dissenting in part:'
Today’s court decides that Defendant’s trial counsel performed so weakly at the sentencing phase that Defendant did not have counsel as required by the Constitution. I do not think so.
Trial counsel based his presentation at the sentencing phase largely on residual doubt over Defendant’s guilt. This case is essentially the same as Tarver v. Hopper, 169 F.3d 710, 715 (11th Cir.1999), in which we concluded that counsel was not ineffective. I believe that today’s decision cannot be squared with Tarver. Because Tarver is the preexisting law of the circuit, today’s decision seems not just mistaken, but also raises worries about full adherence to binding precedent.
Defendant argues that his counsel inadequately prepared for sentencing, spending fewer than 24 hours — the time between conviction and the start of the sentencing phase — and that counsel failed to investí-, gate and to present sufficient character mitigation evidence. This argument is basically the same as the one made by the defendant in Tarver, who claimed that his trial counsel had only spent four hours preparing for sentencing between conviction and sentencing and had failed to investigate and present sufficient character mitigation evidence. We rejected defendant’s claims in Tarver, concluding that counsel reasonably chose to focus his time, energy and efforts on acquittal and, if necessary, on residual doubt at sentencing. Tarver’s lawyer’s acts were not found to be ineffective; but here the court comes down the other way, saying Chandler’s lawyer was ineffective.
In Tarver, this court concluded that the four-hour characterization was “ ‘inaccurate and misleading,’ because of the overlap in preparation for the sentencing and guilt/innocence stages of the trial.” Tarver, 169 F.3d at 715. Likewise, the claim of fewer than 24 hours in this case is inaccurate and misleading. In both cases, the lawyer had met with potential witnesses before and during the guilt phase but focused the limited time and resources on acquittal at trial and, then, on residual doubt at sentencing.
In this case, trial counsel interviewed 67 witnesses during his investigation leading up to trial,1 many of whom were familiar with Defendant’s reputation and were located in Defendant’s home town: the rural *1311Piedmont/Easom Springs area. Trial counsel stated that he conducted these interviews making “mental note[s]” of anything that might be of value later. We have pointed out in the past, that “[a] lawyer’s time and effort in preparing to defend his client in the guilt phase of a capital case continues to count at the sentencing phase.” Tarver, 169 F.3d at 715.
Here, trial counsel did offer the testimony of Defendant’s wife and mother at sentencing to show Defendant’s humanity and the stability of family.2 In Tarver, the lawyer had presented defendant’s uncle. To show evidence of defendant’s prior good works, defense counsel in this case already had some evidence from the guilt phase: that Defendant had given a soon-to-be married man the deed to a lot on which the man could build a house for his new family. And it was stipulated that Defendant had no criminal history.
Defendant’s counsel did not call other character witnesses in mitigation because he (1) doubted whether character evidence would successfully mitigate the crime and (2) was afraid of opening the door to harmful cross-examination and rebuttal witnesses. Defense counsel believed that character evidence inherently would not help much and testified: “it would be at least questionable whether a sufficient impact of character type testimony could overcome a fixed opinion based on other evidence ... [whether it] could change it from life to death. Or death to life.” [R13-^00-01, 365].3 I cannot say this view of how best to proceed is unreasonable.
*1312Trial counsel also testified that based on his interviews he found that “some individuals in the community considered [defendant] to be a drug dealer” and “that there were people in the community [who] were afraid of him.” [R13-433, 399]. Given these facts, it was reasonable for trial counsel not to investigate further personally, especially given the knowledge that he already had other credible mitigation arguments. See Lambrix v. Singletary, 72 F.3d 1500, 1506 (11th Cir.1996) (“[Counsel cannot be held responsible for failing to find mitigating evidence if, after a reasonable investigation, nothing has put counsel on notice of the existence of that evidence.”); see also Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 2066, 80 L.Ed.2d 674 (1984) (“In other words, counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary”); Williams v. Head, 185 F.3d 1223, 1237 (11th Cir.1999) (“to be effective a lawyer is not required to pursue every path until it bears fruit or until all hope withers ... [a] decision to limit investigation is accorded a strong presumption of reasonableness”).4 Highly important, our court already has “rejected the position that strategic decisions can be considered reasonable only if they are preceded by a thorough investigation ... we explained that the correct approach toward investi*1313gation reflects the reality that lawyers do not enjoy the benefit of endless time, energy or financial resources.” Williams, 185 F.3d at 1237 (internal citations and quotations omitted); see also Rogers v. Zant, 13 F.3d 384, 386-88 (11th Cir.1994) (decision not to investigate need only be reasonable to be competent).
Trial counsel in Tarver chose to focus his effort and resources on the guilt phase of the trial “based on his assessment of the likelihood of acquittal.” Tarver, 169 F.3d at 712. The evidence supports the finding that counsel in this case also decided to focus on the guilt phase of the trial. Counsel’s acts at trial are evidence of his intent. As for counsel’s words at the post-conviction proceedings, not only did he testify that time constraints were such that this course is on what he chose to focus, but he thought he had a good case for innocence and consulted a jury selection expert to that effect.5 [R-13-396, 360-61, 328], This decision about how to invest the lawyer’s time and energy is entitled to deference by this court. The evidence of guilt in this case was, by no means, overwhelming: Defendant did not, in fact, kill the victim; the issue was whether the person (Jarrell) who did the killing did so for Defendant or because the actual killer was settling his own score. As in Tarver, I conclude that “the decision to focus on the acquittal of a capital murder was not unreasonable” and does not amount to inadequate preparation for sentencing. Tarver, 169 F.3d at 716.
Despite trial counsel’s decision not to rely much on character evidence,6 trial counsel did present a robust case for mitigation. In addition to the humanizing testimony of family, he highlighted this mitigation evidence: (1) Defendant had no prior criminal record; (2) the triggerman, Charles Ray Jarrell, would not receive the death penalty; and (3) lingering doubt existed of Defendant’s guilt. We have vindicated lawyers who did less. See, e.g., Waters v. Thomas, 46 F.3d 1506, 1511 (11th Cir.1995) (en banc) (“the Supreme Court *1314and this Court in a number of cases have held counsel’s performance to be constitutionally sufficient when no mitigating circumstance evidence at all was introduced, even though such evidence ... was available”); Francis v. Dugger, 908 F.2d 696, 702-04 (11th Cir.1990) (not ineffective to decide to argue value of life and compassion rather than present mitigation testimony about defendant’s poor childhood and mental dysfunction).
When I look at (among other things) the trial transcript, I see that trial counsel in this case — as in Tarver — had a central theme from start to finish. He sought to raise and to rely on doubt, both at trial and at sentencing, about Defendant’s guilt in fact. As we stressed in Tarver, lingering doubt on guilt is “an effective strategy for avoiding the death penalty.” Tarver, 169 F.3d at 715 (noting that scholars— their articles were cited — have presented evidence that raising doubt about guilt is the most effective method of avoiding death penalty).7
The evidence was uncontroverted that someone (Jarrell) other than Defendant had shot the victim outside of Defendant’s presence. At issue was whether the killer had acted on Defendant’s behalf. Trial counsel contended at sentencing that the killer was not motivated by Defendant to kill the victim, stressing that the killer had attempted to kill the victim before; that the killer had had 23 beers that day before shooting the victim; that a history of personal animosity existed between the victim and his actual killer; that the killer’s statements after being arrested were inconsistent; and that a recorded damaging statement that Defendant had to “kill somebody” was made almost three months after the killing took place and did not indicate that Defendant was speaking about the victim in this case. [R15-27, 59-60],
Counsel argued that it would be a mistake and “cruel and unusual punishment” to give Defendant the death penalty considering these circumstances. I am convinced the argument had force and was a valid line to take. Of course, this argument was ultimately unsuccessful; but so are roughly half of all arguments made in American courtrooms. Lack of ultimate success has almost nothing to do with whether the lawyer met his professional obligations as an advocate.
The evidence shows that trial counsel’s conduct at sentencing, in this case, was “within the broad range of reasonable performance we have recognized in other cases.” Tarver, 169 F.3d at 715 (citing cases from this Circuit); see Bolender v. Singletary, 16 F.3d 1547, 1558 (11th Cir.1994) (counsel not ineffective for deciding to forgo arguing defendant’s troubled history and instead arguing at sentencing that defendant should not be treated more harshly than codefendants); Porter v. Singletary, 14 F.3d 554, 556-59 (11th Cir.1994) (not ineffective for failing to present mitigating evidence of difficult background and family relationships); Singleton v. Thigpen, 847 F.2d 668, 670 (11th Cir.1988) (performance not constitutionally deficient even though mitigation efforts consisted of asking mother and wife to identify possible witnesses); see also Elledge v. Dugger, 823 *1315F.2d 1439, 1445 (concluding counsel to be ineffective because of total failure to interview defendant’s relatives where defendant had confessed to crime and when counsel knew mitigation was only defense), partially withdrawn on other grounds, 833 F.2d 250 (11th Cir.1987).
Today’s court gives too much weight to trial counsel’s testimony that he did “basically not anything explicit” to prepare for the sentencing phase. Not what is said by a lawyer in post-conviction proceedings, but what the lawyer did at the trial is the more important. If a trial lawyer preformed adequately at the trial, his unartful explanation years later about that performance does not convert adequate performance at trial into inadequate performance at trial. Furthermore, particularly given the defense counsel’s central theme— doubt about guilt in fact — no clean, clear line separated preparing for the guilt phase and preparing for the sentencing phase. In such cases, even if a lawyer did nothing just for the sentencing phase in particular, it does not follow that he earlier did not do a lot that was useful at sentencing as well as at the guilt phase.
In Tarver, counsel said that he had not adequately prepared for sentencing. In Tamer, however, we recognized that merely because trial counsel did not label his act of concentrating on getting an acquittal as a strategic decision did not negate the evidence that his acts were deliberate. See also Wright v. Hopper, 169 F.3d 695, 707 (11th Cir.1999) (deciding that even though defense counsel denied making strategic decisions about whom to investigate, evidence indicated that he “admitted to the process of making strategic decisions” because he admitted that he had other avenues to pursue concerning the remainder of the defense and also had to make choices in defending a client based on time, information, and what his client told him). Counsel’s acts at trial as well as counsel’s words at the habeas hearing are evidence. Furthermore, this court has repeatedly rejected relying upon counsel’s own admissions of deficient performance: “Because ineffectiveness is a question which we must decide, admissions of deficient performance by attorneys are not decisive.” Wright, 169 F.3d at 707 (quoting Harris v. Dugger, 874 F.2d 756, 761 n. 4 (11th Cir.1989)).
Something different could always have been done in these cases. But the trial of a murder case, like all litigation, requires the lawyer to decide — to make hard choices — what is the best use of his inherently limited time and energy. A lawyer can make a reasonable decision that, no matter what an investigation might produce, he wants to steer clear of a certain course: for example, not to rely on character evidence as the main pillar of mitigation. See generally Rogers, 13 F.3d at 386-87. Because trials involve imponderables, much room always exists after the fact for debate about the wisdom of the lawyer’s choices. But we know that counsel need not investigate or present all available mitigating evidence. Waters, 46 F.3d at 1511; Rogers, 13 F.3d at 387 (“ ‘strategy’ can include a decision not to investigate”). More is not always better; for example, there can be a diluting impact when too many witnesses or arguments are used, the weaker points obscuring and confusing the main point.
Courts are required to presume strongly that defense counsel performed reasonably and adequately. And I conclude the choices evidenced by the record here— particularly the choice to work for an acquittal and then to focus on lingering doubt about guilt at sentencing and not to clutter the case with mediocre character witnesses' — were well within the wide range of what is reasonable. The constitutional requirements were met. Defendant has the burden of persuasion to warrant post-conviction federal relief; and he has not carried it.
Today’s, court decision is out of step with the preexisting law of the circuit: not least — but not just — Tarver. It bothers me. Defense counsel in Tarver, like the lawyer in this case, concentrated on acquittal and residual doubt. We determined that his acts were not ineffective; but today’s court says that counsel in this case was. Cases that are essentially similar— Tarver and this one' — ought to have the same result.
*1316I would affirm the judgment of the district court.
. Trial counsel had met with the Defendant frequently and had the benefit of Defendant’s *1311insight about what might be available. The lawyer interviewed Defendant's brother, wife, mother, former minister, and members of the community: some of the same kinds of persons interviewed in Tarver. For all that we know, trial counsel here interviewed every witness that his client ever suggested; at the evidentiary hearing on habeas corpus, neither side asked the trial lawyer whether his client had suggested witnesses to be interviewed for mitigation or otherwise.
. The lawyer testified that he wanted to show, for example, with testimony about building their own houses from natural materials gotten from their own land, that these "were people who did not intrude themselves on others and who were peaceful people.” The lawyer argued in closing at sentencing that Defendant’s was a family with “tremendous stability,” that Defendant was a man who was skilled with his hands, and all of this demonstrated "lives with some purpose.” Putting up the wife and mother also called to the jury’s attention, of course, that these innocent people would be punished too by the execution of their husband and son.
. The evidence in this record indicates that trial counsel decided to stress lingering doubt about guilt, because he did not think character evidence — even the testimony of good acts offered by Defendant at the evidentiary hearing — would prevent the jury from giving Defendant the death penalty if they were absolutely sure he had procured this murder:
Q:And that specific instances of the defendant helping people is probably the most compelling of that type of evidence, a mere opinion that he’s a good person is one thing, but to say he did this for me is even better. You recognized that, did you not?
A: Well, I recognize its admissibility as a compelling nature, I wouldn’t test it.
Q: You didn’t think that type of testimony would be very helpful, is that your testimony?
A: Well I’m — if that's your question, I say in the whole picture of things as it existed at that time, if that jury was going to give him the death penalty based on what they had heard, their minds were made up as to that, I doubted that compelling would be an appropriate word.
[R13-365].
The conclusion that Defendant's attorney chose to pursue lingering doubt rather than character evidence is also shown by the following passage:
Q: Prior to being able to make a determination of how compelling a witness would be in mitigation you would have to know what the witness was going to say, is that correct?
A: I think that I could assume what a piece of testimony would be, for example.
Q: Right.
A: And assume the most favorable testimony that you might get and then form some judgment, not the most reliable judgment in the world, but some judgment about how compelling it might be.
[R13-368].
From the testimony, it appears that Defendant’s attorney thought that even the best character mitigating evidence would not be *1312that compelling given the facts of this case: hence the decision to pursue the strategy of lingering doubt. (And, therefore, no need to request a continuance to look for more character witnesses). This path is not ineffective given the facts of the case against the Defendant, particularly considering that we are to view counsel's performance, not in hindsight (we now know what was done, at trial, was not successful), but “from counsel’s perspective at the time.” Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 2065, 80 L.Ed.2d 674 (1984).
By the way, the decision to pursue lingering doubt was made by an experienced and respected member of the Alabama bar — a lawyer with over 40 years of experience in civil and criminal litigation who had tried over 1000 cases. "Our strong reluctance to second guess strategic decisions is even greater where those decisions were made by experienced criminal defense counsel.” Provenzano v. Singletary, 148 F.3d 1327, 1332 (11th Cir.1998).
. Based on the information he had about what some of the community thought of the Defendant and the fact that the law enforcement community in Piedmont was hostile to Defendant, trial counsel testified that not only did he fear opening the door to unfavorable character evidence, but he also feared that the government would offer damaging rebuttal witnesses. [R13-399, 416]. Among other things, he testified that deciding whether to pul on character witnesses when there was a possibility of the government cross-examining on various bad acts was a balancing act and that he did consider who and what might have been produced. [R13-416].
Mainly stating that after the guilt phase the jury would not be surprised to hear that some people thought Defendant was a drug dealer and that nothing in the record reveals what new damaging evidence would have come to light, today's court opinion scorns the decision not to focus on character evidence more. I submit, however, it was not unreasonable for trial counsel to be wary of provoking the prosecution to present a string of new witnesses testifying to Defendant's bad reputation and bad acts: evidence which would reflect and solidify any negative opinion the jury already had of the Defendant generally. Trial counsel knew he had other mitigating elements that he could (and did) advance.
More important, today's line of observation — from the vantage of hindsight — by judges on how best to proceed with the defense to the death penalty is, by its nature, troubling: the Supreme Court has said that it is not for this court to second-guess counsel’s decision, but just to decide whether his performance was constitutionally deficient. See Collier v. Turpin, 177 F.3d 1184, 1200 (11th Cir.1999) ("Although we might think that it would have been more prudent for [defendant's] attorneys to investigate other possible avenues for obtaining information about [defendant’s] character and background, ... 'in considering claims for ineffective assistance of counsel, "[w]e address not what is prudent or appropriate, but only what is constitutionally compelled.” ' ” (quoting Burger v. Kemp, 483 U.S. 776, 107 S.Ct. 3114, 3126, 97 L.Ed.2d 638 (1987) and United States v. Cronic, 466 U.S. 648, 104 S.Ct. 2039, 2050 n. 38, 80 L.Ed.2d 657 (1984))). And we must recall, we have already said that there is no "duty on defense counsel to present general character evidence in every capital case.” Stanley v. Zant, 697 F.2d 955, 961 (11th Cir.1983).
. He testified that "you do what's coming up first, and you do what's immediately on you” [R13-395]; that he did not have "that much time to spend” [R13-398]; and that he thought — quite rightly, I believe — Defendant had a realistic chance of not being convicted [R13-330, 360-61],
. “The mere fact that other witnesses might have been available or that other testimony might have been elicited from those who testified is not a sufficient ground to prove ineffectiveness of counsel.” Waters v. Thomas, 46 F.3d 1506, 1514 (11th Cir.1995) (en banc).
In addition, as the district court noted, the witnesses Defendant offered at the evidentiary hearing for post-conviction relief had little or no knowledge of Defendant’s criminal acts. Their ignorance of this part of Defendant's life would significantly minimize the "humanizing effect” of these character witnesses and bring into doubt their reliability. See McQueen v. Scroggy, 99 F.3d 1302, 1314 (6th Cir.1996). On cross-examination at the evi-dentiary hearing, these witnesses testified that knowledge of Defendant's threats to kill and turning a gun on a law enforcement officer would not affect their assessments of Defendant’s character. The district judge — who personally not only heard the witnesses' testimony, but observed their demeanor — found that this "bias” “largely nullified” the value of their testimony: the jury would be unlikely to credit "witness[es][who] believed that drug dealing and violent crimes were irrelevant to a person’s character."
The decision not to call a witness is ineffective ultimately only if it can be said that every reasonable lawyer informed about this witness would have called him. See Rogers v. Zant, 13 F.3d 384, 386 (11th Cir.1994) ("no relief can be granted on ineffectiveness grounds unless it can be shown that no reasonable lawyer, in the circumstances, would have done so”); White v. Singletary, 972 F.2d 1218, 1220 (11th Cir.1992) ("We ask only whether some reasonable lawyer at trial could have acted, in the circumstances, as defense counsel acted”). I think many reasonable trial lawyers would decide not to use these witnesses.
If not every reasonable lawyer would have called a certain witness or witnesses (even among lawyers who were completely informed about the witness or witnesses), I fail to understand why we would order a new trial on the ground that the witness or witnesses were not called in the first trial: at the new trial, the witnesses again might not be called. If the new trial can be just like the first trial, how can the first trial be constitutionally inadequate while the new trial is constitutionally adequate?
. That such residual doubt can easily exist is not just a figment of a lawyer's imagination. The jury in this case was instructed at the guilt phase this way: "[I]t is not necessary that the defendant’s guilt be proven beyond all possible doubt. It is only required that the government's proof exclude any reasonable doubt concerning the defendant's guilt.” As this instruction shows, the law recognizes that jurors who have found a person guilty of a crime may well still have doubt about his true guilt. Thus, the law itself points to and lays the foundation for a good argument based on this lingering doubt when the jury is later asked to impose death, the ultimate and most irremediable punishment. In my view, nothing about the trial lawyer's advancement of this argument signals submissiveness or fatalism; stressing residual doubt is a straightforward and sound defense. And if we know with hindsight that this argument did not prevail, we do not know that an argument based chiefly on character witnesses would have proved to have been one bit more successful.