dissenting.
Because I believe the failure to address a jury with closing remarks in the sentencing phase of a capital ease is a fundamental error that leaves an irreparable taint on subsequent jury deliberations, I must dissent from the section of the majority opinion entitled "Penalty-phase summation waiver.” In that section, the majority denies Mr. Moore’s claim of ineffective assistance of counsel concerning the waiver of argument during the sentencing phase of his trial. Because I believe the waiver was an example of exactly the type of error that should be considered ineffective assistance of counsel, I would *1117grant the relief requested by Mr. Moore and send this case back for re-sentencing.
There is no such thing as a perfect trial, of course, and as a consequence, when reviewing claims of ineffective assistance of counsel, appellate courts afford attorneys a great deal of leeway in how they litigate cases.1 However, the devastating effect of some errors simply should not be overlooked. An attorney’s failure to argue for his client's life during the sentencing phase of a capital case is that type of error. The absence of argument for the defendant in this situation leads inevitably, I believe, to a breakdown of the adversary system and a flawed trial.2
The majority evaluates Mr. Moore’s claim of ineffective assistance of counsel under the standard articulated by the United States Supreme Court in Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). To be successful on a claim of ineffective assistance of counsel under that standard, a defendant must show “counsel’s performance was deficient” and “the deficient performance prejudiced the defense.” Id. at 687, 104 S.Ct. 2052. “The benchmark for judging any claim of ineffectiveness must be whether counsel’s conduct so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result.” Id. at 686, 104 S.Ct. 2052. I disagree with the majority as to whether this benchmark was met.
Prejudice
I will first address the second, or prejudice, prong of the Strickland test. To meet his obligations under this prong, Mr. Moore must show “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Strickland, 466 U.S. at 694, 104 S.Ct. 2052. The majority finds Mr. Moore’s ineffective assistance claim should fail, in part, because “there is no reasonable probability that, had Moore’s defense counsel given a closing argument, the jury would have chosen life over death.”3 In other words, the majority does not believe Mr. Moore has adequately demonstrated prejudice.
I do not think we can make a reasoned determination of “reasonable probability” of prejudice when confronted, as we are in this case, with an error by counsel that completely undermines the adversarial process. I think this question is different in kind from the usual one of determining whether an additional piece of evidence might have made a difference in the overall presentation. Here there was no presentation. How can a judge divine in any meaningful way whether the existence of an argument, as opposed to the absence of an argument, would change the mind’s of the jurors (or at least one juror)? The answer to such a question is pure conjecture. How can I even guess at the likely effect of what, for all practical *1118purposes, appeared to be a concession by Mr. Moore’s own attorneys that he deserved death? Perhaps the jury would have decided upon death no matter what Mr. Moore’s attorney argued, but perhaps most were leaning toward a life sentence and changed their minds when Mr. Moore’s attorney did not even appear to believe he could dispute the prosecution’s argument. There is no way to know, or, more importantly, to even make a meaningful guess. In such a circumstance, this test is no more than a charade. I believe the Strickland analysis breaks down in this context and, for this reason, I would presume prejudice4 when counsel waives argument during the sentencing phase of a capital case.5
The Supreme Court has acknowledged a difference exists when a defendant’s life is at stake. “In death cases doubts [with regard to the prejudicial effect of trial error] should be resolved in favor of the accused.” Andres v. United States, 333 U.S. 740, 752, 68 S.Ct. 880, 92 L.Ed. 1055 (1948). Furthermore,
the penalty of death is qualitatively different from a sentence of imprisonment, however long. Death, in its finality, differs more from life imprisonment than a 100-year prison term differs from one of only a year or two. Because of that qualitative difference, there is a corresponding difference in the need for reliability in the determination that death is the appropriate punishment in a specific case.6
Woodson v. North Carolina, 428 U.S. 280, 305, 96 S.Ct. 2978, 49 L.Ed.2d 944 (1976). Although Strickland was itself a death penalty case, it did not involve the same type of fundamental error as is found in this case. For this reason, its analysis does not fit.
In addition, even if I were not to assume prejudice, I would be unconvinced by the majority’s analysis. In support of its decision, the majority points to the admission during the sentencing phase of strong evidence of Mr. Moore’s guilt from the guilt phase, evidence of a previous robbery conviction, and evidence he once attempted to use a firearm during an altercation. The majority then concludes, based on the relative merit of the evidence presented by each side, there is no reasonable probability a closing argument would have affected the result. I simply do not understand how one gets to the conclusion reached by the majority, given the supporting evidence listed.7
I do not doubt Mr. Moore’s guilt. The murder he committed was horrendous, but it does not strike me as an obvious and unquestionable candidate for the death penalty. The murder was not random — Mr. Moore knew the victim and had several altercations with him. The murder did not involve multiple victims. The murder did not involve torture. Indeed, the only unusual aspect of the murder is the number of times Mr. Moore shot his victim in the head. As aggravating factors, the state merely offered Mr. *1119Moore’s previous conviction and an example of his apparent predisposition to violence.8 Altogether, this strikes me as a typical example of what has become a far too common crime. It most certainly does not strike me as the type of murder for which the proper punishment is so decidedly certain I would be “unable to envision” a result other than death.9 Put another way:
Balancing the factors — mitigating and aggravating — involves the exercise of judicial discretion and judgment. When such balancing is required, it is the duty of defense counsel to argue why the balance should be struck in his client’s favor. No matter how great the odds may seem to be against winning the argument, counsel must make the best case he can. It is not counsel’s role to act as judge as well as advocate, any more than it is the district court’s or this court’s role to say, “Oh well, his client would have lost anyway. He had a bad argument, so there’s no need for him to make any argument at all — no need to advocate — no need for representation by counsel.” That kind of reasoning is squarely contrary to the elementary precepts of the adversary system.
Gerlaugh, 129 F.3d at 1049 (Reinhardt, J., concurring in part and dissenting in part).
I also believe the tone of the.majority discussion evinces a misunderstanding of what I believe to be the most significant problem with a waiver of argument in the sentencing phase of a capital case. The real problem is not what has gone unsaid,10 as the majority suggests, it is the potentially devastating effect of saying nothing at all. Simply put, a waiver of the argument in the sentencing phase is a concession the defendant deserves to die. See Smith v. Stewart, 140 F.3d at 1269 (suggesting failure to argue may be worse than failure to present mitigating evidence at sentencing hearing); Myles, 389 So.2d at 30-31 (reversing death sentence for tepid closing argument). I can think of no stronger signal to a jury that a defendant’s life is not worth sparing, than the failure of anyone to speak for him. When nobody in the courtroom stands -up‘and argues for the defendant’s life, it devalues the jury system, the court, and life itself.
*1120Under Strickland, we determine whether there is a “reasonable probability that, absent the errors, the senteneer ... would have concluded that the balance of aggravating and mitigating circumstances did not warrant death.” Strickland, 466 U.S. at 695, 104 S.Ct. 2052. For most types of mistakes attorneys make, this is an appropriate standard of review. Here, however, we had a trial where the prosecution got to stand up before the jury not just once, but on two separate occasions, without opposition, and argue for putting Mr. Moore to death. As far as the jury knew, Mr. Moore’s own attorney could think of no answer to the prosecution’s arguments. I suggest there is nothing a court can do to evaluate the effect this had on the jurors’ minds, and it simply makes no sense to talk in terms of “reasonable probabilities.” 11
I raise this not out of concern for the rights of cold-blooded murderers like Mr. Moore, but out of concern for the health of the judicial system. I am troubled that our review of cases in some areas essentially has become no review at all. In the context of this issue in this case, the standard of “reasonable probability” is no standard at all. I can think of no situation where we could say there is a “reasonable probability” the outcome would have differed had counsel given a closing argument or there is no “reasonable probability” that the outcome would have differed. Either way, we engage in pure conjecture. We could admit the folly of this type of review and decline to provide review in this context. But I think the better choice is to provide meaningful review and assume prejudice in situations like this one. Requiring counsel to “present whatever argument can be made to counter the state’s request that [the defendant] be sentenced to death ... simply reflects a basic tenet of the adversarial process — that the defendant’s interests be represented before the court.” Gerlaugh, 129 F.3d at 1046 (Reinhardt, J. concurring in part and dissenting in part).
Representation
Under the representation prong of Strickland, we must determine if the “counsel’s performance was deficient.” Strickland, 466 U.S. at 687, 104 S.Ct. 2052. “This requires showing that counsel made errors so serious that counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth Amendment.” Id. Essentially, this means a complaining defendant must show “counsel’s representation fell below an objective standard of reasonableness.” Id. at 687-88, 104 S.Ct. 2052. Courts “indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.” Id. at 689, 104 S.Ct. 2052.
Argument is deemed such an important part of any trial it is afforded constitutional protection. “A defendant’s right to effective assistance of counsel includes the right to present closing argument.” United States v. Stenzel, 49 F.3d 658, 661 (10th Cir.1995) (citing Herring v. New York, 422 U.S. 853, 865, 95 S.Ct. 2550, 45 L.Ed.2d 593 (1975)), cert. denied, 516 U.S. 840, 116 S.Ct. 123, 133 L.Ed.2d 73 (1995). This right may, however, be waived. Id. So the ultimate question is whether Mr. Moore’s attorneys were acting within the “range of reasonable professional assistance” when they encouraged him to waive this important right.
When reviewing for ineffective assistance of counsel, courts must give wide latitude to trial lawyers in the tactics they choose to employ. See Strickland, 466 U.S. at 689-90, 104 S.Ct. 2052. As the majority notes, the district court found “that the conscious, tactical choice of the defense in waiver of closing argument [wa]s to be accorded deferential *1121scrutiny and that the tactic was not unreasonable.” The majority agreed, finding the decision to waive closing argument to be nothing less than an informed strategic choice. There is no real question about whether the decision to waive argument in the sentencing phase of Mr. Moore’s trial was in some sense “tactical;” but, in my mind, there is also no question it was unreasonable. It was tactical to the extent Mr. Moore’s attorneys considered the question and decided to waive argument in order to gain what they thought would be an advantage over the prosecution. By waiving argument, they apparently intended to prevent the prosecution from making its final closing arguments in the form of its rebuttal. This decision, and the way in which it was carried out, however, was so misdirected that it falls well outside the “wide range of reasonable professional assistance” we allow.12
Frankly, I can think of no worse decision than that made by Mr. Moore’s attorneys in this case. Faced with a tough case, they decided their best course of action was to give up in the vain hope the other side would have to stop. This is closer to a complete retreat than a tactic. I agree with the Ninth Circuit that the failure to argue for “leniency ‘amount[s] in every respect to no representation at all.’ ” Smith v. Stewart, 140 F.3d at 1269 (quoting Clabourne v. Lewis, 64 F.3d 1373, 1387 (9th Cir.1995)) (dealing with a case in which no evidence was presented and no argument was made). Years ago a justice on the Louisiana Supreme Court found it “difficult to imagine circumstances under which a defense attorney may render effective assistance in a death penalty proceeding by waiving closing argument.” State v. Smith, 400 So.2d 587, 595 (La.1981) (Dennis, J., concurring). I agree, and the facts of this cáse do not ignite my imagination in this regard. :
Furthermore, this “tactical” decision was premised on a questionable interpretation of state law that was unsuccessful in trial and on appeal. Mr. Moore was totally deprived of any opportunity to persuade the jury to spare his life, in the face of two arguments by the prosecution, because his attorneys miscalculated the effect of their decision under Oklahoma law. In my mind, the question of whether that decision can be characterized as legitimate is completely foreclosed by their failure to prepare a closing just in case their risky tactic failed, as it ultimately did.13 Put simply, at the time Mr. Moore’s counsel made this decision, they were failing to act as “the ‘counsel’ guaranteed ... by the Sixth Amendment.” Strickland, 466 U.S. at 687, 104 S.Ct. 2052. Our system allows for poor decisions by trial attorneys; it does not, however, allow for tragically wrong-headed ones.
In conclusion, I would find counsel’s performance in this case to be deficient. Counsel’s decision to waive argument during the sentencing phase denied Mr. Moore his day *1122in court and denied the jury an opportunity to weigh the aggravating and mitigating circumstances on a fully informed basis. This is a breakdown in the adversary process that, I believe, renders the resulting death sentence unreliable. See Strickland, 466 U.S. at 687, 104 S.Ct. 2052. In other words, this failure by Mr. Moore’s attorneys so .undermined the basic foundation underlying our trial process, I would assume it prejudiced the defense. Accordingly, I would remand this case for re-sentencing.
. The waiver was not the only serious flaw in this trial. Remarkably, the trial court sent the jury out to begin its penalty-phase deliberations at 9:55 p.m. Although this was not contested as reversible error in this case, I am gravely concerned a judge would fail to consider the effect this may have on the jury's deliberations, especially when a man’s life hung in the balance. The jury came back with a sentence of death at 11:25 p.m.’ '
.The majority cites with favor a Fifth Circuit rule that in order to establish ineffective assistance of counsel for failure to give a closing argument "a habeas petitioner must make some type of showing of what defense counsel might have said at closing that would have had a reasonable probability of changing the result.” This is not a rule I would embrace, for it serves no useful purpose. Any one of us could compose an argument to spare a man's life. The exact form of such an argument is beside the point; what is critical is that someone makes the argument. "To put it starkly, counsel’s failure [to argue] was a virtual admission that the death penalty should be imposed upon ,his client.” Smith v. Stewart, 140 F.3d 1263, 1270 (9th Cir.1998). I see no reason for imposing this requirement unless we intend to adopt the ridiculous approach of granting or denying appeals based on the rhetorical force of proposed arguments.
. Considering the same question, Judge Reinhardt, of the Ninth Circuit, concluded the “failure to make a closing argument during the penalty phase of the trial denie[s] [a defendant] representation by counsel at a critical stage of the sentencing proceeding in violation of the basic constitutional principles” found in United States v. Cronic, 466 U.S. 648, 104 S.Ct. 2039, 80 L.Ed.2d 657 (1984). Gerlaugh v. Stewart, 129 F.3d 1027, 1049 (9th Cir.1997) (Reinhardt, J., concurring in part and dissenting in part). While I take a somewhat different approach, I agree with the ultimate conclusion that the failure to make a closing in such a proceeding is an error so fundamental it cannot be properly analyzed under the Strickland standard.
. I would draw a distinction between the failure to argue during the guilt phase and during the sentencing phase. "The overwhelming reality of certain cases may make a finding of guilt inevitable. That inevitability, however, is never the case at the penalty hearing. At that stage, there is no set of facts and no set of legal principles that in any given case requires a rational person to vote for the death penalty.” Hopkinson v. State, 632 P.2d 79, 195 (Wyo.1981) (Rose, C.J., dissenting in part and concurring in part) (quoting other sources), cert. denied, 455 U.S. 922, 102 S.Ct. 1280, 71 L.Ed.2d 463 (1982).
. "From the point of view of the defendant, [the death penalty] is different in both its severity and its finality. From the point of view of society, the action of the sovereign in taking the life of one of its citizens also differs dramatically from any other legitimate state action.” Gardner v. Florida, 430 U.S. 349, 357-58, 97 S.Ct. 1197, 51 L.Ed.2d 393 (1977) (plurality opinion).
. Furthermore, a closing argument is far more than the sum of the proceeding evidence. Limiting the possible persuasive value of an argument to that of the evidence presented evidences a misconception of the value of argument itself.
. And recall, the "especially heinous, atrocious or cruel” aggravating circumstance found by the jury was struck as invalid in this case on direct appeal. Moore v. State, 809 P.2d 63, 65 (Okla.Crim.App.1991), cert. denied, 502 U.S. 913, 112 S.Ct. 313, 116 L.Ed.2d 255 (1991).
. “Given the vast spectrum of individual opinions and feelings about the appropriateness of the death penalty, and given the basic humanity of every defendant no matter how heinous his crime, there is in eveiy case a rational basis for a juror to vote for life imprisonment.” Hopkinson, 632 P.2d at 195 (Rose, C.J., dissenting in part and concurring in part) (quoting other sources). See also Goodpaster, Gary, The Trial For Life: Effective Assistance of Counsel in Death Penalty Cases, 58 N.Y.U. L.Rev. 299, 300 (1983) (describing case of Bernardo Sierra, who committed twelve robberies, two maimings, and three murders within eight hours and did not receive death sentence). One need only consider the outcome of Oklahoma bombing defendant Terry Nichols’ trial to doubt there are any cases where imposition of the death penalty is a sure bet.
. I do not want to minimize the problems inherent in what went unsaid in this case. That too is a significant problem. One court described the types of statement an attorney should have made in a similar circumstance:
He did not ask the jury to spare the defendant's life. He did not remind the jury that [the defendant] is a human being or urge the jurors to be mindful of their awesome responsibility in deliberately choosing whether he should live or die. Nor did he emphasize to the jurors any of their legal obligations designed to prevent the arbitrary or capricious imposition of the death penalty, e.g., the requirement that they base their findings upon a beyond a reasonable doubt certainty; their duty to weigh any aggravating circumstance found against any and all mitigating circumstances; the duty of each individual juror to hold fast to his honest convictions and to vote to prevent a unanimous verdict in the event he is convinced that the death penalty is inappropriate.
State v. Myles, 389 So.2d 12, 30 (La.1979) (on rehearing). Furthermore, in this case, defense counsel did not even reiterate the evidence of mitigating factors presented during the sentencing phase. I do not see how this court can measure the probable effect such statements would have had on the jurors. Some types of errors cut so closely to the core of the jury decision making process that application of "reasonable probability” analysis strikes me as an absurd exercise.
. "To attempt to assess the effect of error in this legal vacuum [of what affects a juror in the sentencing phase] is to superimpose one untestable, surmise upon another. We must not pile conjecture upon conjecture and posit the decision of life or death upon a pyramid of guesses.” People v. Terry, 61 Cal.2d 137, 37 Cal.Rptr. 605, 390 P.2d 381, 392 (1964), cert. denied, 379 U.S. 866, 85 S.Ct. 132, 13 L.Ed.2d 68 (1964). I believe it draws into question the fundamental fairness of the process when courts affirm death sentences on the basis of conjecture as to how twelve jurors were affected by a breakdown in the adversary system. Through its death penalty jurisprudence, the Supreme Court has created a system whereby juries in capital cases essentially make life and death decisions for any and all reasons. It is therefore illogical, to me, to attempt to determine the probable effect of an egregious error, like the failure to make a closing argument, because we can never know the basis for the jury’s initial decision.
. The majority declines to consider Mr. Moore's new claim that the use of waiver as a tactic was prompted by an initial failure by one of his attorneys to prepare a closing. I too ignore this new claim for purposes of my analysis. I do not think we have to go this far (essentially abandonment by counsel) to find ineffective assistance in this case. Forgoing a critical, constitutionally protected part of trial where the defendant’s life is at stake, on the basis of a questionable interpretation of the law, in order to gain a dubious (nonexistent I would say) advantage, without taking any precautionary measures in case the plan fell through, and without, at least, making a spontaneous argument before the prosecution's rebuttal, once the judge made it clear the plan would not be successful, is far enough.
To be fair, defense counsel requested an opportunity to give an argument after the prosecution concluded its “rebuttal.” However, such a last-minute attempt to correct for their error does not obviate it. Upon learning of defendant’s intention to waive argument, the court held a conference in chambers to ensure the waiver was knowing and informed. During that conference, the judge indicated he would allow rebuttal if the defense waived argument. Once back in the courtroom, the judge asked the defense one last time if it intended to waive argument.. At this point, although they were unprepared, counsel should have requested to argue (as they eventually did). As their "tactical" goal had failed, anything would have been better than nothing. It makes no sense to waive at this point, and then request argument later when there are no legal grounds to support the request.
. It could be argued defense counsel misunderstood the legal effect, under Oklahoma law, of a defense waiver of argument on the right of the prosecution to present rebuttal (as both the trial and appellate courts ruled against counsel’s position). At best, one could say counsel failed in an attempt to create a new rule.