JOHN M. WALKER, JR., Circuit Judge:

John M. Purdy, Jr., has moved for a certificate of appealability in order to appeal from a July 27, 1999 judgment of the United States District Court for the District of Connecticut (Alfred V. Covello, Chief Judge) denying his petition for a writ of habeas corpus. In 1996, Purdy was convicted of conspiring to violate the Anti-Kickback Act, 41 U.S.C. § 53, 54, in violation of 18 U.S.C. § 371, and we affirmed his conviction, see United States v. Purdy, 144 F.3d 241 (2d Cir. 1998). In his habeas corpus petition, Purdy claims that he was denied effective assistance of counsel when his attorney at trial failed to communicate to him certain representations of the federal prosecutor and failed to specifically and explicitly advise him to accept a guilty plea. After denying his petition, the district court refused to grant Purdy's motion for a certificate of appealability under 28 U.S.C. § 2253(c), prompting Purdy to move for one in this court. Because Purdy has failed to make a substantial showing of the denial of a constitutional right, see 28 U.S.C. § 2253(c)(2), we deny the motion.

BACKGROUND

Unless otherwise indicated, the facts pertinent to the petition are substantially undisputed. In 1995, the government commenced an investigation into whether Purdy, along with other military parts suppliers, had paid kickbacks to federal contractors to obtain supply contracts. Purdy retained attorney Jacob Zeldes to represent him. The Assistant U.S. Attorney in charge of the investigation, Mark Califano, met with Zeldes on several occasions to discuss the possibility of a plea agreement and Purdy's cooperation. It was Zeldes's practice to record what transpired at these meetings and then to communicate to Purdy, verbally and often in writing, the substance of what was said. Purdy's principal objective was to avoid imprisonment, but the prosecutor made it clear to Zeldes that even if Purdy were to plead guilty and cooperate, the government would still seek a term of imprisonment in accordance with the sentencing guidelines. Throughout this period and continuing into the subsequent trial, Purdy proclaimed to Zeldes his innocence and his ignorance of matters as to which the government was seeking his cooperation. With this background, we turn to the alleged deficiencies in Zeldes's representation about which Purdy complains.

On April 26, 1995, Zeldes and Califano held one of their meetings to review the situation. Califano said that the case would probably be assigned to District Judge Burns and that he "didn't know of any case where a white collar criminal with those guidelines had gone to jail with Judge Burns or Judge Dorsey." Zeldes admits that he did not convey this particular remark to Purdy. Zeldes testified that he did not consider Califano's observation to be worth mentioning to Purdy because there was no certainty that the case would be assigned to Judge Burns, and because in a recent white collar case before Judge Dorsey, Zeldes's client had received a prison sentence of four and one-half years. Zeldes also testified that he was concerned about improperly coercing Purdy, a client who had consistently maintained his innocence, into pleading guilty.

Purdy also complains that Zeldes withheld from him a discussion with Califano concerning a possible sentence range of either 12 to 18 or 15 to 21 months, depending on Purdy's precise role in the kickback scheme. Zeldes disputes that the 12 to 18 month range was ever discussed, but admits that Califano discussed a range of 15 to 21 months, a proposal that Zeldes considered to be "in the nature of an offer," and maintains that he told Purdy of the government's position. Purdy denies that Zeldes communicated these ranges. Sometime thereafter, Zeldes informed Purdy, this time in writing, that his exposure to imprisonment after a trial would be 27 to 33 months, but upon a plea of guilty the government had offered between 18 and 24 months, "with the possibility, but not a certainty, for a lesser sentence."

On December 8, 1995, Purdy met with Zeldes and his associates, who grilled their client in a mock cross examination that was designed to highlight the strengths of the government's case. Afterwards, Zeldes again stressed the difficulty of the case from Purdy's perspective and Purdy's option of pleading guilty on the terms Zeldes had previously outlined. Purdy did not waver in his insistence on his innocence and his desire to go to trial.

On January 16, 1996, Zeldes again outlined to Purdy in writing an exposure of 27 to 33 months following a trial or one of 18 to 24 months upon a guilty plea. Zeldes also impressed upon Purdy the "troublesome" aspects of the case, taking care to advise Purdy and not pressure him to admit guilt, so long as he continued to maintain his innocence. On February 19, 1996, Zeldes wrote to his client again, urging him to "give very careful consideration" to the case and pointing out once more the strength of the case against him. He advised Purdy that "it will be difficult to convince the jury that [certain payments] were not made for purposes of kickbacks." Undeterred by these communications, Purdy maintained his innocence.

At trial, the jury found Purdy guilty and Judge Covello sentenced him to 37 months' imprisonment. After Purdy agreed to cooperate, his sentence was reduced to 18 months.

In denying Purdy's habeas petition after an evidentiary hearing, the district court concluded that it need not decide whether Zeldes's performance had fallen below an objective standard of professional reasonableness because nothing in Zeldes's representation had prejudiced Purdy in his defense, and that therefore the second part of the controlling test for ineffective assistance of counsel had not been met. See Strickland v. Washington, 466 U.S. 668, 687 (1984).

The district court found that Zeldes had substantially communicated the government's plea offers to Purdy; had not prejudiced Purdy by keeping Califano's remark about Judge Burns and Judge Dorsey to himself, since his own experience cast those remarks into doubt; had adequately informed Purdy about the consequences of pleading guilty and cooperating with the government; had stressed the strength of the government's case; and had urged Purdy to carefully weigh his options of pleading guilty and going to trial. Critically, the district court refused to credit Purdy's post-conviction testimony that if only he had been aware of Califano's comments, he would have foregone the trial and pled guilty at that time. Judge Covello found that "[t]o the contrary, the record reflects that Purdy stoutly maintained his innocence of the alleged charges and his total lack of knowledge" of occurrences that the government sought to uncover by his cooperation. The district court therefore found that Purdy had failed to establish that Zeldes's alleged deficiencies had prejudiced him and thus that his ineffective assistance claim could not be maintained under Strickland.

The district court denied a certificate of appealability. Purdy now moves for one in this court.

DISCUSSION

Following the enactment of the Anti-Terrorism and Effective Death Penalty Act of 1996, no appeal may be taken from a denial of a habeas corpus petition in the district court unless either the district court or the court of appeals grants a certificate of appealability. The certificate will be granted only if defendant has made a substantial showing of the denial of a constitutional right. See 28 U.S.C. § 2253(c)(2). To demonstrate that he was denied his Sixth Amendment right to effective assistance of counsel, Purdy must show both that his counsel's performance was deficient as measured by objective professional standards, and that this deficiency prejudiced his defense. See Strickland, 466 U.S. at 687; McKee v. United States, 167 F.3d 103, 106 (2d Cir. 1999).

The performance inquiry is contextual; it asks whether defense counsel's actions were objectively reasonable considering all the circumstances. See Strickland, 466 U.S. at 688. In determining what constitutes objective reasonableness, courts look for guidance to "[p]revailing norms of practice as reflected in American Bar Association standards." Id. These standards provide at least two benchmarks for the representation of a client who is deciding whether to accept a plea offer.

On the one hand, defense counsel "`must give the client the benefit of counsel's professional advice on this crucial decision'" of whether to plead guilty. Boria v. Keane, 99 F.3d 492, 497 (2d Cir. 1996) (quoting Anthony G. Amsterdam, Trial Manual 5 for the Defense of Criminal Cases (1988)) (emphasis omitted); see also Cullen v. United States, 194 F.3d 401, 404 (2d Cir. 1999) (" Boria recognizes a lawyer's general duty to advise a defendant concerning acceptance of a plea bargain."); Model Rules of Professional Conduct Rule 1.4(b) (1995) ("A lawyer shall explain a matter to the extent reasonably necessary to permit the client to make informed decisions regarding the representation."). As part of this advice, counsel must communicate to the defendant the terms of the plea offer, see Cullen, 194 F.3d at 404, and should usually inform the defendant of the strengths and weaknesses of the case against him, as well as the alternative sentences to which he will most likely be exposed, see United States v. Gordon, 156 F.3d 376, 380 (2d Cir. 1998) ("[K]nowledge of the comparative sentence exposure between standing trial and accepting a plea offer will often be crucial to the decision whether to plead guilty.") (internal quotation marks omitted); Jones v. Murray, 947 F.2d 1106, 1110-11 (4th Cir. 1991) (interpreting Standards for Criminal Justice 14-3.2(b) (2d ed. 1986 Supp.)).

On the other hand, the ultimate decision whether to plead guilty must be made by the defendant. See Model Rules of Professional Conduct Rule 1.2(a) (1995) ("In a criminal case, the lawyer shall abide by the client's decision, after consultation with the lawyer, as to a plea to be entered. . . ."). And a lawyer must take care not to coerce a client into either accepting or rejecting a plea offer. See Jones, 947 F.3d at 1111 ("[V]arious [ABA] Standards place upon counsel an affirmative duty to avoid exerting `undue influence on the accused's decision' and to `ensure that the decision . . . is ultimately made by the defendant.'" (quoting Standards for Criminal Justice 4-5.1(b) 14-3.2(b))).

Counsel's conclusion as to how best to advise a client in order to avoid, on the one hand, failing to give advice and, on the other, coercing a plea enjoys a wide range of reasonableness because "[r]epresentation is an art," Strickland, 466 U.S. at 693, and "[t]here are countless ways to provide effective assistance in any given case," id. at 689. Counsel rendering advice in this critical area may take into account, among other factors, the defendant's chances of prevailing at trial, the likely disparity in sentencing after a full trial as compared to a guilty plea (whether or not accompanied by an agreement with the government), whether defendant has maintained his innocence, and the defendant's comprehension of the various factors that will inform his plea decision.

We think Zeldes represented Purdy in a manner that successfully steered a course between the Scylla of inadequate advice and the Charybdis of coercing a plea. Nothing to which Purdy draws our attention persuades us otherwise. First, Purdy complains that Zeldes misrepresented his sentence exposure upon a plea of guilty. Califano says he proposed a sentence of 12 to 18 or 15 to 21 months in prison, depending on Purdy's role in the offense, and with a possibility of a reduction if Purdy cooperated with the investigation. Zeldes denies that he ever discussed a range of 12 to 18 months with Califano, but he testified that he did communicate Califano's offer of 15 to 21 months to Purdy. Zeldes also informed Purdy in writing that upon a plea of guilty he could expect a sentence of "between 18 to 24 months, with the possibility, but not the certainty, of a lesser sentence." Because Zeldes did not place an absolute floor under the sentence of 18 months but suggested the "possibility, but not the certainty" of a lower term, the district court concluded, we think correctly, that Zeldes's communications adequately conveyed the government's position.

Second, we think it was reasonable for Zeldes not to have told Purdy of Califano's comment that he "didn't know of any case where a white collar criminal with those guidelines had gone to jail with Judge Burns or Judge Dorsey." Zeldes did not think that this information was accurate, since he had recently represented a client in a similar situation before Judge Dorsey, who had sentenced the client to a prison term. Under the circumstances, including Purdy's persistent insistence on his innocence despite his understanding of his case and its shortcomings, and considering the proscription against an attorney's coercion of his client's guilty plea, we think that Zeldes's judgment not to convey Califano's comment to Purdy conformed to professional standards.

Zeldes's decision to forgo specifically and explicitly telling Purdy whether he should accept the government's plea offer was also within the range of professional reasonableness. Zeldes repeatedly advised Purdy of the strengths of the government's case against him and went so far as to demonstrate these strengths in a mock cross examination. This strategy evidently worked, because Purdy himself testified that he understood that his task at trial would be "very difficult." Furthermore, as previously noted, Zeldes adequately advised Purdy of the likely sentence he would face after trial. Again in light of Purdy's steadfast protestations of innocence, we think these actions reasonably fulfilled Zeldes's obligation to advise Purdy while avoiding the danger of improper coercion.

Purdy argues that our decision in Boria establishes a per se rule that counsel not only must inform each client of the probable costs and benefits of accepting a plea offer, but also, in so many words and regardless of the particular circumstances, must advise the defendant to either plead guilty or not. We disagree. Boria did not fashion such a blunt legal instrument.

In Boria the government offered the defendant a plea bargain whereby he would receive a sentence of one to three years in prison. Boria rejected the plea offer and, after a trial on charges arising from a "buy and bust" drug operation, received a prison term of 20 years to life. Defense counsel did not "in any way or at any time discuss . . . the advisability of accepting or rejecting the offered plea." Boria, 99 F.3d at 495. He confined his discussions with Boria to trial strategy, even though he personally believed that rejection of the plea offer would be "suicidal." Id. To justify his failure to either inform or advise his client, Boria's attorney relied solely on the fact that Boria had repeatedly told him that he did not want to plead guilty because it would embarrass him in front of his children. We held that in such circumstances the actions, or more precisely the inactions, of Boria's defense attorney departed from objective standards of professional conduct and thus were constitutionally deficient. See id. at 497.

Because Purdy's situation differs from that of Boria in several crucial respects, we think the holding of that case does not control here. The Boria court's task was to determine the parameters of "a criminal defense lawyer's duty when a defendant's best interests clearly require that a proffered plea bargain be accepted, but the defendant, professing innocence, refuses to consider the matter." Id. at 496. Although we quoted with approval a professional responsibility text that instructed lawyers to "`advise [their] client[s] fully on whether a particular plea to a charge appears to be desirable,'" Boria, 99 F.3d at 496 (quoting Model Code of Professional Responsibility EC 7-7 (1992)), we ultimately held only that under the circumstances of that case, Boria's lawyer had failed to adequately advise his client, see id. at 497. We held that a defense lawyer must do more than Boria's counsel did; we did not hold that a lawyer in that position must do everything Purdy now proposes. To the extent that there is language in Boria that seems to go further than its holding, it is non-binding dicta. Because in Boria defense counsel's representation clearly departed from objective norms of professional conduct, in contrast to Zeldes's conduct here, we think that case is easily distinguished from this one.

First, Boria's defense counsel not only failed to give an ultimate opinion as to the wisdom of a plea, but also "never gave his client any advice or suggestion as to how to deal with the People's offered plea bargain." Id. at 498. In contrast, Zeldes informed Purdy at length and on several occasions of the strength of the government's case, going so far as to stage a mock cross examination, and of the benefits of pleading guilty and cooperating. Unlike Boria, Purdy, a sophisticated businessman, understood that the case against him was formidable. The fact that Zeldes never advised Purdy to plead guilty in so many words does not equate to a professional deficiency in such circumstances.

Second, unlike Boria, Purdy's best interests did not necessarily demand a guilty plea. The sentence disparity facing Purdy (27 to 33 months after trial and 18 to 24 months after a plea, or even 12 to 18 months after a plea as Purdy claims) was not nearly as vast as the one facing Boria (20 years to life after trial and one to three years after a plea). The evidence against Purdy, although strong, was considerably more complex than that against Boria, for whom there was no "reasonable chance of acquittal." Id. at 498; see also id. at 495 n. 4 (noting Boria's lawyer's testimony that "getting an acquittal in a drug case in upstate New York was almost impossible"). If Purdy's decision to go to trial turned out to be unwise, it was at least reasonable; Boria's was not.

Third, it is far from clear to us that Purdy refused to consider the possibility of a plea. Rather, it seems quite plain from Purdy's own testimony that his decision not to plead guilty was based on a rational calculation: since the government's plea offer was not significantly less than his exposure would be at trial, Purdy thought it preferable to risk a somewhat greater sentence at trial in the hope of getting an acquittal and thus achieving his primary goal of avoiding incarceration altogether. He made this decision fully cognizant that his case was "very difficult." For these reasons, we think Boria factually distinguishable.

The three sources from which the Boria court drew its standards of professional responsibility do not compel a different result. First, the American Bar Association's Model Code of Professional Responsibility counsels that "[a] defense lawyer in a criminal case has the duty to advise his client fully on whether a particular plea to a charge appears to be desirable." Boria, 99 F.3d at 496 (quoting Model Code of Professional Responsibility EC 7-7 (1992)). Although we have acknowledged above that a defense attorney must advise a client who is facing a plea decision, we also noted that counsel's choice of how to do so will be guided by many factors, including the duty to avoid coercing a plea from an unwilling client. Moreover, the Model Code has been superceded by the Model Rules of Professional Conduct. See Geoffrey C. Hazard, Jr., The Future of Legal Ethics, 100 Yale L.J. 1239, 1251 (1991). The Model Rules delete the language cited in Boria. Cf. Model Rules of Professional Conduct Rule 1.2 (1995) (replacing EC 7-7).

Second, the Amsterdam trial manual instructs that, although the important decision whether to plead guilty must be left ultimately to the client, the defense attorney " must give the client the benefit of counsel's professional advice." Boria, 99 F.3d at 497 (quoting Anthony G. Amsterdam, Trial Manual 5 for the Defense of Criminal Cases (1988)) (emphasis from Boria omitted but from Amsterdam retained). But as we have already noted, Zeldes did give Purdy "professional advice" about the quality of the government's case against him.

Finally, the Boria court cited an admonition from the Supreme Court that counsel should "make an independent examination of the facts, circumstances, pleadings and laws involved and then . . . offer his informed opinion as to what plea should be entered." Boria, 99 F.3d at 497 (quoting Von Moltke v. Gillies, 332 U.S. 708, 721 (1948)). But in Von Moltke, the defense lawyer not only failed to give an opinion with the specificity and explicitness that Purdy urges, but also failed to educate the defendant as to her legal rights. See Von Moltke, 332 U.S. at 721-22. Here, by contrast, Zeldes took reasonable steps to be sure Purdy was cognizant of his rights.

We note these disparities not to cast doubt on the outcome of Boria, but only to recognize that reasonable professional conduct does not under all circumstances require a lawyer to give an explicit opinion as to whether a client should take a plea offer. See Jones, 947 F.2d at 1110-11 (citing ABA standards in support of the conclusion that defense counsel who reviewed the strengths and weaknesses of the prosecution's case but did not ultimately recommend that defendant accept the plea did not render ineffective assistance).

Nor is this case controlled by our recent decision in Cullen. In that case, the defense attorney failed to communicate to defendant even the terms of the government's offer, let alone an accurate comparison of the sentencing ranges Cullen would have faced upon a plea of guilty and after a trial. See Cullen, 194 F.3d at 404. In fact, we explicitly left open the question of whether "there might be circumstances where defense counsel need not render advice as to acceptance of a plea bargain." Id. Because Cullen's attorney failed even to render the most basic advice, our decision in his case does not require us to conclude that in this case Zeldes's significantly more competent representation was ineffective under the Sixth Amendment.

Per se rules like the one Purdy advocates — and mistakenly reads into Boria — are not well calibrated to gauge the ineffective assistance of counsel. The purpose of Sixth Amendment protection, after all, is to guarantee the right to a fair trial, see Strickland, 466 U.S. at 684, or, stated a bit more broadly, to ensure a fairly arrived at outcome of the criminal proceeding. With this constitutional objective properly understood, it is "[m]ost important," the Strickland Court stressed, that "in adjudicating a claim of actual ineffectiveness of counsel, a court should keep in mind that the principles we have stated do not establish mechanical rules. . . . [T]he ultimate focus of inquiry must be on the fundamental fairness of the proceeding." Id. at 696; see also Jones, 947 F.2d at 1111 n. 3 ("[I]f we accepted petitioner's argument [that we should adopt] a binding rule of conduct [for defense counsel] in all cases, we would run afoul of the Supreme Court's explicit warning that `[n]o particular set of detailed rules for counsel's conduct can satisfactorily take account of the variety of circumstances faced by defense counsel. . . .'" (quoting Strickland, 466 U.S. at 688-89)). The Strickland Court thus repeatedly emphasized that the "performance inquiry must be whether counsel's assistance was reasonable considering all the circumstances," Strickland, 466 U.S. at 688, that a court "must judge the reasonableness of counsel's challenged conduct on the facts of the particular case," id. at 690, and that "[t]here are countless ways to provide effective assistance in any given case," id. at 689.

Given these cautionary statements by the Supreme Court, we think it unwise to read Boria to have established a per se rule that defense counsel must always expressly advise the defendant whether to take a plea offer. Under the circumstances of this case, Zeldes acted reasonably when he informed Purdy fully of the strength of the government's case against him, together with the nature of the government's plea offer, without specifically advising Purdy to take the plea.

Our inquiry could end here, since Purdy has failed to demonstrate that his counsel's performance was deficient, the first independent showing required by Strickland. See id. at 697. Nevertheless, we also agree with the district court that even if Zeldes's failure to give a specific recommendation had constituted unreasonable lawyering, the deficiency would not have prejudiced Purdy. Under this second part of Strickland's test, Purdy must demonstrate a reasonable probability that but for Zeldes's deficiencies, he would have pled guilty. See Cullen, 194 F.3d at 404. The district court did not find credible Purdy's testimony either that he would have pled guilty had he known of Califano's comments regarding Judge Burns and Judge Dorsey, or that he would have cooperated had he known of the government's interest. There is no reasonable basis for us to upset the district court's credibility determination of Purdy's post-conviction testimony, or its holding that Purdy has failed to demonstrate a reasonable probability that the outcome would have been different had Zeldes performed in the manner Purdy now says he should have.

CONCLUSION

Because Purdy has failed to make a substantial showing that his constitutional right to effective assistance of counsel was violated, his motion for a certificate of appealability is denied.