Glen J. Breese was convicted of murder in the first degree on February 1, 1979, after being tried on theories of deliberate premeditation and extreme atrocity or cruelty. We affirmed the conviction in a decision rendered June 26, 1980. See
Commonwealth
v.
Breese,
We hold that
Perry
is the law only in cases tried after our decision in
Commonwealth
v.
Gould,
In deciding whether
Perry
is retroactive, we first must resolve whether the decision announced a new rule. Decisional law usually is retroactive. When a decision announces a new rule, however, the issue arises whether it will be applied only prospectively. The issue is resolved by the application of a three-part test which we discuss later in this
*542
opinion. That test is not triggered, however, unless the rule announced is new. The requirement of newness has been stated in various ways. In
Chevron Oil Co.
v.
Huson,
Perry did not announce a new rule. The decision was “clearly foreshadowed” by Gould. In Gould, supra at 684-685, we reasoned that, “if a malicious mind may be considered as evidence that a defendant committed a murder with extreme atrocity or cruelty, then fairness requires that an impaired mind may also be considered as evidence bearing on whether or not the defendant committed the murder with extreme atrocity or cruelty.” We reasoned further that “[t]he jurors’ broad discretion [in determining whether a murder was committed with extreme atrocity or cruelty] will more accurately reflect the community’s conscience, goals, and norms, if the jurors are . . . permitted to consider the defendant’s peculiar mental state . ...” Id. at 685. Finally, we remarked that “[i]mpairment of a defendant’s ability to make a decision in a normal manner may have a direct bearing on the degree of murder, and consequently, on the issue of extreme atrocity or cruelty.” Id. at 686. In Perry, we recognized that a defendant’s mental state may be affected by intoxication as well as by mental illness. Our holding that the jury must be permitted to consider the defendant’s intoxication in determining his degree of culpability was based exclusively on Gould.
*543 Since Perry did not announce a new rule, but was clearly foreshadowed by Gould, we hold that the law as stated in Perry applies retroactively to all trials conducted after the date of the Gould decision. See Whitinsville Plaza, Inc. v. Kotseas, 378 Mass. 85, 97-98 (1979). Reaching this conclusion, however, does not help Breese, who was tried before Gould was decided. Consequently, we must consider whether Perry is retroactive to cases tried prior to Gould. In making this determination, we consider whether Gould is retroactive. Perry was a mere application of principles set forth in Gould. The issue of their retroactive effect should be resolved consistently.
In determining whether
Gould
is retroactive, we first examine the opinion. In
Gould,
we reversed the first degree murder conviction of a mentally impaired defendant who had been tried under a theory of extreme atrocity or cruelty. We did so, not because the judge’s instructions on extreme atrocity or cruelty were erroneous, but pursuant to our duty under G.L. c. 278, § 33E, “to consider broadly the whole case on the law and the facts to determine whether the verdict is ‘consonant with justice.’”
Gould,
In five direct appeals argued subsequent to
Gould,
defendants convicted prior to
Gould
sought either a reversal or a modification of their convictions on the ground that the instructions at their trials on deliberate premeditation or extreme atrocity or cruelty were not consistent with the principles announced in
Gould.
In
Commonwealth
v.
Cunneen, ante
216 (1983),
Commonwealth
v.
Brown, 387
Mass. 220,
*544
225-227 (1982), and
Commonwealth
v.
Shelley,
We have never given retroactive effect to Gould. This case requires us to decide whether the law as stated in Gould applies retroactively to the defendant’s trial. If Gould is retroactive to any date before it was decided, it must be retroactive to all trials that took place on or after that date.
As the first step in our analysis, we consider whether
Gould,
in holding that the jury may consider a defendant’s mental illness in determining whether a murder was committed with extreme atrocity or cruelty, announced a new rule. Prior to
Gould,
the court long had held that neither specific intent nor a knowledge of the nature of one’s acts was necessary for conviction of murder with extreme atrocity or cruelty. In 1858, the Legislature classified the three types of murder in the first degree in language substantially the same as that now appearing in G. L. c. 265, § 1. See St. 1858, c. 154, § 1. Two years later, this court held that proof of murder committed “with extreme atrocity or cruel
*545
ty” did not require proof of premeditation.
Commonwealth
v.
Desmarteau,
In
Commonwealth
v.
Gilbert,
The court’s conclusions in
Gilbert
were echoed by decisions in the decade preceding
Gould.
In
Commonwealth
v.
Appleby,
In addition to discussing the nature of the crime,'the court in decisions rendered prior to
Gould
detailed certain factors the jury were permitted to consider in determining whether a murder was committed with extreme atrocity or cruelty. These factors reflected the court’s view that the defendant’s guilt was to be determined by objective criteria, not focusing on the defendant’s mental state. The jury were permitted to consider the consciousness and degree of suffering of the victim, the disproportion between the means actually needed to inflict death and those employed, the instrumentalities employed, and the extent of the physical injury or indignities inflicted. See
Commonwealth
v.
Gould,
In Gould, the court concluded that the jury should no longer be “restricted to considering only the defendant’s course of action, but [should be] permitted to consider the defendant’s peculiar mental state as an additional factor to be weighed in determining whether the murder was committed with extreme atrocity or cruelty.” Commonwealth v. Gould, supra at 685-686. This ruling was a “clear break with the past,” representing a major change in Massachusetts law. Accordingly, we consider whether Gould should apply only prospectively.
We look to decisions of the United States Supreme Court for guidance in determining whether a new rule affecting
*547
the rights of criminal defendants should be applied only prospectively. We begin by noting that the Supreme Court has not distinguished between constitutional and non-constitutional decisions in addressing the question of retro-activity. In
Linkletter
v.
Walker,
In
Halliday
v.
United States,
To decide whether
Gould
applies retroactively to trials conducted prior to the date of that decision, we employ a test which “contemplates the consideration of three criteria: ‘(a) the purpose to be served by the new standards, (b) the extent of the reliance by law enforcement authorities on the old standards, and (c) the effect on the administration of justice of a retroactive application of the new standards.’
Stovall
v.
Denno,
*549
The threshold question in this case is whether the major purpose of the rule announced in
Gould
was to overcome an aspect of the criminal trial that substantially impaired its truth-finding function and so raises serious questions about the accuracy of guilty verdicts in past trials. In making this inquiry, we bear in mind that the extent to which a practice infects the integrity of the truth-determining process of trial “is necessarily a matter of degree.”
Johnson
v.
New Jersey,
In
Reddick
v.
Commonwealth,
Reliance on the old rule by law enforcement authorities and the impact retroactivity would have on the administration of justice also support the application of Gould prospec *550 lively only. 1 Since the rule established by Gould was a “clear break with the past,” reliance by the Commonwealth on pre-Gould standards was justified. In addition, although we cannot predict how many defendants convicted before Gould of murder in the first degree on the basis of extreme atrocity or cruelty would seek to attack their convictions on the ground that the trial judge erred in failing to instruct the jury in accordance with Gould, we are confident that the number of such attacks and resulting appeals and retrials would be enough to burden substantially the administration of criminal justice in the Commonwealth. We conclude that Gould is not retroactive. Consequently, we also conclude that Perry is not retroactive beyond the date of Gould.
We summarize as follows. The question before the court is whether the judge at Breese’s trial committed error by instructing the jury that they should not consider Breese’s intoxication in determining whether Breese murdered with extreme atrocity or cruelty. The answer depends on whether Perry is retroactive to Breese’s trial, that is, whether the law as stated in Perry was the law at the time of Breese’s trial. Since Perry did not announce a new rule, but was clearly foreshadowed by Gould, our holding in Perry applies retroactively to trials conducted after Gould. Perry is not retroactive to trials conducted prior to Gould, however, since Gould itself is not retroactive. Breese was tried be *551 fore the Gould decision. Consequently, the rule expressed in Perry was not the law at his trial, and there was no error in the judge’s instructions on extreme atrocity or cruelty.
Order denying motion for a new trial affirmed.
Notes
Compare
Diaz
v.
Eli Lilly & Co.,
