Convicted of two capital murders and sentenced to death, James Tenner seeks a writ of habeas corpus. Believing that his former business partner Albert Sauls had alienated the affections of his former girlfriend Shirley Garza, Tenner planned and carried out an execution. Sauls and Tenner operated trucking businesses from the same building in South Chicago Heights. In his space Tenner set up three ropes and nooses, each just the right height for the neck of the intended victim. Then with shotgun in hand he invaded the adjoining premises, finding (as he expected) Sauls and his employee Alvin Smith. He held the two at bay until Garza and Donna Sauls, Albert's wife, arrived. Tenner directed Garza and Donna Sauls to bind the men hand and foot; when that had been accomplished, Tenner compelled Garza to bind Donna Sauls the same way. All four moved to Tenner's garage, where the three bound victims were restrained in the nooses designed for them. Tenner then tied up Garza and constructed a fourth noose for her. After covering his victims' mouths with duct tape, Tenner delivered a harangue (lasting more than two hours) about the victims' supposed misconduct regarding both his business and his personal life, and the Saulses' treatment of Garza. Tenner then moved Garza to his office and returned to shoot the three remaining victims in their faces at point-blank range. Smith and Donna Sauls died instantly. Albert Sauls had been able to free his hands and used his arms to shield his face; he was left for dead but survived and testified against Tenner at trial. So did Garza, who was rescued by police after Tenner kidnapped her and drove off. Testifying on his own behalf, Tenner conceded that he prepared the ropes and shot the victims; his defense was that he sought to protect Garza from the Saulses and therefore committed second rather than first degree murder.

Tenner's convictions and sentence were affirmed on direct appeal. People v. Tenner, 157 Ill. 2d 341, 626 N.E.2d 138 (1993). A collateral attack in state court was unsuccessful. People v. Tenner, 175 Ill. 2d 372, 677 N.E.2d 859 (1997). The district judge rebuffed Tenner's federal collateral attack. Tenner v. Gilmore, 1998 U.S. Dist. Lexis 16188 (N.D. Ill. 1998). In this court Tenner presents many contentions, with multiple subparts. Most can be resolved with no analysis beyond citations to Holman v. Gilmore, 126 F.3d 876 (7th Cir. 1997), and Gosier v. Welborn, No. 98-2806 (7th Cir. Apr. 15, 1999). Four require discussion, but to obtain relief Tenner must demonstrate that the Supreme Court of Illinois made "a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States". 28 U.S.C. sec.2254(d)(1). This is the standard of the Antiterrorism and Effective Death Penalty Act, which applies because Tenner filed his federal petition after April 24, 1996. See Lindh v. Murphy, 521 U.S. 320 (1997).

1. The Supreme Court of Illinois reversed Tenner's conviction for the attempted murder of Albert Sauls, because the judge gave a confusing response to a question the jury posed during its deliberations. 626 N.E.2d at 155. This set up an argument that the convictions for first degree murder of Donna Sauls and Alvin Smith also should be reversed. According to the instructions, the jury could convict of first degree murder if it found that Tenner fired the fatal shots and that any one of four other conditions obtained: that Tenner intended the victims' deaths; that his acts created a strong probability of death; that the deaths occurred in the course of aggravated unlawful restraint (activating the felony-murder doctrine); or that Tenner attempted to murder Albert Sauls (another felony-murder possibility). The jury returned general verdicts and therefore did not reveal which one or more of these four circumstances it had found. Tenner argued that the jury might have relied exclusively on the attempt to murder Albert Sauls. With that conviction gone, the argument ran, the others fell automatically. The state judiciary did not agree:

With respect to Alvin Smith and Donna Sauls, the jury returned general verdicts finding the defendant guilty of their murders. A "general finding of guilty is presumed to be based on any good count in the indictment to which the proof is applicable. [Citations.]" (People v. Lymore (1962), 25 Ill. 2d 305, 307-08, 185 N.E.2d 158.) Thus, even if the defendant's conviction for attempted murder is reversed, which means that he could not have been found guilty of felony murder based on that offense, his convictions for murder may still stand. There were other good charges of murder to which the proof was applicable: first degree intentional murder, first degree murder based on acts creating a strong probability of death, and felony murder based on aggravated unlawful restraint. As no challenge has been made to these three theories of the offense, there is no need to reverse the defendant's convictions for first degree murder.

626 N.E.2d at 155 (citations omitted; brackets in original). Tenner now contends that Lymore, which allows a verdict to stand if the proof supports any theory of guilt, is inconsistent with federal law exemplified by Griffin v. United States, 502 U.S. 46 (1991), and Yates v. United States, 354 U.S. 298 (1957) (overruled on other grounds by Burks v. United States, 437 U.S. 1 (1978)).

Griffin and Yates deal with cases in which one theory of culpability is flawed while another is sound, and the jury does not reveal which theory it adopted. In one set of cases the flaw is legal. A jury may be told, for example, that the law forbids the doing of either A or B, and the verdict shows that the jury found that the defendant did one of these things, but not which. If one of the two is not a crime, Yates concludes, a new trial must be held. For all the court can tell, the jury found that the defendant committed the act that the law does not condemn. In the other set of cases the flaw is in the proof. The jury is told correctly that the law forbids the doing of either A or B, and the record permits a rational trier of fact to find that the defendant did A, but it does not permit a conclusion that he did B. In these cases, Griffin holds, the verdict is sound. A reviewing court should assume that the jury found that the defendant committed the act that the facts support; anything else would attribute irrationality to the jury. Juries apply to the facts of the case the law articulated by the judge. Courts assume that juries can distinguish good proof from bad, but juries do not separate good law from bad. That's the line between Yates and Griffin. When the facts could justify a conclusion that the defendant did both A and B, but only A is illegal, a court has no reason to believe that the jury found A, and a new trial ensues. Tenner contends that his case fits the Yates pattern rather than the Griffin pattern and that the Supreme Court of Illinois failed to appreciate the difference between these situations. As he sees it, the attempted murder charge was factually supported but legally erroneous, so that a well-functioning jury might well have convicted him of murder because it believed that he shot Donna Sauls and Alvin Smith (not intending to kill either one) in the course of attempting to kill Albert Sauls.

Tenner's description is not quite right. The attempted murder charge was not legally defective; the law of Illinois proscribes attempted murder. Only a garble in the judge's response to a question from the jury during deliberations spoiled the conviction. This fluff did not affect the instructions on the murder charges. Let us assume, nonetheless, that the situation is close enough to Yates as to fall within its gravitational influence. That raises the question whether the distinction between factual and legal shortcomings is one of constitutional law, binding on state as well as federal courts. Both Griffin and Yates were direct appeals in the federal system. Yet we need not decide whether these cases clearly establish a rule of constitutional law, as opposed to sound practice, because Tenner can not prevail either way.

Recall the point of Griffin: a jury may be relied on to get the facts right and avoid logical errors. If that is so, then the murder verdicts are unassailable. Trussing up four people, putting their heads in nooses, binding their mouths with duct tape, and forcing them to listen to a tirade laced with promises of their imminent deaths-- conduct that Tenner admitted on the witness stand-- meets anyone's definition of aggravated unlawful restraint. The jury convicted Tenner of four counts of this crime, which makes the killings first-degree murder under the felony-murder doctrine even if the jury believed (what is exceptionally unlikely) that Tenner did not intend the death of Donna Sauls or Alvin Smith. The four convictions for unlawful aggravated restraint are the equivalent of a special verdict by the jury, identifying a circumstance that supports a first-degree murder conviction. Special verdicts avoid the Yates problem, because the court then can be confident that the facts as the jury believed them to be are a legally proper basis of conviction.

Asked at oral argument why the convictions for unlawful restraint do not show that the murder verdicts are reliable, Tenner's counsel replied that the jury may have rendered a compromise verdict or disregarded the unlawful restraint convictions and relied only on the attempted murder conviction. Yet why should we attribute random or irrational conduct to this jury? If that were appropriate, then Griffin should have come out the other way--for it is possible that the jury in Griffin convicted the defendants on the factually deficient theory while finding against the prosecutor on the one supported by the evidence. What Griffin holds is that perverse factfinding should not be attributed to juries. Sure, it is possible; in a large enough sample of cases weird things are bound to happen. But verdicts are not annulled just because illogical behavior might have occurred. A jury that found aggravated unlawful restraint based on Tenner's confession in open court--a confession corroborated by the physical evidence and the testimony of the two survivors--would have been out of its mind to forget about the restraint and convict Tenner of murder based only on the attempted-murder charge. For that to happen, the jury would have had to conclude that Tenner did not intend to kill either Donna Sauls or Alvin Smith, that shotgun blasts in the face at close range do not create a strong probability of death, and that Tenner did not unlawfully restrain any of the four victims, contradicting the rest of its verdicts. Few things in litigation are as certain as the conclusion that the murder verdicts in Tenner's case were supported by the law and the evidence. Griffin rather than Yates rules the day here.

2. More than three years after his trial, Tenner was examined by a psychiatrist in connection with the post-conviction proceedings in state court. Dr. Lyle Rossiter concluded that Tenner suffers from a "thought disorder" that "grossly impairs his ability to communicate in a manner that allows [his post-conviction] counsel to ascertain the basis, if any, for a complaint that he has been deprived of his constitutional rights [at trial]." This conclusion forms the basis for Tenner's contention that he was not competent to stand trial. If he could not communicate effectively with counsel in 1995, during the post-conviction proceedings, then he could not have communicated effectively during trial; and ability to assist in one's defense is undermined by inability to communicate. The constitutional question is whether the accused had "a reasonable degree of rational understanding [and] a rational as well as factual understanding of the proceedings against him." Godinez v. Moran, 509 U.S. 389, 396 (1993). See also Drope v. Missouri, 420 U.S. 162 (1975); Pate v. Robinson, 383 U.S. 375 (1966); Dusky v. United States, 362 U.S. 402 (1960); United States v. Grimes, No. 98-1828 (7th Cir. Apr. 19, 1999); Eddmonds v. Peters, 93 F.3d 1307 (7th Cir. 1996). Inability to communicate with counsel may affect this understanding.

Rossiter also concluded that Tenner suffers from a paranoid delusional disorder, marked by the imputation of malevolent motives to others. Although Dr. Rossiter did not conclude that this rendered Tenner insane at the time of the crime (or trial), his current lawyers contend that paranoia would have compounded the difficulty in communicating with or assisting his trial counsel. Trial testimony by both Albert Sauls and Shirley Garza describes Tenner as ranting, sometimes incoherently, during the hours while the victims were bound prior to the shootings. Interview notes show that Tenner told his trial lawyer that he heard voices during 1981 (six years before the crimes) and "while in custody." All of this calls into question Tenner's ability to understand the proceedings and assist in his defense at trial, according to his current lawyers. But the district judge declined to hold a factual hearing on this subject, ruling that the facts available to the state trial judge would not have created a substantial doubt about Tenner's competence.

This way of phrasing the question is important. The record certainly contains evidence that Tenner has mental problems. But people who suffer from paranoia display "different levels of skill and understanding at different times. . . . Persons afflicted by paranoia often are intelligent and skillful." Gosier, 175 F.3d at 509. What was Tenner's level of understanding and cooperation near the time of his trial? More to the point, what should the trial judge have understood about this issue? Tenner's current lawyers want us to look principally at Dr. Rossiter's evaluation, but this was not available to the trial judge. "Only when the facts at the time of trial create a bona fide doubt about an accused's fitness is a hearing required." Gosier, 175 F.3d at 507 (emphasis in original). Hindsight does not entitle a federal court to upset the result of a state proceeding that was conducted in a fundamentally fair manner.

To justify an evidentiary hearing at trial to explore the possibility of incompetence to stand trial, a defendant must produce "clear and convincing evidence [raising] threshold doubt about his competency." Nguyen v. Reynolds, 131 F.3d 1340, 1346 (10th Cir. 1997), quoting from Lokos v. Capps, 625 F.2d 1258, 1261 (5th Cir. 1980); see also Grimes, 173 F.3d at 635-36; Walker v. Attorney General of Oklahoma, 167 F.3d 1339, 1342-47 (10th Cir. 1999). Cf. Cooper v. Oklahoma, 517 U.S. 348 (1996) (preponderance standard governs the decision once competence has been drawn into question). The Supreme Court of Illinois concluded that Tenner had not demonstrated the threshold doubt about competence. It observed:

Nothing in the defendant's record or personal history disclosed any history of mental illness or emotional disturbance; there was nothing to prompt a separate inquiry into the defendant's mental condition. In rejecting this portion of the defendant's post-conviction petition, the judge below, who had also presided at trial, explained:

"There was nothing to suggest a psychiatric exam in this young Defendant's background. If there had been anything that would have suggested [a] psychiatric exam, I probably would have ordered it sua sponte. I've been known to do that to satisfy myself as to the mental competency of a defendant in front of the bench. There's none of that in this record."

677 N.E.2d at 864. Although the state court wrote this when assessing Tenner's argument that failing to obtain an earlier psychiatric assessment was ineffective assistance of counsel, it is equally apropos to the competence issue. And under sec.2254(d) this finding--that nothing suggested to the trial judge that Tenner had any mental problem (beyond the kind of problem implied by the nature of the offense)--blocks collateral review unless it "was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or . . . was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding." We do not think that the state court's assessment is either an unreasonable understanding of the facts or an unreasonable application of the law.

This is the sort of issue on which the legal change made by the AEDPA matters. The parties do not debate the legal standards; they were established in cases such as Drope and Pate. It involves, rather, the application of law to the facts of a particular case. "[W]hen the dispute lies not in the meaning of the Constitution, but in its application to a particular set of facts-- when it is, in the standard phrase, a 'mixed question of law and fact'--sec.2254(d)(1) restricts the grant of collateral relief to cases in which the state's decision reflects 'an unreasonable application of' the law. . . . [W]hen the constitutional question is a matter of degree, rather than of concrete entitlements, a 'reasonable' decision by the state court must be honored." Lindh v. Murphy, 96 F.3d 856, 870-71 (7th Cir. 1996) (en banc), reversed on other grounds, 521 U.S. 320 (1997). (Our decision on remand in Lindh shows how the AEDPA mattered. Applying the de novo review under pre-AEDPA law, we disagreed with the state courts' application of law to the facts of that case and issued the writ. Lindh v. Murphy, 124 F.3d 899 (7th Cir. 1997).)

Did the Supreme Court of Illinois act unreasonably in concluding that the trial judge lacked a good reason to doubt Tenner's mental soundness? We think that its assessment was eminently reasonable. Even now, Tenner's able counsel do not point to any circumstance beyond the facts of the crime itself that should have alerted the judge to any problem. The judge had ample opportunity to evaluate Tenner, not only in pretrial proceedings but also during his testimony. Tenner responded intelligently to questions posed by his lawyer (and those during cross-examination by the prosecutor). His testimony was lucid and suggested an understanding of the legal process and the nature of the charges against him. See Grimes, 173 F.3d at 635-36. Tenner now relies principally on a psychiatric evaluation that came long after trial--and even this evaluation is of little use, for Dr. Rossiter did not offer an opinion about Tenner's ability to understand the proceedings or assist his lawyer at the time of trial. Rossiter concluded that Tenner had little understanding of, or ability to assist his post-trial counsel with, constitutional issues. That's true of most entirely competent defendants; indeed, many graduates of law school have trouble with constitutional questions from time to time. Rossiter did not suggest that Tenner was unable to understand the charges laid against him or to assist in a defense against them. That is the critical issue; so even with the benefit of hindsight it is not possible to condemn the state court's approach as "unreasonable."

3. Tenner asked the jury to show mercy during the sentencing phase, adducing evidence that he had redeeming human qualities. For example, he presented the testimony of a prison guard to the effect that Tenner was a peaceful influence who helped resolve quarrels among other inmates, did not join fights, and helped keep the unit clean. Tenner wanted to present the testimony of Eric Stokes, another guard, expecting it to be similarly helpful. He had a constitutional right to present this testimony in mitigation. Hitchcock v. Dugger, 481 U.S. 393 (1987). But Stokes did not show up to testify, and Tenner's lawyer did not ask the court to enforce the subpoena. In post-conviction proceedings Tenner argued that the prison had a policy against testimony by guards, but the state court was not impressed. It observed that Tenner had not established the existence of this supposed policy, that the appearance of another guard undercut the assertion that the warden had any such policy, and that judges rather than wardens determine who is to testify. 677 N.E.2d at 868- 69. The court added that Tenner could not establish prejudice in any event, for he "has provided no written offer of proof of what the witness in question would have testified to." Ibid.

Because the state court resolved this claim without invoking a rule of forfeiture--despite Tenner's failure to ask the trial judge to enforce the subpoena, and despite his omission of this issue on direct appeal--we too must address the merits. Harris v. Reed, 489 U.S. 255 (1989). Tenner contends that Stokes' failure to appear enabled the prosecutor to paint the guard who had testified as a disgruntled employee who was helping Tenner only out of dissatisfaction with her superiors. Testimony from a second guard would have undercut this response, the argument goes. But would it? We can't evaluate the possibility without knowing what testimony Stokes would have given. The Supreme Court of Illinois pointedly remarked on the absence of an offer of proof. Tenner did not use the federal proceedings as an opportunity to repair the deficiency. Like the state court, we have no idea what Stokes would have said or how that testimony would have affected the prosecutor's closing argument. Perhaps Tenner's lawyer let the subject pass at trial precisely because he did not expect Stokes' testimony to be helpful and valued the events more as a possible claim of error. What his lawyer may have had in mind is unimportant, however. Section 2254(d) tells us that collateral relief must be denied unless the state court's decision "was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or . . . was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding." The state court's demand that Tenner make an offer of proof in order to establish prejudice was neither contrary to clearly established federal law nor an unreasonable evaluation of a factual issue.

4. Ineffective assistance of counsel is the final issue we discuss. Tenner complains about his lawyer's decision not to secure psychiatric evaluations in order to explore the possibility of an insanity defense (or to use his mental condition as a mitigating factor in sentencing). He contends that his lawyer erred by failing to give the pretrial notice necessary before relying on a claim of self-defense. Finally, he submits that counsel fell below professional standards by conceding that Tenner unlawfully restrained his victims. The Supreme Court of Illinois carefully analyzed these and related contentions, both on direct appeal, 626 N.E.2d at 153-54, and on collateral review, 677 N.E.2d at 862-68. None of that court's analysis can be condemned as "contrary to, or . . . an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or . . . based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding." 28 U.S.C. sec.2254(d). Although the meaning of sec.2254(d) as applied to ineffective- assistance claims is now before the Supreme Court, see Williams v. Taylor, 163 F.3d 860 (4th Cir. 1998), cert. granted, 119 S. Ct. 1355, 143 L.Ed.2d 516(1999), we do not think that the Court's selection from among the interpretations can make a difference to Tenner, who loses under any plausible approach to sec.2254(d). See also Ashford v. Gilmore, 167 F.3d 1130, 1134-35 (7th Cir. 1999) (discussing how sec.2254(d) applies to ineffective-assistance claims in this circuit). Arguments concerning a possible mental evaluation are related to the question whether Tenner was competent to stand trial and do not require additional analysis when seen through the lens of an ineffective-assistance claim. Of the many errors that Tenner's current lawyers lay at the doorstep of his former ones, only one calls for comment.

At trial, Tenner's lawyer conceded the unlawful- restraint charge as part of a plan to argue that the jury should be charged on second-degree murder as a lesser included offense (and one that would avoid exposure to the death penalty). A conviction for second-degree murder is proper if the killer acted "under a sudden and intense passion resulting from serious provocation by the individual killed or another whom the offender endeavors to kill" or if "[a]t the time of the killing he believes the circumstances to be such that, if they existed, would justify or exonerate the killing under the principles stated in Article 7 of this Code, but his belief is unreasonable." 720 ILCS 5/9-2(a). Tenner testified that he believed that the Saulses had taken Garza captive and were using her as a sex slave, and that he acted to defend her. Although Garza testified that the Saulses had not done anything untoward, Tenner asserted that his belief was genuine and mitigated the offense even if held unreasonably. He could not use the supposed relationship between Garza and the Saulses as a defense without explaining what he had done in response, so he essentially admitted all of the prosecution's factual allegations while denying that he acted with the mental state the prosecution attributed to him. The strategy did not succeed: the trial judge declined to give a second-degree-murder instruction, a decision vindicated on appeal, 626 N.E.2d at 151-53, because Tenner's testimony established that he did not act under intense passion: preparation of the nooses showed that his acts were calculated. Moreover, the unreasonable-belief-in- justification defense is unavailable to aggressors, the court held. One who acts as a vigilante when it is easy to seek aid from the police cannot claim mitigation. Cf. United States v. Haynes, 143 F.3d 1089 (7th Cir. 1998).

Having lost the fight for a second-degree-murder instruction, Tenner switched grounds and insisted that the battle should not have been waged. He chastised his former lawyer for thinking that it was even conceivable that a lesser-included- offense instruction could have been obtained; and if that was impossible, then silence would have been the best defense. The Supreme Court of Illinois responded:

The State had overwhelming proof of the defendant's commission of the acts involved here, including the testimony of two persons present throughout the offenses. Given the strength of that evidence, counsel could have reasonably concluded that the best option was to attempt to portray the defendant's conduct as the result of sudden and intense passion resulting from provocation or an unreasonable belief in the need for self- defense, which, under either theory, would establish second degree murder. That the trial judge ultimately refused to instruct the jury on either form of second degree murder does not demonstrate that counsel acted unreasonably in pursuing this line of defense. Nor can it be said that the presentation of this theory was useless. By portraying the defendant's actions as the result of provocation or an unreasonable belief in self-defense, counsel could attempt to evoke sympathy for the defendant in the minds of the jurors, which might possibly have led to imposition of a sentence other than death. For these reasons, we do not believe that trial counsel was deficient in this choice of strategy, or that the defendant was prejudiced by it.677 N.E.2d at 866-67 (citation omitted). The state court thus found that counsel's performance was adequate and that there was no prejudice from any inadequacy. See Strickland v. Washington, 466 U.S. 668 (1984). We bypass the former possibility and say only that we agree with the latter: Tenner has not established prejudice.

Tenner and his lawyer were in a fix. Two dead bodies were found in Tenner's business premises. Ballistics tests established that the murder weapon was Tenner's. Nooses hanging from a beam showed preparation. Albert Sauls and Shirley Garza testified for the prosecution and related not only Tenner's acts but also the suffering the victims endured while confined more than two hours with their necks in nooses, waiting to die. What was counsel to do? Outright denial of the accusations would accomplish little other than lead the jury to conclude that Tenner was a liar as well as a murderer. Even a multiple murderer can fall in a jury's estimation (and thus take a greater risk of capital punishment) by coming across as an unrepentant, perjurious, multiple murderer. Was it not better to admit the obvious and give some explanation in the hope that at least one juror might think it good enough to avoid execution? Let us suppose that counsel fell below professional standards in thinking that a second-degree-murder instruction could be secured. Still, how was Tenner harmed by the effort? He had no defense other than his claim of justification. Tenner's current lawyers assume that a good lawyer must be able to find some resource for every accused, no matter what the evidence shows. Not so. Some crimes are so heinous, and so well documented, that there is little to be done for the defendant. This is one of those crimes.

Affirmed.