropriate in isolation. Judge Garippo did just that — and proceeded to follow up with more pointed questions when appropriate. Judge Grady’s opinion details several of these sequences, which we need not repeat. Moreover, as Judge Grady pointed out, general questions about attitudes toward homosexuality were beside the point. Gacy admitted killing many persons. His defense was insanity. Gacy’s lawyer did not ask the judge to propound questions along the line of “Do you believe that homosexuals are more likely than heterosexuals to be sane?” or “Are you disposed to convict an insane person whose murders were related to homosexuality, although the law and the evidence require acquittal?” Of course questions this blunt would have been useless, but Gacy’s lawyer also did not propose ways to get at these subjects indirectly. Which is not to fault his lawyer: the subjects are not easy to broach. Even in hindsight it is not clear what more counsel, or the court, could have done. All that we need hold is that the trial judge was entitled to resist the invitation to turn voir dire into a trial of jurors’ attitudes about homosexuality. B Gacy presented his defense of insanity through six expert witnesses: four psychiatrists and two psychologists. Judge Garippo declined to allow Gacy’s lawyers to use these witnesses to relay to the jury verbatim statements Gacy made to them, ruling that these were hearsay when offered for the truth of the matter asserted. Gacy did not testify at trial, and the prosecutor used the hearsay objections to prevent Gacy from getting the more favorable portions of his story before the jury indirectly. The trial judge nonetheless permitted the expert witnesses to recount the substance of what Gacy had said. The prosecutor did not hesitate to ask these six witnesses, and the state’s own experts, about Gacy’s incriminating statements. These came in as admissions of a party opponent. Gacy contends that this one-sided use of his words violated his rights under the due process clause. On direct appeal, the Supreme Court of Illinois concluded that Gacy had defaulted this claim by failing to make appropriate offers of proof. 82 Ill.Dec. at 420-22, 468 N.E.2d at 1200-02. On collateral review, however, that court addressed the merits of Gacy’s contentions. In light of an intervening decision, People v. Anderson, 113 Ill.2d 1, 12-13, 99 Ill.Dec. 104, 495 N.E.2d 485 (1986), overruling People v. Hester, 39 Ill.2d 489, 237 N.E.2d 466 (1968), on which Judge Garippo had relied, the Supreme Court of Illinois held that Gacy’s psychiatric experts should have been permitted to repeat his exact words to the jury, the better to illuminate their diagnoses. Nonetheless, the court concluded, exclusion of this testimony was harmless error in light of the fact that all six witnesses fully explained the bases of their diagnoses by paraphrasing what Gacy had said to them. 125 Ill.Dec. at 775, 530 N.E.2d at 1345. The Supreme Court of Illinois did not reiterate the procedural holding from the first case, which clears the way for Gacy to argue the merits in federal court. Ylst v. Nunnemaker, — U.S. -, 111 S.Ct. 2590, 115 L.Ed.2d 706 (1991); Harris v. Reed, 489 U.S. *316 255, 109 S.Ct. 1038, 103 L.Ed.2d 308 (1989). After a painstaking review of the evidence, Judge Grady concluded that all six witnesses had presented the jury with full assessments of Gacy’s condition and complete statements of their reasons. Nothing material was withheld, so the error was indeed harmless, just as the Supreme Court of Illinois had concluded. Without detracting in any way from the care Judge Grady lavished on this issue, or the correctness of his decision that any error was harmless, we hold that there was no error at all — no constitutional error, that is. Anderson was based on state law. Beyond explicit rules such as the privilege against self-incrimination and the confrontation clause, none of which applies here, the Constitution has little to say about rules of evidence. Estelle v. McGuire, — U.S. -, 112 S.Ct. 475, 116 L.Ed.2d 385 (1991). The hearsay rule and its exception for admissions of a party opponent are venerable doctrines; no serious constitutional challenge can be raised to them. A challenge would lie if a state used its evidential^ rules to blot out a substantial defense. See Chambers v. Mississippi, 410 U.S. 284, 298-303, 93 S.Ct. 1038, 1047-50, 35 L.Ed.2d 297 (1973); Green v. Georgia, 442 U.S. 95, 99 S.Ct. 2150, 60 L.Ed.2d 738 (1979). These cases hold that states must permit defendants to introduce reliable third-party confessions when direct evidence is unavailable. No court has extended them to require a state to admit defendants’ own out of court words. A defendant is available to himself as a witness. Nothing in the Constitution gives an accused the privilege of proffering, through hearsay, his self-serving statements while denying the state access to the rest of the story that could be got at by cross-examination. Gacy had a right to offer expert evidence. Ake v. Oklahoma, 470 U.S. 68, 105 S.Ct. 1087, 84 L.Ed.2d 53 (1985). He did so, six times over. That these witnesses had to beat around the bush in order to avoid hearsay does not create a constitutional problem. C At last we reach the inevitable attack on trial counsel. It is, as Judge Grady concluded, unpersuasive. Gacy received a skillful, vigorous defense by lawyers who were prepared to the nines. Gacy’s lead counsel, Sam Amirante, is now a judge of the court that conducted his trial. For this appeal, Gacy’s current legal team has discarded all but two objections. First comes the argument that Amirante raised the insanity defense over Gacy’s objection, depriving him of the ability to control decisions vital to the defense. Second, Gacy insists that Amirante’s loyalties were compromised by his pursuit of revenues from a book and other publicity. The lure of lucre led him to induce Gacy to confess, the argument goes, the better to get an interesting story — but with devastating effect on the defense. Although such claims could in principle raise problems, they fail for want of proof. Let us start with the latter claim. Ami-rante filed an affidavit denying that he was pursuing profits from publicity. No book by Amirante or any of his associates appeared after the trial. Although an investigator may have leaked some tape recordings to Tim Cahill, author of Buried Dreams (1985), Ami-rante denied authorizing or knowing of this misconduct. A tape recording of a conversation between Gacy and Amirante 10 months before trial contains a brief discussion of publication possibilities. Amirante offered to refer Gacy to another lawyer to pursue that possibility — exactly the right thing to do, and the antithesis of a conflict of interest. All that remains is one paragraph in an order entered by the trial court requiring Gacy’s lawyers to continue their work at public expense (Amirante had been retained privately at the outset): *317 The record does not reveal who raised the subject of royalties (the prosecutor, the judge, or defense counsel), .or the basis of the objection. For all we know, only the judge had movies on his mind. So Amirante asserted in his affidavit. In the absence of any contrary evidence, this is a blind alley. *316 That, over the objection of Defense Counsel, it is hereby ordered that Attorneys Amirante and Motta reimburse Cook County to the extent of fees received for services rendered from any royalties received as a result of book or movie rights hereafter acquired, excluding any professionally oriented works, lectures, treatises, or the like. *317 As for the contention that counsel barged ahead with an unwanted insanity defense: again the evidence gets in the way. Gacy cooperated with extended interviews and tests by six experts for the defense and another six for the state, not the behavior you would expect of a person who wanted to stand on a plain denial of guilt. In mid-trial, Gacy threatened to stop cooperating with his attorneys, complaining: “I’m not running the trial.” When the judge asked what the problem was, Gacy continued: “I was against the insanity defense from the beginning.” This assertion, never heard again during the trial, 5 has become the foundation for the attack on counsel. In an affidavit dated July 25, 1990, Gacy at last furnished a reason: “I couldn’t believe that anyone could go insane 33 different times and then run a successful business, [and] if I didn’t believe it how could [Amirante] expect 12 jurors to believe that”. Good question — but how did Gacy expect to persuade the jury to disregard his confessions, plus the damning evidence of the 28 skeletons under his house, a 29th under his driveway, and 4 more recovered from the river, not to mention the testimony of witnesses who barely survived their encounters with him? His current story, that he was out of town on every occasion, is unsupported by evidence and less plausible than his insanity defense. As Judge. Grady remarked, Gacy’s only real choice was between an insanity defense and a guilty plea. It may be that Gacy could have obliged Amirante to desist from the insanity defense and conduct a defense limited to guilt, trying (as Amirante did not) to suppress the confessions and fob off the significance of the human remains. We say “it may be” because several of Gacy’s experts testified that he was not competent to assist in his defense. Although Judge Garippo rejected that position and .ordered Gacy to stand trial, the duties of counsel representing a client of borderline competence are not so clearly established as the duties of counsel representing a normal defendant. However that may be, Gacy did not tell Amirante to stop. A statement such as “I was against the insanity defense from the beginning” is some distance from “I directed Amirante to drop that defense, and he refused.” Being “against” a defense at the outset is consistent with yielding to the judgment of those who know better. Even the' affidavit of 1990 does not assert that Amirante refused to carry out any direct instructions from his client. There is consequently no material dispute requiring an evidentiary hearing. Affirmed. Footnotes [1] Because the decision in Gacy’s case may affect many prisoners on Death Row in Illinois, we have accepted amicus briefs submitted by the MacArthur Justice Center and a representative of Prof. Zeisel, by Bernard Williams (of Williams v. Chrans, still litigating despite McCleskey), and by Free himself. These briefs oddly do not explain or defend either the legal or the factual analysis in Free; instead they urge this court to wait for Free to wend its way here. There is, however, no principle that legal issues must be resolved for the first time on appeal in whatever case reached the district court first. [2] Although Prof. Zeisel’s study of the Illinois pattern instructions postdates Gacy's conviction, studies of other jury instructions, reaching similar conclusions, were published in the 1970s. See Robert P. Charrow & Veda R. Charrow, Making Legal Language Understandable: A Psycholinguistic Study of Jury Instructions, 79 Co-lum.L.Rcv. 1306 (1979); Amiram Elwork, Bruce Sales & James Alfini, Jundic Decisions: In Ignorance of the Law or in Light of It, 1 L. & Human Behavior 163 (1977). That jurors had trouble coping with gobbledygook in instructions is no news. Strangely, the parties make nothing of these studies predating Gacy's trial. [3] By contrast, the studies cited in note 2 tried out variations of the instructions to determine which were more comprehensible and the degree of improvement attainable from rewording. [4] We arc of course aware that the Constitution does not require Illinois to use juries in capital sentencing decisions. Still, the assumptions un-dergirding the use of juries in trials are no less appropriate when states elect to give juries a role in sentencing.