Strickland v. Washington: The Two-Prong Test for Ineffective Counsel

By Elizabeth Franklin-Best, Esq., and Christopher Zoukis, JD, MBA

What does Strickland v. Washington require?

Strickland v. Washington, 466 U.S. 668 (1984), requires a defendant claiming ineffective assistance of counsel to prove two things: that counsel’s representation “fell below an objective standard of reasonableness,” and that there is “a reasonable probability” the result would have been different. A reasonable probability is one “sufficient to undermine confidence in the outcome.” Both prongs are required.

Key takeaways on the Strickland test

• The test has two prongs and you need both. Strickland holds that a claim “has two components. First, the defendant must show that counsel’s performance was deficient. … Second, the defendant must show that the deficient performance prejudiced the defense.” Failing either one ends the claim.
• The performance prong starts with the presumption against you. “A court must indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.” The movant’s job is to overcome it, not to argue in a vacuum.
• “Reasonable probability” is lower than “more likely than not.” It is “a probability sufficient to undermine confidence in the outcome” — but it is higher than “some conceivable effect,” which Strickland expressly rejects as insufficient.
• A court can skip the first prong entirely. Strickland directs that “if it is easier to dispose of an ineffectiveness claim on the ground of lack of sufficient prejudice, which we expect will often be so, that course should be followed.” Most denials you will read are prejudice denials.
• Presumed prejudice exists but is very narrow. United States v. Cronic, 466 U.S. 648 (1984), presumes prejudice where counsel is absent at a critical stage or “entirely fails to subject the prosecution’s case to meaningful adversarial testing” — and Bell v. Cone, 535 U.S. 685 (2002), holds that failure must be complete, not a failure at particular points.
• The “doubly deferential” standard is a state-prisoner rule. Harrington v. Richter, 562 U.S. 86 (2011), layers 28 U.S.C. §2254(d) deference on top of Strickland for state prisoners. A federal prisoner filing under 28 U.S.C. §2255 gets Strickland once, not twice.

What is the Strickland test?

The Strickland test is the constitutional standard for deciding whether a criminal defense lawyer’s performance was so poor that it violated the Sixth Amendment right to counsel. It comes from Strickland v. Washington, 466 U.S. 668 (1984), and it is the framework a federal court applies to nearly every claim that a lawyer failed a client — at trial, at a plea, at sentencing, and on appeal.

The Court stated it in two sentences that every later case quotes:

“A convicted defendant’s claim that counsel’s assistance was so defective as to require reversal of a conviction or death sentence has two components. First, the defendant must show that counsel’s performance was deficient. … Second, the defendant must show that the deficient performance prejudiced the defense.”

Both showings are mandatory. Strickland says that “[u]nless a defendant makes both showings, it cannot be said that the conviction or death sentence resulted from a breakdown in the adversary process that renders the result unreliable.” A movant who proves a genuinely terrible piece of lawyering, but cannot connect it to the outcome, loses. A movant who identifies a devastating outcome, but cannot show the lawyer’s choice was objectively unreasonable, also loses.

The Sixth Amendment text itself says only that “[i]n all criminal prosecutions, the accused shall enjoy the right … to have the Assistance of Counsel for his defence.” U.S. Const. amend. VI. Strickland is the Court’s answer to the question the text leaves open: what does “Assistance of Counsel” require in substance, and how does a court measure a shortfall after the fact? What sorts of attorney failures courts have actually accepted as deficient — missed suppression motions, uninvestigated witnesses, unadvised immigration consequences — belongs to ineffective assistance of counsel, which catalogs them.

One procedural point matters before the standard does. In federal court, an ineffective-assistance claim is a §2255 claim, not a direct-appeal claim. Massaro v. United States, 538 U.S. 500 (2003), holds: “We hold that an ineffective-assistance-of-counsel claim may be brought in a collateral proceeding under § 2255, whether or not the petitioner could have raised the claim on direct appeal.” That is why Strickland is the single most-litigated doctrine in federal post-conviction relief, and why the ordinary rules about procedural default do not bar an ineffectiveness claim merely because appellate counsel did not raise it. Massaro is also the reason ineffectiveness is one of the few claims that belongs in a §2255 motion rather than in a direct appeal, where the trial record almost never contains the evidence the claim needs.

What was Strickland v. Washington about?

Strickland arose from a Florida capital case. David Leroy Washington pleaded guilty, against his lawyer’s advice, to three capital murders and a series of related crimes committed during a ten-day crime spree, and waived his right to an advisory jury at sentencing. His experienced criminal-defense lawyer made a strategic judgment about the sentencing hearing: he did not request a psychiatric examination, did not seek out character witnesses, and did not request a presentence report. His reasoning was that Washington’s own statements of acceptance of responsibility and remorse, made directly to the sentencing judge, were the best available mitigation, and that opening the door to a broader inquiry would let the State cross-examine and introduce Washington’s criminal history in rebuttal. The trial judge imposed death on each count.

Washington then sought collateral relief in the Florida courts and, when that failed, federal habeas relief, arguing that counsel’s decisions at sentencing amounted to ineffective assistance. The federal district court denied relief. A divided en banc Fifth Circuit set out its own standard and remanded. The Supreme Court granted certiorari, announced the two-component test, applied it to the record, and reversed — holding that Washington had shown neither deficient performance nor prejudice, because counsel’s strategic judgment was reasonable on the facts he knew and the aggravating evidence was overwhelming.

That disposition is worth sitting with, because it explains the shape of the opinion. Strickland is not a case where the movant won. It is a case where the Court built a framework and then used it to deny relief, which is why the opinion is full of language protecting counsel’s judgment. Reading Strickland as a neutral test misreads the document. It was written by a Court that was worried about “the temptation for a defendant to second-guess counsel’s assistance after conviction or adverse sentence” and about “intrusive post-trial inquiry into attorney performance.” Those concerns are built into both prongs.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: The first thing we do with a potential ineffectiveness claim is read the case the way the Strickland Court read Washington’s: from the lawyer’s chair, on the day the decision was made, with only the information the lawyer had. If a decision looks defensible from that seat — even if it turned out badly — the claim usually needs a different theory or more record development, not more adjectives. The claims that move are the ones where the record shows the lawyer did not have the information at all, and had no reasonable basis for not going to get it.

What does “deficient performance” mean under Strickland?

Deficient performance means representation that “fell below an objective standard of reasonableness.” That is the whole of the legal test, and each word in it carries weight.

Objective. Harrington v. Richter, 562 U.S. 86 (2011), makes the point directly: “Strickland … calls for an inquiry into the objective reasonableness of counsel’s performance, not counsel’s subjective state of mind.” A lawyer’s affidavit saying he had a strategy is evidence, not an answer; a lawyer’s admission that he had no strategy is likewise not the end, because the question is whether a reasonable lawyer could have done what he did.

Standard of reasonableness. Strickland declined to write a checklist. It says the proper standard “is that of reasonably effective assistance,” measured against “[p]revailing norms of practice as reflected in American Bar Association standards and the like,” which Wiggins v. Smith, Warden, 539 U.S. 510 (2003), confirms “are guides to determining what is reasonable.” Guides, not rules. A violation of an ABA standard is not automatically a constitutional violation, and compliance with one is not automatically a defense.

The strong presumption, and what actually overcomes it

Before a court weighs any of that, it applies a thumb to the scale:

“A court must indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.”

Strickland continues: “that is, the defendant must overcome the presumption that, under the circumstances, the challenged action ‘might be considered sound trial strategy.'” Notice the conditional — might be. The court is not asking whether this lawyer had a strategy. It is asking whether the conduct could be characterized as one. Richter puts the same idea in operational terms: “Counsel was entitled to formulate a strategy that was reasonable at the time and to balance limited resources in accord with effective trial tactics and strategies.”

This is the single largest reason strong-sounding claims fail, and it is why the most productive question in a case evaluation is not “was this a mistake?” but “is there any reasonable professional explanation for this on the record as it stood?”

The presumption is not unbeatable, and Wiggins shows exactly where it breaks. Counsel in Wiggins looked at a presentence report and social-services records, then stopped, and later characterized the decision to present a different sentencing theme as strategy. The Court refused the characterization because the investigation behind it was inadequate, quoting Strickland‘s own limit:

“[S]trategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable; and strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation. In other words, counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.”

That is the crack in the presumption. Deference runs to informed choices. A choice made without the information a reasonable lawyer would have gathered is not a strategic choice at all — it is the absence of one, and the court says so.

Kimmelman v. Morrison, 477 U.S. 365 (1986), is the same principle in a non-capital setting. Counsel there failed to file a timely suppression motion, not for tactical reasons but because he had conducted no pretrial discovery at all, operating on the mistaken beliefs that the State had to volunteer its inculpatory evidence and that the victim’s preferences would determine whether the case went to trial. The Court found the decision “unreasonable, that is, contrary to prevailing professional norms” — even while applying “a heavy measure of deference” to counsel’s judgment. Ignorance of a basic legal obligation is not a judgment the Constitution defers to.

Hindsight is expressly excluded

Strickland commands that “every effort [must] be made to eliminate the distorting effects of hindsight, to reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the conduct from counsel’s perspective at the time” — language Bell v. Cone, 535 U.S. 685, 698 (2002), quotes back when applying it. Practically, this means the fact that a defense failed is not evidence that the lawyer was unreasonable to try it, and the fact that a better witness surfaced years later does not make it unreasonable that the lawyer did not call that witness at trial. It also means a movant’s affidavit describing what he now wishes had happened is worth very little; contemporaneous documents — the file, the discovery log, the correspondence, the notes — are worth a great deal.

What is a “reasonable probability” of prejudice?

Prejudice under Strickland means “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,” and the Court then supplied the definition that decides most cases:

“A reasonable probability is a probability sufficient to undermine confidence in the outcome.”

Two boundaries sit on either side of that sentence, and both come from the opinion itself.

The floor. Strickland rejects the argument that any conceivable effect is enough: “It is not enough for the defendant to show that the errors had some conceivable effect on the outcome of the proceeding,” because “[v]irtually every act or omission of counsel would meet that test.”

The ceiling. Strickland also rejects a preponderance standard. The movant does not have to prove the error “more likely than not altered the outcome.” That is why “reasonable probability” is a term of art rather than a percentage, and why a court can find prejudice while acknowledging the movant would probably still have been convicted.

The prejudice inquiry is a whole-record inquiry

Strickland directs that “[i]n making the determination whether the specified errors resulted in the required prejudice, a court … must consider the totality of the evidence before the judge or jury.” It draws the practical consequence itself: a verdict only weakly supported by the record is more likely to have been affected by counsel’s errors than one with overwhelming support.

The opinion also fences off what a court may not consider. The prejudice assessment “should proceed on the assumption that the decisionmaker is reasonably, conscientiously, and impartially applying the standards that govern the decision.” It “should not depend on the idiosyncracies of the particular decision-maker, such as unusual propensities toward harshness or leniency,” and evidence about a particular judge’s sentencing practices “should not be considered in the prejudice determination.” A movant cannot argue prejudice on the theory that his judge was unusually severe. He must argue it against a hypothetical, law-abiding decisionmaker on the record as it would have been.

Prejudice at sentencing has no minimum

Because the prejudice question is about the “result of the proceeding,” and sentencing is a proceeding, an error that produced additional prison time is prejudicial even if the amount is small. Glover v. United States, 531 U.S. 198 (2001), rejected a Seventh Circuit rule requiring the increase to be “significant,” holding that “any amount of actual jail time has Sixth Amendment significance” and that the amount “cannot serve as a bar to a showing of prejudice” under a determinate system like the Sentencing Guidelines. The Court noted the practical reason: “there is no obvious dividing line by which to measure how much longer a sentence must be for the increase to constitute substantial prejudice.”

This matters more than it sounds. Guidelines-calculation errors that counsel failed to object to are among the most tractable claims on the list of grounds for a §2255 motion, precisely because Glover removes the argument that a few months do not count.

Table 1. The Strickland framework, prong by prong — the Court’s operative language and the specific showing a movant must make.

Element	The Court’s operative language (Strickland, 466 U.S. 668)	What the movant must actually show	Where the claim usually dies
Prong 1 — Deficient performance	“the defendant must show that counsel’s representation fell below an objective standard of reasonableness”	A specific act or omission, identified in the record, that no objectively reasonable defense lawyer would have made or omitted on the information available at that time.	The court supplies a reasonable explanation the lawyer never gave, because the conduct “might be considered sound trial strategy.”
Prong 1 — The presumption	“A court must indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance.”	Affirmative rebuttal — record evidence that forecloses every reasonable professional explanation, not merely an assertion that the choice was wrong.	The movant argues the decision was bad without addressing why it could not have been reasonable.
Prong 1 — Point of view	“every effort [must] be made to eliminate the distorting effects of hindsight … and to evaluate the conduct from counsel’s perspective at the time”	The state of counsel’s knowledge on the date of the decision: the file, the discovery produced, what the client told him, what a reasonable investigation would have turned up.	The claim is built from evidence that surfaced after trial and is judged as if counsel had it.
Prong 1 — Duty to investigate	“counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary” (Strickland, quoted and applied in Wiggins v. Smith, Warden, 539 U.S. 510 (2003))	That the investigation actually performed was too thin to support the choice counsel now calls strategic — with proof of what a reasonable investigation would have produced.	The movant shows the investigation was limited but never shows what it would have found.
Prong 2 — Prejudice	“there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different”	A concrete, record-anchored account of the different result: the suppressed evidence, the acquittal-supporting witness, the lower guideline range, the plea not taken.	The movant alleges a different result without identifying what specifically would have changed.
Prong 2 — The definition	“A reasonable probability is a probability sufficient to undermine confidence in the outcome.”	Enough to shake confidence — less than “more likely than not,” more than “some conceivable effect on the outcome.”	The showing sits at the “conceivable effect” end, which Strickland expressly holds insufficient.
Prong 2 — Scope of review	a court “must consider the totality of the evidence before the judge or jury”	That the error mattered against the whole record, including evidence the movant would rather not discuss.	Overwhelming untouched evidence of guilt absorbs the error.
Prong 2 — At sentencing	“any amount of actual jail time has Sixth Amendment significance” (Glover v. United States, 531 U.S. 198, 203 (2001))	Any increase in the sentence traceable to the error; there is no de minimis threshold.	The movant cannot tie the error to a specific, calculable change in the range or sentence.
Order of decision	“there is no reason for a court … to address both components of the inquiry if the defendant makes an insufficient showing on one”	Nothing — this is the court’s option, not the movant’s. Both prongs should be briefed as if the court will reach only one.	The motion develops performance in detail and treats prejudice as an afterthought.

Source: 28 U.S.C. §2255; Strickland v. Washington, 466 U.S. 668 (1984); Wiggins v. Smith, Warden, 539 U.S. 510 (2003); Glover v. United States, 531 U.S. 198 (2001). Quoted language is from the opinions.

Does the court have to address both prongs?

No. Strickland expressly relieves courts of any obligation to reach both, and tells them which one to prefer:

“Although we have discussed the performance component of an ineffectiveness claim prior to the prejudice component, there is no reason for a court deciding an ineffective assistance claim to approach the inquiry in the same order or even to address both components of the inquiry if the defendant makes an insufficient showing on one. … If it is easier to dispose of an ineffectiveness claim on the ground of lack of sufficient prejudice, which we expect will often be so, that course should be followed.”

The Court adds the reason: “The object of an ineffectiveness claim is not to grade counsel’s performance.” A court has no institutional interest in publishing a finding that a named lawyer performed unreasonably, and a prejudice ruling avoids it entirely.

This has a direct drafting consequence. Because most denials are prejudice denials, a §2255 motion that develops the performance prong across fifteen pages and gives prejudice a paragraph is built to lose on the ground the court is most likely to use. Prejudice should be pleaded with the same specificity as performance: what the alternative record would have contained, how it interacts with the evidence that was presented, and why confidence in the result is undermined. Some courts will still reach performance — the Eleventh Circuit noted in an April 2026 unpublished decision that while the Supreme Court permits resolution on prejudice alone, a reviewing court “[is] not precluded from addressing both” — but no movant should count on it.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: When we build a §2255 motion around ineffectiveness, prejudice gets its own factual record, not a concluding sentence. That usually means an affidavit or declaration from the witness who was never called, the report from the expert who was never retained, or the corrected guideline calculation with the arithmetic on the page. A judge deciding whether confidence in the outcome is undermined needs to see the alternative proceeding, not be told it existed.

How is Strickland prejudice different from harmless-error review?

They ask a similar question and allocate the burden in opposite directions, which is why conflating them is a costly error.

On direct appeal, once a defendant shows a preserved constitutional trial error, the government bears the burden of proving the error harmless. Under Strickland, the movant bears the burden of affirmatively proving prejudice — the opinion says so: the defendant must “affirmatively prove prejudice.” There is no stage at which the government must disprove it, and no presumption that an unreasonable act by counsel harmed the defense.

Strickland also declines to adopt an outcome-determinative test, which is what makes the standard workable at all. And the Court situated its prejudice standard alongside, not inside, harmless-error doctrine: prejudice under Strickland is an element of the constitutional violation itself, not a remedial filter applied after a violation is found. As the Court has put it elsewhere, a Sixth Amendment violation of this kind “is not ‘complete’ until the defendant is prejudiced.”

The most important application of that distinction involves structural error — the small category of defects that, when preserved and raised on direct review, produce automatic reversal without any inquiry into harm. Weaver v. Massachusetts, 582 U.S. 286 (2017), holds that when a defendant does not preserve a structural error and instead raises it later through an ineffective-assistance claim, “the defendant generally bears the burden to show deficient performance and that the attorney’s error ‘prejudiced the defense.'” The structural label does not travel with the claim into collateral review. A movant whose lawyer failed to object to a courtroom closure does not get automatic reversal; he gets Strickland.

That is a trap worth naming, because a great deal of published guidance still describes structural errors as automatically reversible without noting that the automatic-reversal rule is a direct appeal, preserved-error rule. If the error reaches a federal court through §2255 as an ineffectiveness claim — which is how most of them reach a federal court — the Strickland burden applies. It is also a reminder that choosing the wrong vehicle costs the claim its most favorable standard, which is why the direct appeal versus §2255 decision has to be made deliberately rather than by default.

When is prejudice presumed? United States v. Cronic and its narrow exceptions

United States v. Cronic, 466 U.S. 648 (1984), was decided the same day as Strickland and answers the companion question: are there failures so complete that a court should not bother measuring their effect? The Court said yes, and identified a short list:

“There are, however, circumstances that are so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified.”

Cronic names three:

1. Complete denial of counsel at a critical stage. “Most obvious, of course, is the complete denial of counsel. The presumption that counsel’s assistance is essential requires us to conclude that a trial is unfair if the accused is denied counsel at a critical stage of his trial.”
2. Total failure to test the prosecution’s case. “[I]f counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing, then there has been a denial of Sixth Amendment rights that makes the adversary process itself presumptively unreliable.”
3. Circumstances where no lawyer could have helped. Where “the likelihood that any lawyer, even a fully competent one, could provide effective assistance is so small that a presumption of prejudice is appropriate without inquiry into the actual conduct of the trial” — the situation in Powell v. Alabama, where counsel was effectively designated on the morning of a capital trial.

Cronic itself found none of them present. A lawyer given twenty-five days to prepare a complex mail-fraud case, with no prior trial experience, did not trigger the presumption; the claim had to go through Strickland.

The critical limit came eighteen years later. Bell v. Cone, 535 U.S. 685 (2002), addressed the second exception and held the failure must be total:

“When we spoke in Cronic of the possibility of presuming prejudice based on an attorney’s failure to test the prosecutor’s case, we indicated that the attorney’s failure must be complete. We said ‘if counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing.'”

The Court framed the distinction categorically: “For purposes of distinguishing between the rule of Strickland and that of Cronic, this difference is not of degree but of kind.” Counsel in Bell participated in the sentencing hearing but presented no mitigating evidence and waived closing argument. Those are “specific attorney errors” governed by Strickland, not a Cronic collapse. The practical upshot: a lawyer who showed up, cross-examined, and argued is almost never a Cronic case, however badly he did it.

Conflicts of interest occupy a middle position. Cronic noted that ineffectiveness “is also presumed when counsel ‘actively represented conflicting interests,'” citing Cuyler v. Sullivan, 446 U.S. 335 (1980). Cuyler holds that a defendant who raised no objection at trial “must demonstrate that an actual conflict of interest adversely affected his lawyer’s performance,” and that a defendant who makes that showing “need not demonstrate prejudice in order to obtain relief.” That is a lighter burden than Strickland‘s but not a free pass — the adverse-effect showing has to be made, and the reach of Cuyler beyond multiple representation is genuinely unsettled. That analysis belongs to conflict of interest in a criminal case.

Is Strickland reviewed the same way in a §2255 motion and a §2254 petition?

No, and this is the most consequential distinction on the page for a federal prisoner.

Harrington v. Richter, 562 U.S. 86 (2011), involved a state prisoner seeking federal habeas relief under 28 U.S.C. §2254. Because a state court had already rejected the ineffectiveness claim on the merits, AEDPA’s §2254(d) barred relief unless that rejection was contrary to, or an unreasonable application of, clearly established Supreme Court law. The Court explained what happens when the two standards stack:

“Establishing that a state court’s application of Strickland was unreasonable under § 2254(d) is all the more difficult. The standards created by Strickland and § 2254(d) are both ‘highly deferential,’ and when the two apply in tandem, review is ‘doubly’ so.”

And then the operative reformulation: “When § 2254(d) applies, the question is not whether counsel’s actions were reasonable. The question is whether there is any reasonable argument that counsel satisfied Strickland‘s deferential standard.” Clark v. Sweeney, 607 U.S. 7 (2025) (per curiam), restates the same rule as recently as November 2025.

None of that applies to a federal prisoner’s §2255 motion. A §2255 movant is asking the sentencing court to review its own judgment in the first instance. There is no prior state-court merits adjudication to defer to, so §2254(d) has no work to do. The district court applies Strickland directly, once. Competitor pages routinely blur this, quoting Richter‘s “doubly deferential” line as if it described every ineffectiveness claim in federal court. It does not. What AEDPA did change for federal prisoners — the one-year clock, the successive-motion gate, the certificate of appealability — is covered at AEDPA, and the federal-versus-state divide itself is worked out at §2255 versus §2254.

Clark v. Sweeney also supplies a rule that applies with full force in §2255 practice. The Fourth Circuit had granted relief on a theory the petitioner never advanced, recasting a narrow failure-to-investigate claim as a “combination of extraordinary failures from juror to judge to attorney.” The Supreme Court reversed, holding that the “radical transformation” of the claim “departed so drastically from the principle of party presentation as to constitute an abuse of discretion,” and reminding lower courts that they “call balls and strikes; they don’t get a turn at bat.” The lesson for a movant is blunt: no court is going to improve your claim for you. The theory has to be pleaded, and pleaded precisely.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: Clark v. Sweeney is the reason we plead each ineffectiveness theory as a separate, self-contained ground with its own facts and its own prejudice showing, rather than as one omnibus “counsel failed at every stage” claim. A court reading an omnibus claim will pick the weakest version of it and deny that. A court reading five discrete grounds has to address five discrete grounds — and the record we build for each one is the record that carries into any certificate of appealability request later.

How does Strickland apply at a guilty plea, at trial, at sentencing, and on appeal?

The two prongs never change. What changes is the content of the prejudice question, because prejudice is defined by reference to “the result of the proceeding,” and each stage has a different result.

At a guilty plea. Hill v. Lockhart, 474 U.S. 52 (1985), holds “that the two-part Strickland v. Washington test applies to challenges to guilty pleas based on ineffective assistance of counsel,” and defines prejudice as “a reasonable probability that, but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.” Jae Lee v. United States, 582 U.S. 357 (2017), then corrected a widespread misreading: the Hill inquiry “focuses on a defendant’s decisionmaking, which may not turn solely on the likelihood of conviction after trial.” A movant need not show he would probably have been acquitted. But Lee also imposed a discipline: “Courts should not upset a plea solely because of post hoc assertions from a defendant about how he would have pleaded but for his attorney’s deficiencies. Rather, they should look to contemporaneous evidence to substantiate a defendant’s expressed preferences.” What kinds of plea advice have been held deficient, and how contemporaneous evidence is assembled, is the subject of ineffective assistance in a guilty plea.

At a lost or rejected plea offer. Missouri v. Frye, 566 U.S. 134 (2012), holds “that, as a general rule, defense counsel has the duty to communicate formal offers from the prosecution to accept a plea on terms and conditions that may be favorable to the accused,” and that the right “extends to the consideration of plea offers that lapse or are rejected.” Lafler v. Cooper, 566 U.S. 156 (2012), sets the prejudice showing where bad advice caused a rejection: a reasonable probability that the offer would have been presented to the court, that the court would have accepted its terms, and that the resulting conviction or sentence “would have been less severe than under the judgment and sentence that in fact were imposed.” Lafler also describes the remedy, which is unusual — a court may order the plea reoffered and then decide whether to vacate the trial conviction. A collateral-attack waiver in the plea agreement complicates but frequently does not bar these claims; that question belongs to appeal waivers and collateral-attack waivers.

At trial. This is the setting Strickland was written for, and the standard operates as stated: an objectively unreasonable act or omission, plus a reasonable probability of a different verdict measured against the totality of the evidence. Kimmelman v. Morrison adds an important wrinkle for suppression-based claims. A Fourth Amendment claim itself is generally unavailable on federal habeas, but Kimmelman holds that limitation does not extend to a Sixth Amendment claim founded on counsel’s incompetent handling of a suppression issue. The price is a compound showing: the movant “must also prove that his Fourth Amendment claim is meritorious and that there is a reasonable probability that the verdict would have been different absent the excludable evidence.”

At sentencing. The proceeding is the sentencing hearing and the result is the sentence, so Glover v. United States controls: any additional prison time attributable to counsel’s error is prejudice. Wiggins is the capital analogue — where counsel failed to investigate mitigation, the court “reweigh[s] the evidence in aggravation against the totality of available mitigating evidence.”

On appeal. Two different rules operate here, and the distinction matters. Where counsel filed the appeal but omitted an issue, Strickland applies in the ordinary way, and the prejudice question is whether there is a reasonable probability the omitted issue would have succeeded — the subject of ineffective appellate counsel. Where counsel’s failure cost the defendant the appeal itself, Roe v. Flores-Ortega, 528 U.S. 470 (2000), applies a presumption: “when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken, the defendant has made out a successful ineffective assistance of counsel claim entitling him to an appeal.” The movant need not identify the issues he would have raised or show the appeal had merit. Flores-Ortega also recognizes a duty to consult about an appeal where “a rational defendant would want to appeal” or the defendant “reasonably demonstrated to counsel that he was interested in appealing.” Garza v. Idaho, 586 U.S. 232 (2019), confirms that “prejudice is presumed” in “certain Sixth Amendment contexts,” including “when counsel’s constitutionally deficient performance deprives a defendant of an appeal that he otherwise would have taken.” That remedy is worked out at my lawyer never filed the appeal I asked for.

Why is Strickland so hard to win?

Because the opinion is built to be hard to win, and it is honest about why.

The Court was explicit that “[j]udicial scrutiny of counsel’s performance must be highly deferential,” and it explained the institutional reason: an unrestrained ineffectiveness inquiry “would encourage the proliferation of ineffectiveness challenges,” would chill vigorous advocacy, and would make the lawyer’s post-trial role adverse to the client’s. Every design feature of the test flows from that concern. The presumption of reasonableness, the exclusion of hindsight, the permission to skip the performance prong, the totality-of-the-evidence framing of prejudice — each one narrows the opening.

Layer on the procedural setting and the difficulty compounds. An ineffectiveness claim usually arrives on collateral review, after a conviction is final, when the finality interest is at its highest. It has to clear the one-year §2255 deadline, and if that date has passed the movant needs equitable tolling or the narrow actual innocence gateway. It generally requires an evidentiary record the movant does not yet have, which means the fight is often about whether the court will hold a §2255 evidentiary hearing at all. And a second §2255 motion raising an ineffectiveness claim the movant did not raise the first time runs into the second or successive §2255 gate, which ineffectiveness claims rarely clear.

None of that means the claim is hopeless. It means the claim is a record-building exercise before it is an argument. The distribution of outcomes across all federal §2255 filings, and what the available data does and does not show, is set out at §2255 grant rates and timing. If you have not filed yet, the mechanics are at how to file a §2255 motion; if you have, the sequence and timing are at what happens after filing a §2255.

Applied Insight — Elizabeth Franklin-Best, Esq., Principal Attorney: The cases that survive Strickland almost always share a structural feature: there is a document, or a witness, or a calculation that exists in the world and did not make it into the proceeding. Not a different argument that could have been made — a piece of reality that was available and absent. When we evaluate a case, that is what we look for first, because it is what a court can see. An argument about how counsel should have framed things reads as second-guessing. A report counsel never obtained reads as a record.

Why work with Elizabeth Franklin-Best, P.C.

Elizabeth Franklin-Best, P.C. is a boutique federal criminal defense and appellate firm founded in 2019, running a national federal practice on an intentionally small caseload. Elizabeth Franklin-Best has more than 20 years of federal and appellate criminal defense experience, has handled more than 330 federal proceedings, over 100 of them appeals, and has sixteen published decisions as counsel of record. She is admitted to the U.S. Supreme Court and twelve of the thirteen U.S. Courts of Appeals — every regional circuit (1st–11th) plus the D.C. Circuit — and has been admitted pro hac vice in federal courts from Montana to Florida.

She is recognized by Best Lawyers in America 2026 for Appellate Practice and ranked by Chambers USA 2026 for Litigation: White-Collar Crime & Government Investigations, was named South Carolina Public Defender of the Year in 2010, and is the author of Reversing Your Criminal Conviction: Develop Your Winning Strategy (2023). She serves as Board Chair of Justice 360 and as Co-Vice Chair of the NACDL Amicus Curiae Committee for the Fourth Circuit.

The firm’s post-conviction work is built around the discipline this page describes: reading the trial record from counsel’s seat, identifying the specific decision that cannot be explained as reasonable, and assembling the evidentiary record that makes the prejudice showing concrete rather than argumentative. With offices in Columbia and Mount Pleasant, South Carolina, the firm represents clients nationwide — a §2255 motion is filed in the district that imposed the sentence, wherever that is.

Talk to a federal post-conviction attorney about a Strickland claim

If you are evaluating whether a lawyer’s performance in your federal case can support a §2255 claim, the two questions that matter first are what the record shows about the decision you are challenging and how much time is left on your filing deadline. Both are answerable, and the second one is urgent.

Schedule a consultation directly, or call (843) 620-1100. The firm handles federal post-conviction matters in every district and circuit. If you are still deciding what kind of representation your case needs, start at federal 2255 motion lawyer, which covers what post-conviction counsel actually does and how these engagements work, or at post-conviction attorney if the direct appeal is already finished. Every deadline that may be running right now is inventoried at post-conviction deadlines.

Frequently Asked Questions About Strickland v. Washington

Reviewed for legal accuracy by Elizabeth Franklin-Best, Esq., Principal Attorney·September 2026