Opinion

FOTI, J.

The petitioner, Efrain Hernandez, appeals following the habeas court’s denial of his petition for certification to appeal from the denial of his amended petition for a writ of habeas coipus. On appeal, the petitioner claims that the court improperly (1) denied his petition for certification to appeal, (2) concluded that he had effective assistance of counsel and (3) con-*703eluded that his plea was voluntary, knowing and intelligent. We agree with the petitioner and reverse the judgment of the habeas court.

The following facts and procedural history are relevant to the resolution of the petitioner’s claims. On July 26, 1993, the first day of testimony at the petitioner’s criminal trial, the petitioner withdrew his plea of not guilty and entered a plea of nolo contendere to one count of murder in violation of General Statutes § 53a-54a. On September 17, 1993, in accordance with the plea agreement, the petitioner was sentenced to a term of twenty-five years incarceration. On October 5, 1995, the petitioner filed a petition for a writ of habeas corpus, which he amended on July 6,1999, claiming ineffective assistance of counsel and that his plea was involuntary.

On January 31, 2002, the habeas court held a hearing on the petition. At the hearing, the petitioner testified that during a recess on the first day of testimony at the criminal trial, a plea offer was made by the state that entailed an agreed on sentence of twenty-five years in exchange for the petitioner’s plea. The petitioner further stated that his attorney, James J. Ruane, had advised him that he would be eligible for parole after serving 50 percent of his sentence. Last, the petitioner testified that the only reason he accepted the plea was that he would be eligible for parole and that had he known that he would be ineligible for parole, he would not have entered a plea of nolo contendere and would have continued with the trial.

Ruane also testified at the hearing on the petition. He testified that he had informed the petitioner “that on a twenty-five year sentence . . . his release would be governed by, at the time, what [he] thought was the parole authorities and, as a general rule, parole can be granted after serving half the sentence, but that their regulations change, and so it would be up to the parole *704board.” Ruane farther testified that although he is now aware that in 1992, there was no parole eligibility for a person convicted of murder, he was not aware of it at the time he advised the petitioner.1

On May 29, 2002, the court issued a memorandum of decision denying the petitioner’s amended petition for a writ of habeas corpus, concluding that the petitioner had failed to show that he was prejudiced by counsel’s deficient performance. Although the court concluded that “Ruane’s performance fell below the standard of reasonableness under prevailing professional norms and that . . . [his] advice regarding the petitioner’s parole eligibility was deficient,” it also concluded that the petitioner failed “to affirmatively show that he would have proceeded with the trial had he known that he would be ineligible for parole .... [and that the] petitioner has not shown that but for counsel's deficient performance, the result of the proceedings would have been different.”

On June 7, 2002, the petitioner sought to appeal to this court from the habeas court’s judgment by filing a petition for certification to appeal, which the habeas court denied on August 30, 2002. On September 16, 2002, this appeal was filed. Additional facts relevant to the petitioner’s claims will be set forth as necessary.

We begin by setting forth our standard of review. “Generally, [t]he conclusions reached by the [habeas] court in its decision to dismiss the habeas petition are matters of law, subject to plenary review. . . . Thus, [w]here the legal conclusions of the court are challenged, we must determine whether they are legally and *705logically correct . . . and whether they find support in the facts that appear in the record. ... In a habeas appeal, although this court cannot disturb the underlying facts found by the habeas court unless they are clearly erroneous, our review of whether the facts as found by the habeas court constituted a violation of the petitioner’s constitutional right to effective assistance of counsel is plenary.” (Internal quotation marks omitted.) Perez v. Commissioner of Correction, 80 Conn. App. 96, 99, 832 A.2d 1210, cert. denied, 266 Conn. 934, 837 A.2d 805 (2003).

The petitioner first claims that the court improperly concluded that he had effective assistance of counsel.2 Specifically, the petitioner argues that the court improperly found that he was not prejudiced by Ruane’s deficient performance. We agree.

“In Strickland v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984), the United States Supreme Court enunciated the two requirements that must be met before a petitioner is entitled to reversal of a conviction due to ineffective assistance of counsel. First, the [petitioner] must show that counsel’s performance was deficient. . . . Second, the [petitioner] must show that the deficient performance prejudiced the defense. . . . Unless a [petitioner] makes both showings, it cannot be said that the conviction . . . resulted from a breakdown in the adversarial process that renders the result unreliable.” (Internal quotation marks omitted.) Henderson v. Commissioner of Correction, 80 Conn. App. 499, 503-504, 835 A.2d 1036 *706(2003) , cert. denied, 267 Conn. 918, 841 A.2d 1190 (2004) .

“For ineffectiveness claims resulting from guilty pleas, we apply the standard set forth in Hill v. Lockhart, 474 U.S. 52, 106 S. Ct. 366, 88 L. Ed. 2d 203 (1985), which modified Strickland’s prejudice prong. ... To satisfy the prejudice prong, the petitioner must show a reasonable probability that, but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.” (Internal quotation marks omitted.) Perez v. Commissioner of Correction, supra, 80 Conn. App. 99. “A reasonable probability is a probability sufficient to undermine confidence in the outcome. ... A different result must be sufficiently probable to undermine confidence in the actual outcome.” (Citation omitted; internal quotation marks omitted.) Falby v. Commissioner of Correction, 32 Conn. App. 438, 443, 629 A.2d 1154, cert. denied, 227 Conn. 927, 632 A.2d 703 (1993).

Here, the court concluded that the petitioner did not satisfy the second prong3 of the Hill-Strickland test requiring a showing of prejudice because the petitioner’s decision to enter a plea of nolo contendere was not based solely on parole eligibility, but rather on the likelihood that he would have been convicted of murder at trial and possibly subjected to an even longer sentence. The court’s conclusion is incorrect and finds no support in the record because a different result at trial was “sufficiently probable.” Specifically, we note that (1) the petitioner’s self-defense claim and (2) the exclu*707sion of a state’s witness create a sufficient probability of a different result had the petitioner not entered a plea of nolo contendere and continued with the trial.4

First, the petitioner’s self-defense claim creates a reasonable probability of a different result at trial. Specifically, the petitioner testified at the habeas trial that he was a former member of a gang and that members of the gang had been tiying to kill him since he left the gang.5 The petitioner further testified that on the date of the shooting at issue he saw an individual in the victim’s car with what appeared to be a gun and that he shot his gun at the victim’s car before the person in the car could shoot at him first. Additionally, Ruane testified that the petitioner was carrying a gun on the night of the incident and that he fired many times at the car as it was driving away, but he believed that the jury could find that he was acting in self-defense. Furthermore, Ruane testified that there was “bad blood” between the petitioner and the petitioner’s family and the victim’s family, which would bear on a self-defense claim, both positively and negatively. Although that testimony does not create a certainty that a self-*708defense claim would succeed at trial, we conclude that it does create a probability sufficient to undermine a murder conviction.

Next, the exclusion of testimony from one of the state’s two witnesses further supports the conclusion that there was a reasonable probability of a different result had the trial continued. Specifically, the state sought to introduce certain testimony of Ruben Gonzalez to prove the petitioner’s motive to commit murder, but the court, Ford, J., excluded that testimony because it was not relevant to motive and intent. Because the state was scheduled to call only two witnesses, the exclusion of the only testimony as to intent, which is an essential element in proving murder, creates a sufficient probability that had the petitioner continued with the trial, he would not have been convicted of murder.

For the foregoing reasons,6 we conclude that because there is a probability sufficient to undermine confidence in the murder conviction, it is reasonably, probable that, but for counsel’s deficiencies, the petitioner would not have entered a plea of nolo contendere and instead would have continued with the trial. We therefore conclude that the petitioner was prejudiced by counsel’s deficiencies and, furthermore, that he had ineffective assistance of counsel.

The petitioner next claims that his plea was involuntary because (1) his attorney erroneously advised him *709of his parole eligibility and (2) he believed he was entering a plea to a charge of manslaughter and not to the charge of murder. We agree with the petitioner’s first argument and, therefore, need not address the second claim.

“Guilty pleas must be intelligent, voluntary and knowing. ... A defendant must be aware of all direct consequences of his plea.” (Citation omitted.) Falby v. Commissioner of Correction, supra, 32 Conn. App. 444-45. Where “the petitioner relied on gross misadvice about an indirect consequence, his plea would have been involuntary, unintelligent and, therefore, invalid. . . . Examples of gross misadvice that invalidate a guilty plea include an attorney’s poor research and incorrect calculation of a defendant’s parole eligibility.” (Citation omitted.) Id., 446-47.

Here, the habeas court explicitly found, and we agree, that “Ruane’s performance fell below the standard of reasonableness under prevailing professional norms and that [his] advice regarding the petitioner’s parole eligibility was deficient.” For that reason, and for the reasons discussed in part I as to why a reasonable probability exists that, but for counsel's deficiencies, the petitioner would not have entered a plea of nolo contendere, we conclude that the court improperly determined that the petitioner’s plea was intelligent, voluntary and knowing.

The judgment is reversed and the case is remanded with direction to render judgment granting the petition for a writ of habeas corpus and for further proceedings in accordance with law.

In this opinion the other judges concurred.

See General Statutes (Rev. to 1991) § 54-125a (b), which provides in relevant part: “No person convicted of any of the following offenses, which was committed on or after July 1, 1981, shall be eligible for parole under subsection (a) of this section . . . murder, as defined in section 53a-54a . . . .” The event at issue in this case took place in June, 1992.

The court denied the petitioner’s petition for certification to appeal to this court. We agree with the petitioner that the issues raised in this appeal warrant review by this court and, accordingly, reach the merits of the appeal. See Copas v. Commissioner of Correction, 234 Conn. 139, 150-52, 662 A.2d 718 (1995).

We note that the petitioner has satisfied the first prong of the Hill-Strickland test requiring a showing that counsel’s performance was deficient. Specifically, we agree with the court’s conclusion that “Ruane’s performance fell below the standard of reasonableness under prevailing professional norms and that [his] advice regarding the petitioner’s parole eligibility was deficient.” Because we conclude that the petitioner has satisfied the first prong of the Hill-Strickland test, we need only address the second prong.

The court also based its conclusion on the notion that Ruane did not convey any certainty to the petitioner about the actual time he would have to serve because a parole board is not required to “ ‘consider any inmate’s eligibility for parole, even if he has fulfilled the statute’s . . . requirements.’ ” The court’s conclusion, however, is misguided and finds no support in the record. Even though the petitioner would not be guaranteed parole or consideration for parole if he were “eligible,” eligibility in general still puts the petitioner in a better position than he would have been in if he were statutorily ineligible, where he would have no chance of even being considered for parole status. The petitioner, therefore, was reasonable in basing his decision to plead on the fact that he would be “eligible” for parole even though he was not guaranteed that he would obtain it successfully or be considered for it.

The petitioner’s belief that members of his former gang were trying to kill him was based on the fact that he had been shot at several times in 1992 by individuals the petitioner believed to be gang members. Further, because he feared for his safety, the petitioner began carrying a gun for protection after leaving the gang.

We also note that the trial judge and the sentencing judge made comments that cast further doubt on whether a murder conviction would result if the trial continued. On the first day of trial, after the state put forward its last scheduled witness, the court, Ford, J., stated that “if all the facts are as is that we heard, and I accept them as true, the best you have is a manslaughter under the extreme indifference to [human life].” Next, at sentencing, the court, Damiani, J., stated that “[b]ased upon the facts of the case, if the case had gone to a conclusion through a trial, a jury could have come back with a guilty of manslaughter or they might have bought a self-defense.”