DIANE P. WOOD, Circuit Judge.
Sherman Howard was convicted in 1989 of sexually abusing his daughter when she was a toddler. After appealing through the Illinois state court system and seeking post-conviction relief in Illinois’s courts, he filed this habeas corpus petition. He contends that his appellate counsel was constitutionally defective in her selection of issues to appeal, and he argues that the district court abused its discretion by refusing to appoint counsel to represent him in these proceedings. The district court rejected Howard’s arguments, concluding that his counsel’s performance was inadequate, but that he suffered no prejudice as a result. We agree that Howard has not shown prejudice, so we affirm.
I
Tamika Howard was born in 1981 into an extraordinarily turbulent and traumatic home. Her mother was addicted to drugs and frequently left the home. During her mother’s absences, Tamika was occasionally alone with her father, Sherman, and sometimes stayed with her aunt, Linda Fletcher. Between April 1982 and May 1983, Tamika lived with her aunt, and Howard visited her on weekends. During that period, Fletcher noticed swelling and a discharge near Tamika’s genital area. Concerned, Fletcher starting asking questions. Tamika’s response left no doubt in her aunt’s mind that her father had been sexually abusing her. Tamika (using childish language) described to Fletcher how Howard would come into her room, expose himself, undress her, and attempt to have intercourse with her. Moreover, he had her perform oral sex on him and fondled her genitals. Finally, Fletcher says that Tamika told her that Howard would give her strange pink and yellow pills and allow her to drink beer, though Fletcher admitted that she never smelled
alcohol on Tamika’s breath during Howard’s visits.
Fletcher tried to take Tamika to a doctor, but she was initially unsuccessful. Medical personnel told her that because she was not the child’s legal guardian, she could not use public emergency services; private help was not feasible, because she had no means of paying for it. Finally, she took Tamika to Dr. Simon Rosen and represented herself as Tamika’s mother. Dr. Rosen examined Tamika and found that her hymen was not intact. There was an inquiry by the Illinois Department of Children and Family Services (DCFS), but no criminal proceedings were instituted at that point. Later, Howard and Tamika’s mother reconciled, at which time Tamika left Fletcher’s care and again spent the bulk of her time with her parents.
According to Fletcher and Tamika, there was substantial drug use and ongoing sexual abuse in Tamika’s home until December 1986, when Tamika’s mother gave birth to Tamika’s little sister. At this point, the details of Tamika’s life become sketchy, but it appears that Fletcher lost track of both Tamika and Howard until February 1987, when she found Tamika living with a friend of her mother. Fletcher took Tamika back to her home in Hammond, Indiana. At this point, Fletcher began seeking legal guardianship of Tamika. After attempting to enlist the aid of the Indiana child protection authorities, Fletcher moved Tamika back to Illinois in June 1987, contacted a child abuse hotline, and took her to Cook County Hospital, where another examining physician, Dr. Shetty, examined her on June 10, 1987, and again found that the hymen was not intact. Five months later, Fletcher once more took Tamika to Cook County Hospital (her efforts to obtain the guardianship having stalled as a result of procedural missteps). This time, the examining physician, Dr. Constance Blade-Schlessinger, found no hymen damage, nor any neovas-cularization (which, if present, would have signaled irritation or inflammation that could have occurred from sexual contact or other disruptive causes such as an infection).
Despite her observation of the intact hymen, Dr. Blade-Schlessinger concluded that Tamika had been sexually abused. (She explained that in incestuous situations, the abuser often refrains from actions that would be forceful or painful, and so the injuries normally associated with forcible rape are often not present.) In 1989, Howard was charged with two counts of aggravated criminal sexual assault in conjunction with his abuse of Tamika. At trial, Tamika and Fletcher testified to the course of events described earlier, while Dr. Blade-Schlessinger offered her expert opinion regarding Howard’s abuse of Tamika. She based her conclusion that Tamika had suffered abuse on the observations of three other doctors — the two who had previously found hymen damage, as well as a psychiatrist who had been working with Tamika during 1987. When asked to reconcile that conclusion with her own examination that indicated no hymen damage, Dr. Blade-Schlessinger said that she believed that Tamika’s hymen may have regenerated itself. She asserted that “when children are removed from situations where they have been sexually abused, they do heal, that there is a constriction process whereby the hymen begins to close down to the normal size associated with that aged child.” When questioned, she said that this phenomenon was “well documented in the literature,” but that studies had not yet adequately shown how long the healing process took. Her own estimate was that two or three months was typical. Finally, she cited to an article that had appeared in
Pediatrics
magazine shortly before the trial (but well after her examination of Tamika). That article detailed a study of sexually abused children whose hymenal injuries had repaired themselves. See M.A. Finkel, “Anogenital Trauma in Sexually Abused Children,” 84
Pediatrics
317 (1989). Howard objected
to the “accuracy and reliability” of this testimony. The Illinois trial judge initially sustained an objection to the foundation for her reference to the article (which had apparently been furnished to Howard’s lawyer three days before Dr. Blade-Schlessinger’s testimony), but after the prosecutor discussed it further with the doctor, Howard’s lawyer never reiterated an objection to either the article or the way that Dr. Blade-Schlessinger referred to it. Both the prosecutor and Howard’s lawyer examined her on the substance of the article.
During the course of proceedings, the prosecutor made a variety of inappropriate remarks. Several times, the prosecutor suggested that Howard had been charged with sexual abuse in 1983 when in fact all that had occurred was the DCFS inquiry. Additionally, the prosecutor made a series of references to the involvement of the DCFS itself in Tamika’s case. Principally, the questions suggested that Howard’s access to Tamika had been restricted; in fact, there was no such limitation. Howard objected to each of these and, in all cases, the trial judge sustained the objection and ordered the testimony disregarded. But most troubling is a question put to Fletcher: “Drawing your attention to 1985, who kidnapped Tamika?” Howard again objected and again the objection was sustained.
Finally, during closing statements, the prosecutor told the jury that “Tamika’s testimony alone would convict this man of this crime.” Her co-counsel later added that “[t]he testimony of one witness who is clear and convincing and credible is enough right there to convict.” Howard objected, claiming that this misstated the burden of proof. The trial judge overruled the objection, ruling that the remark was not designed to state the burden of proof but rather merely that one witness’s testimony (if believed by the jury) can suffice in appropriate circumstances to establish guilt beyond a reasonable doubt. See
People v. D’Angelo,
30 Ill.App.3d 86, 333 N.E.2d 525, 530-31 (1975).
Howard claimed that all of the testimony against him was false and just a result of Fletcher’s jealousy and desire to obtain custody over Tamika, but the jury obviously did not buy that story. After his conviction, Howard appealed in the Illinois courts. While awaiting appeal, he received a letter from his court-appointed appellate counsel, Julie Campbell, indicating that she planned only to challenge the lack of a mental state requirement in the statute under which Howard was convicted and (possibly) a defect in the indictment. Campbell also noted that if Howard so chose, he could file a supplemental brief with the Illinois Appellate Division. He does not appear to have done so. In any event, the appeals court dismissed his appeal and the Illinois Supreme Court declined review.
Howard then sought collateral relief in the Illinois courts. This, too, was unsuccessful. In February 1994, he petitioned the federal district court under 28 U.S.C. § 2254. He also asked for the appointment of counsel, but the district court ruled on the merits of the petition without mentioning this motion. Initially, the district court denied relief because it believed that adequate state grounds existed for the conviction, but we reversed and remanded,
Howard v. Detella,
No. 95-3123, 1996 WL 405212 (7th Cir. July 16, 1996) (unpublished order), for farther consideration because the Illinois Appellate Division had said that his case presented “no issues of arguable merit,” not that Howard had procedurally defaulted. The district court then considered the merits of Howard’s claim, but concluded that even if his appellate counsel’s performance was unreasonably deficient, Howard was not prejudiced and therefore not entitled to habeas relief.
II
A.
As with all claims of ineffective assistance of counsel, we evaluate Howard’s
argument that his appellate lawyer was incompetent according to the two-part test of
Strickland v. Washington,
466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). That means that he .must show both that his lawyer’s performance was unreasonably deficient and that this inadequacy prejudiced him in the sense that there is a reasonable probability that his case would have been remanded for a new trial or that the decision of the state trial court would have been otherwise modified on appeal.
Id.
at 687, 694, 104 S.Ct. 2052;
Blacharski v. United States,
215 F.3d 792, 794 (7th Cir.2000). Also, since Howard filed his habeas petition before the effective date of the Antiterrorism and Effective Death Penalty Act (AEDPA), Pub.L. No. 104-132, 110 Stat. 1214 (1996), we conduct a
de novo
inquiry into these questions.
Jenkins v. Nelson,
157 F.3d 485, 491 (7th Cir.1998);
Lieberman v. Washington,
128 F.3d 1085, 1091 (7th Cir.1997).
Howard’s theory of ineffective assistance is simple — he believes that his appellate lawyer chose the wrong issues for appeal. The district court identified four issues that Howard believes his lawyer should have raised on direct appeal: (1) Dr. Blade-Sehlessinger’s expert testimony (and especially the “hymen regeneration” theory) was not generally accepted in the medical community, (2) the accumulation of prosecutorial misconduct we outlined above led to an unfair trial, (3) the jury instructions failed to remedy the prosecutor’s erroneous statements about the “reasonable doubt” standard, and (4) Fletcher’s report of Tamika’s conversations with her should have been excluded as inadmissible hearsay. Instead, the only issue counsel actually raised was that the indictment was defective for failing to allege a mental state. This court’s opinion in
Mason v. Hanks,
97 F.3d 887 (7th Cir.1996), demonstrates the method by which we apply the
Strickland
test to a claim that counsel failed to raise the correct issues on appeal. We deem performance insufficient when counsel omits a “significant and obvious issue” without a legitimate strategic reason for doing so.
Mason, id.
at 893, quoting
Gray v. Greer,
800 F.2d 644, 646 (7th Cir.1986), and
Hollenback v. United States,
987 F.2d 1272, 1275 (7th Cir.1993). We find prejudice “when that omitted issue ‘may have resulted in a reversal of the conviction, or an order for a new trial.’ ”
Mason,
97 F.3d at 893, quoting
Gray,
800 F.2d at 646.
Since the choice of issues on appeal naturally turns on the relative likelihood of success of each one, there is obviously some overlap between the performance and prejudice inquiries. For example, because reversal is much more likely when the appellate court’s review is
de novo
rather than deferential, the standard of review for a particular issue is a factor in determining the adequacy of appellate counsel’s performance. See,
e.g., Mapes v. Coyle,
171 F.3d 408, 427 (6th Cir.1999),
cert. denied,
— U.S. -, 120 S.Ct. 369, 145 L.Ed.2d 284 (1999);
Bethea v. Artuz,
126 F.3d 124, 127 (2d Cir.1997). The performance half of the
Strickland
test, in this context, requires us to decide whether it was within
the realm of competent appellate representation to decline to raise a particular point on appeal, considering both the arguments that might be made and the governing standard of review. Assuming that a competent lawyer would have taken an appeal at all (a question we need not consider here, since there was no threat of a sentence-enhancing remand that might make the risks of an appeal outweigh its potential benefits), we evaluate counsel’s performance by looking at the issues that the defendant had available and determining whether counsel’s choice of the best of them represented the same kind of strategic choice we permit for trial decisions. See
Mason,
97 F.3d at 893;
Gray,
800 F.2d at 646. The prejudice inquiry is, in a sense, a more absolute one. There we ask only whether there is a reasonable probability that raising the issue would have made a difference in the outcome of the appeal. In other words, performance is about picking the battles; prejudice looks at whether winning the battle would have made a difference in-the outcome of the war.
In Howard’s case, the district court found that his appellate counsel performed inadequately because it believed that there was “no reason why Howard’s attorney could not simply have added the issues Howard wanted raised to the single issue she already planned to present.” The court indicated that this was a function not only of the mechanical ease of adding on more issues, but also, more importantly, “the dictates of legal strategy” and the fact that “at least some of Howard’s issues had a better factual and legal basis than the issue that his counsel presented.” Appellate lawyers are clearly not incompetent when they refuse to follow a “kitchen sink” approach to the issues on appeals. To the contrary, one of the most important parts of appellate advocacy is the selection of the proper claims to urge on appeal.
Schaff v. Snyder,
190 F.3d 513, 526-27 (7th Cir.1999). Throwing in every conceivable point is distracting to appellate judges, consumes space that should be devoted to developing the arguments with some promise, inevitably clutters the brief with issues that have no chance because of doctrines like harmless error or the standard of review of jury verdicts, and is overall bad appellate advocacy. On the other hand, counsel can’t throw the baby out with the bath water. That is what concerned the district court, and it concerns us as well. Howard’s attorney chose to appeal only one issue, in spite of the fact that Howard himself had called counsel’s attention to the other points, the record supported an appeal on those points, and Howard was not asking for an inordinate number of issues. We therefore have no cause to disagree with the district court’s conclusion that Howard has shown defective performance on the part of appellate counsel, for
Strickland
purposes, and we turn to the prejudice question.
B.
Turning to the specifics of Howard’s ease, his first and most substantial argument is that his appellate lawyer was constitutionally ineffective for failing to contest on appeal the district court’s decision to allow Dr. Blade-Schlessinger to testify that she believed that Tamika’s hymen had repaired itself. His basic argument is that the “hymen regeneration” testimony lacked foundation, was not believable, and was therefore highly prejudicial because it explained away an otherwise highly exculpatory fact — Dr. BladeSchlessinger’s observation that Tamika’s hymen was normal.
The principal reason the district court gave for rejecting this point concerned the performance of Howard’s
trial
counsel. Because trial counsel did not object to the admission of this particular expert opinion (and the trial court had qualified Dr. Blade-Schlessinger as an expert both in child abuse and as a doctor), Illinois courts would find the issue waived on appeal. See
People v. Mahaffey,
166
Ill.2d 1, 209 IIl.Dec. 607, 651 N.E.2d 1055, 1067 (1995). Nor did defense counsel ever perfect an objection to Dr. Blade-Schles-singer’s reference to the then-recent August 1989 article in
Pediatrics
— an article that buttressed her 1987 conclusions, but which obviously did not play a role in her contemporaneous diagnosis. To the contrary, counsel cross-examined her about the substance of the article in some detail, bringing out for the jury both the paucity of evidence it revealed for regeneration of damaged hymens (only one case involved this situation) and the fact that scar tissue was evident after healing (at least in early examinations).
Under the circumstances, appellate counsel would have had a serious uphill battle to convince an Illinois appellate court to overturn the admission of this testimony. As is generally true, evidentiary decisions are left to the discretion of the trial courts in Illinois, so appellate review is deferential.
People v. Jones,
306 Ill.App.3d 793, 239 IIl.Dec. 811, 715 N.E.2d 256, 261 (1999). Second, using the distinction between waiver and forfeiture articulated by the United States Supreme Court in
United States v. Olano,
507 U.S. 725, 733, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993), an Illinois court may well have found that Howard waived his right to challenge Dr. Blade-Schlessinger’s theory and that it was not simply forfeited and fair ground for something like our plain error review. Even if it could have reached the merits of the point — which is at best unclear — the court would not have found reversible error unless it considered the admission of Dr. Blade-Schlessinger’s opinion and her references to the
Pediatrics
article so beyond the pale of acceptable conduct as to constitute an abuse of discretion. In determining the admissibility of expert testimony, Illinois continues to apply the “general acceptance” test, meaning that the article on which Dr. BladeSchlessinger relied could serve as a proper foundation for Dr. Blade-Schlessinger’s testimony if it is “generally accepted by the relevant scientific community.”
People v. Oliver,
306 Ill.App.3d 59, 239 Ill.Dec. 196, 713 N.E.2d 727, 734 (1999), citing
Frye v. United States,
293 F. 1013 (D.C.Cir.1923). Howard does not present any reason to believe that an article in
Pediatrics,
which is the official journal of the American Academy of Pediatrics, would not meet Illinois’s evidentiary standards for foundation for expert testimony.
The bulk of Howard’s attack on Dr. Blade-Schlessinger’s testimony questions the sensibleness of her regeneration theory. He makes a variety of arguments that all revolve around the same theme — it is simply not believable that a girl as young as Tamika could be subjected to intercourse with an adult male and show no signs of hymenal damage or scarring. He further maintains that the Finkel article and even Dr. Blade-Schlessinger’s testimony itself actually support his position. But that does not reflect the entirety of Dr. Blade-Schlessinger’s testimony. She also told the jury, as we have already noted, that incestuous situations are different from ordinary sexual violence and tend to show less in the way of physical injury. Taken as a whole, the jury could have inferred that Howard’s physical contact with Tamika was enough to make her hymen “not intact” — that is, lacerated or torn somehow — and more than enough to irritate the genital area, but not enough to destroy it beyond repair. That would have explained both the earlier medical findings of injury, and Dr. Blade-Schlessinger’s later finding of a normal body. Howard offers a strong critique of Dr. BladeSchlessinger’s testimony, but all of his arguments were for the jury, not the appellate judges, to consider. Howard’s trial counsel vigorously contested the logic of Dr. Blade-Schlessinger’s conclusions, but the jury put- the totality of the evidence together in a different way. Since Illinois’s higher courts do not second-guess jury conclusions regarding questionable testimony,
People v. Kirwan,
96 Ill.App.3d 121, 51 IIl.Dec. 809, 421 N.E.2d 317, 320
(1981);
People v. Dunn,
49 Ill.App.3d 1002, 7 Ill.Dec. 879, 365 N.E.2d 164, 170 (1977), there is no reason to think that presenting these arguments to the Illinois appellate courts would have increased the likelihood of reversal. Consequently, Howard cannot show prejudice from his appellate counsel’s failure to contest Dr. Blade-Schlessinger’s testimony on appeal.
C.
The statements from the prosecutor set forth above were indeed troublesome, and once again, we do not disagree either with the district court’s assessment that they were “clearly deliberate and improper prosecutorial comments,” nor its implicit conclusion that appellate counsel should have presented this ground to the Illinois appellate courts. Once again, the critical part of the
Strickland
inquiry for Howard is the prejudice question. The district court concluded that these remarks, inappropriate though they were, did not deprive Howard of his right to a fair trial.
The leading Supreme Court decision on the question whether prosecutorial misconduct is so egregious that a new trial is required, as a matter of constitutional law, is
Darden v. Wainwright, 477
U.S. 168, 106 S.Ct. 2464, 91 L.Ed.2d 144 (1986). In
Darden,
the Court set forth six factors that should be considered in deciding this question: (1) whether the prosecutor misstated the evidence, (2) whether the remarks implicate specific rights of the accused, (3) whether the defense invited the response, (4) the trial court’s instructions, (5) the weight of the evidence against the defendant, and (6) the defendant’s opportunity to rebut. 477 U.S. at 181-82, 106 S.Ct. 2464; see also
United States v. Pirovolos,
844 F.2d 415, 426 (7th Cir.1988) (reciting the factors outlined in
Darden).
Here, we must filter that inquiry through the question whether appellate counsel’s failure to raise this argument prejudiced Howard. Once again, we conclude that the answer is no.
First, Illinois courts do not reverse a jury’s verdict because of this kind of prosecutorial misconduct “absent a showing that the prosecution’s improper remarks resulted in substantial prejudice to the defendant and that, without those remarks, the verdict would have been different.”
People v. Modrowski,
296 Ill.App.3d 735, 231 Ill.Dec. 299, 696 N.E.2d 28, 39 (1998) (citation omitted). Second, as the district court here observed, the most important of the
Darden
factors is the weight of the evidence against the defendant.
United States ex rel. Howard v. DeTella,
959 F.Supp. 859, 867 (N.D.Ill.1997). That evidence included Tamika’s own testimony, which the district court found to be a “clear and straightforward description of repeated sexual abuse.” Added to that was the testimony of Linda Fletcher, who reported both her own observations of Tamika’s body and Tamika’s reports of abuse. Dr. Blade-Schlessinger testified that she concluded Tamika had been abused, both in reliance on the earlier medical reports and the report of the treating psychiatrist Tamika was seeing, and on her own examination.
Howard is certainly right that the jury was not compelled to believe this evidence, and that some of the
Darden
factors cut in his favor. The prosecutor misstated the evidence and nothing Howard did invited the comments. On the other hand, Howard’s attorney objected to four of the remarks in question, his objections were sustained, and the jury was instructed to disregard the remarks. Howard’s attorney also had the opportunity to rebut, even though for understandable strategic reasons (we assume) he chose not to do so. In the end, we think that the Illinois appellate court, following
Darden,
would have concluded that the remarks were improper, but that they did not go to the heart of the prosecution’s case — whether Howard abused Tamika. We do not believe that this point would have struck it as reversible error, and thus we cannot
say appellate counsel’s decision not to give it a try was prejudicial.
D.
Last, we come to Howard’s point concerning the district court’s failure to rule on his motion for appointment of counsel, and the related point that he should have received counsel for the presentation of his habeas corpus petition. We are concerned about the “fall-between-the-cracks” sense we get about the way this motion was handled. Ordinarily, we would review a decision not to appoint counsel for abuse of discretion,
Zarnes v. Rhodes,
64 F.3d 285, 288 (7th Cir.1995), but in this instance we think it preferable to give the question
de novo
review, since it is quite possible that there is no underlying decision to review.
In general, a refusal to appoint counsel calls for reversal “only ‘if, given the difficulty of the case and the litigant’s ability, [he] could not obtain justice without an attorney, [he] could not obtain a lawyer on his own, and [he] would have had a reasonable chance of winning with a lawyer at [his] side.’ ”
Winsett v. Washington,
130 F.3d 269, 281 (7th Cir.1997), quoting
Forbes v. Edgar,
112 F.3d 262, 264 (7th Cir.1997) (alterations in
Winsett).
We apply liberal standards to this inquiry, because there is a certain circularity to the argument. An unskilled lay defendant may have trouble showing the court which of his arguments has serious legal merit, whereas a lawyer may be able to see right away which parts of the case have possibilities.
In this case, Howard was able to present his
Strickland
arguments to the district court, and from that point, the district court’s task was to review the state proceedings in light of both Illinois law and the Sixth Amendment standard for effective counsel. Counsel could not have changed the strong evidence against Howard; he made no claim that other newly discovered evidence would have exonerated him; and we are not convinced that further exploration into the medical testimony would have made a difference given the applicable standards of review. Howard’s able counsel on appeal, to whom we give thanks, has done the best she could to show prejudice from the performance of his state appellate counsel and to indicate how the federal proceedings might have gone better for Howard if he had received counsel right away in the district court, but we do not find that Howard’s rights were substantially affected by his lack of counsel at the district court, and thus we decline the request to send this case back for further proceedings.
Ill
We have considered the other arguments Howard has raised and find no ground for reversal in them. We therefore Affirm the judgment of the district court.