The State of New Mexico appeals from the district court’s grant of a writ of habeas corpus to prisoner Donald Duncan. This appeal raises two issues: (1) whether the doctrine of res judicata bars a defendant’s postconviction claim of ineffective assistance of counsel when the same claim has been denied on the merits in a direct appeal, and (2) whether the trial court erred in granting petitioner’s writ of habeas corpus on the grounds of ineffective assistance of counsel. We hold that the doctrine of res judicata does not apply in this particular case, and affirm the trial court’s grant of habeas corpus relief on the grounds of ineffective assistance of counsel.
Donald Duncan was convicted on six counts of criminal sexual penetration and incest in 1985. On direct appeal with a new attorney, Duncan argued that his conviction should be reversed because he did not receive effective assistance of counsel at trial. The Court of Appeals assessed and rejected the ineffective counsel claim on the merits and affirmed the conviction. Duncan then petitioned the district court for a writ of habeas corpus, requesting postconviction relief on the grounds of ineffective assistance of counsel. After holding an evidentiary hearing on the issue of *346counsel’s effectiveness, District Judge William J. Schnedar, who also had presided over Duncan’s jury trial, determined that Duncan had not received effective assistance of counsel at trial, granted the writ of habeas corpus, set aside Duncan’s conviction, and ordered a new trial. The judge found that material lapses in the trial attorney’s performance leading to a conclusion of incompetent representation included the failure to construct an available and crucial alibi defense, to call key witnesses, and to move to sever offenses.
I. RES JUDICATA IN HABEAS CORPUS PROCEEDINGS
The State argues that Duncan’s ineffective assistance of counsel claim is barred by the doctrine of res judicata in this postconviction action because the same claim was raised and rejected on direct appeal. We have held that New Mexico postconviction procedures are not a substitute for direct appeal and that our statutes do not require collateral review of issues when the facts submitted were known or available to the petitioner at the time of his trial. See State v. Gillihan, 86 N.M. 439, 440, 524 P.2d 1335, 1336 (1974). When a defendant should have raised an issue on direct appeal, but failed to do so, he or she may be precluded from raising the issue in habeas corpus proceedings. Id. at 440, 524 P.2d at 1336.1 A habeas corpus petitioner will not be precluded, however, from raising issues in habeas corpus proceedings that could have been raised on direct appeal either when fundamental error has occurred, Gillihan, 86 N.M. at 440, 524 P.2d at 1336, or when an adequate record to address the claim properly was not available on direct appeal, State v. Gomez, 112 N.M. 313, 314-15, 815 P.2d 166, 167-68 (Ct.App.), cert. denied, 112 N.M. 279, 814 P.2d 457 (1991). Our habeas corpus cases applying preclusion have not involved instances in which a defendant could not make an evidentiary case for himself on direct appeal because facts supporting his claim were not part of the record. Id. at 314, 815 P.2d at 167. The proper exception to the application of the doctrine of res judicata in such circumstances has been described as follows:
If an application [for a writ of habeas corpus] is grounded in facts beyond the record previously presented on appeal, and if the additional facts are those which could not, or customarily would not, be developed in a trial on criminal charges, there should be no issue preclusion____ When a post-conviction application makes a substantial showing that due process or another fundamental right has been abridged — and the application is supported by facts ill-suited for development in the original trial — it should be addressed on its merits. Res judicata does not apply.
Two New Mexico cases endorse SCRA 1986, 5-802 habeas corpus proceedings as the preferred avenue for adjudicating ineffective assistance of counsel claims. See State v. Powers, 111 N.M. 10, 12, 800 P.2d 1067, 1069 (Ct.App.), cert. denied, 111 N.M. 16, 801 P.2d 86 (1990) (stating that Court of Appeals cannot assess validity of ineffective assistance of counsel claim without the benefit of an evidentiary hearing on the issue, and that postconviction proceedings are recommended for such fact-finding); State v. Stenz, 109 N.M. 536, 539, 787 P.2d 455, 458 (Ct.App.), cert. denied, 109 N.M. 562, 787 P.2d 842 (1990) (same). The rationale is that even assuming that a criminal defendant has a new attorney to handle his direct appeal, the record before the trial court may not adequately document the sort of evidence essential to a determination of trial counsel’s effectiveness be*347cause conviction proceedings focus on the defendant’s misconduct rather than that of his attorney. Consequently, an evidentiary hearing on the issue of trial counsel’s effectiveness may be necessary. While this may be done through a remand to the trial court during direct appeal when unusual circumstances exist, habeas corpus is specifically designed to address such postconviction constitutional claims and is the procedure of choice in this situation. It would be inconsistent to recommend habeas corpus for ineffective assistance of counsel claims and then to deny a convicted defendant the ability to use habeas procedure because he was fortunate enough to obtain new counsel on direct appeal and raise the issue at that time. By expressly recommending habeas corpus for claims of ineffective counsel, Powers and Stem imply that res judicata will not bar such action.
The State calls our attention to Manlove v. Sullivan, 108 N.M. 471, 775 P.2d 237 (1989), as part of its argument that Duncan is precluded from relitigating his ineffective assistance of counsel claim in habeas corpus proceedings. Manlove observed that when a petitioner brings successive petitions for habeas relief, “collateral estoppel principles may, at the discretion of a subsequent habeas corpus court, prevent relitigation of issues argued and decided on a previous habeas corpus petition____” Id. at 475, 775 P.2d at 241. Clearly, Manlove is procedurally and theoretically distinguishable from this case because it deals with the preclusive effect of issues raised in successive petitions for writs of habeas corpus rather than with the preclusive effect of issues raised initially on direct appeal. The successive-writ petitioner has already enjoyed the opportunity to fully explore his constitutional claims in the post-conviction setting, whereas the petitioner who makes his initial claim on direct appeal has not, and consequently, the successive-writ petitioner is in a weaker position to argue that equity confers yet another post-conviction opportunity to make his claim.
Our view of the nature of res judicata in the habeas corpus setting as an equitable, discretionary, and flexible judicial doctrine leads us to conclude that the decision to give preclusive effect to an ineffective assistance of counsel claim rejected on direct appeal but raised again in habeas corpus proceedings is within the sound discretion of the reviewing court. Courts reviewing habeas corpus petitions should decide whether particular claims are res judicata based on relevant facts and circumstances, including whether or not a full evidentiary hearing on counsel’s effectiveness was previously conducted when that claim is at issue. See SCRA 5-802(E)(3) (Repl.Pamp.1992). While redundant claims may be precluded, courts should not impede posteonviction litigation that will provide necessary fuller or fairer procedural opportunities to examine alleged constitutional defects when consideration of an issue on direct appeal is based upon facts which could not, or customarily would not, be developed at trial. While the doctrine of res judicata or claim preclusion may apply in other postconviction circumstances, applying the doctrine in this habeas corpus action would be inconsistent with New Mexico precedent, and it would deprive the habeas corpus petitioner of the opportunity to meaningfully litigate his constitutional claim, thereby thwarting an essential purpose of habeas corpus procedure.
II. STANDARD OF REVIEW
When this Court addresses the propriety of a lower court’s grant or denial of a writ of habeas corpus based on ineffective assistance of counsel, findings of fact of the trial court concerning the habeas petition are reviewed to determine if substantial evidence supports the court’s findings. Cf. State v. Goss, 111 N.M. 530, 533-34, 807 P.2d 228, 231-32 (Ct.App.) (stating that the factual determination of whether a search and seizure is reasonable is reviewed on appeal for support by substantial evidence, and implying that although the existence of probable cause is a legal question, it is for the trial court to resolve disputed factual issues prior to that determination), cert. denied, 111 N.M. 416, 806 P.2d 65 (1991). Questions of law or questions of mixed fact and law, however, including the assessment of effective assis*348tance of counsel, are reviewed de novo. This approach conforms to our past approach in the evaluation of ineffective assistance of counsel claims in habeas corpus proceedings, see Manlove v. Sullivan, 108 N.M. 471, 475, 775 P.2d 287, 241 (1989) (deciding petition for writ of habeas corpus de novo on the merits, including claim of ineffective assistance of counsel), and parallels federal habeas corpus appellate review practice, see United States v. Owens, 882 F.2d 1493, 1501-02 n. 16 (10th Cir.1989) (stating that in federal habeas corpus cases, findings of fact are reviewed under the clearly erroneous standard, and application of the Strickland test is reviewed de novo).2 Such an approach provides logical deference to the trial court fact-finder as first-hand observer, while assuring that higher courts perform their sanctioned role as arbiter of the law.
III. INEFFECTIVE ASSISTANCE OF COUNSEL
Duncan’s claim of ineffective assistance of counsel was first presented on his direct appeal. Although the grounds alleged in support of Duncan’s ineffective assistance of counsel argument were basically the same on direct appeal and in the habeas corpus proceedings, the evidence available in the two proceedings was dramatically different. Duncan’s principal contention was that his trial counsel was ineffective because he did not investigate or adequately present Duncan’s alibi defense. Reviewing the record to evaluate this argument, the Court of Appeals found on direct appeal that, “the record does not contain any indication whether counsel did or did not investigate this [alibi] defense. It only indicates that counsel chose not to corroborate the defense through witnesses other than defendant.” State v. Duncan, No. 9218, slip op. at 3 (N.M.CtApp. Dec. 2, 1986).
The evidence available concerning counsel’s effectiveness was significantly more extensive at the three-day habeas corpus hearing than at Duncan’s direct appeal. Evidence presented at the habeas corpus hearing supporting Duncan’s claim included the following; Mr. Arnold Miller, Duncan’s trial attorney, did not call five credible and available alibi witnesses; these five witnesses were known to Miller and were prepared to testify that Duncan was with them when the crimes allegedly occurred; Miller did not call other witnesses who would have impeached key prosecution witnesses, especially the victims; and several of the defense witnesses who were not called were standing outside the courtroom during Duncan’s trial, waiting to be called. The State’s evidence at the habeas hearing consisted of Miller’s testimony, which attempted to rationalize his alleged failures.
Substantial evidence supports the lower court’s conclusion that Duncan did not receive effective assistance of counsel at trial. To succeed on a constitutional claim of ineffective assistance of counsel, the defendant or habeas corpus petitioner must prove that defense counsel did not exercise the skill of a reasonably competent attorney and that such incompetent representation prejudiced his case, rendering the trial court’s result unreliable. Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 2064, 80 L.Ed.2d 674 (1984); State v. Taylor, 107 N.M. 66, 72-73, 752 P.2d 781, 787-88 (1988), overruled on other grounds, Gallegos v. Citizens Ins. Agency, 108 N.M. 722, 731, 779 P.2d 99, 108 (1989). In considering an ineffectiveness claim, the evaluation should focus on whether an attorney exercised “the skill, judgment and diligence of a reasonably competent defense attorney” in light of the entire proceeding allegedly tainted by incompetent representation. State v. Orona, 97 N.M. 232, 233, 638 P.2d 1077, 1078 (1982). Be*349cause substantial evidence supports the district court’s findings of fact, we presume these findings to be correct, and undertake de novo review of counsel’s effectiveness.
Duncan’s trial attorney did not represent him competently. Based on the evidentiary hearing at the trial level of this habeas corpus action, Judge Schnedar found that Miller inexcusably failed to give notice of the alibi defense or call alibi witnesses crucial to Duncan’s defense. Miller also did not investigate possible attacks on the credibility of State’s witnesses or move to sever counts as related to each victim. Miller’s trial preparation was not adequate, and his intentional failure to disclose his mistake of not giving notice of alibi defense compounded his shortcomings, breaching his duty of loyalty to the defendant. We disagree with the State’s characterization of these failures as reasonable litigation tactics and find that such a performance falls below an objective level of competent legal representation.
We further agree with the district court that Duncan’s incompetent representation prejudiced his case such that but for counsel’s error, there is a reasonable probability that the result of the conviction proceedings would have been different. Although the State contends that the district court did not find prejudice as required by the New Mexico standard for ineffective assistance of counsel, the finding of prejudice is clearly implicit in the lower court’s findings. The trial court found that “Miller’s conduct undermined the proper function of the adversarial procedure and it cannot be relied upon as having produced a just result.” It further found that had Duncan received competent representation, his trial “could have resulted in a prosecution verdict, a defense verdict or a hung jury.” These findings sufficiently express the concept that but for counsel’s incompetence, the result of the trial probably would have been different. Consequently, the district court did find prejudice, and we will not elevate form above substance by accepting the State’s contention that prejudice was not found. Our own review of the trial court’s findings enforces our belief that Duncan’s defense was prejudiced by his counsel’s performance and that his conviction, therefore, cannot stand.
For the foregoing reasons, the decision of the district court granting habeas relief is affirmed.
IT IS SO ORDERED.
RANSOM, C.J., and FRANCHINI, concur.