I respectfully dissent.
This case involves the calculated and systematic mass robbery of helpless migrant farm workers by professional bandits. The many victims were attacked in their isolated rural bunk house in the middle of the night, terrorized, robbed at gunpoint of their meager possessions. Martinez had predicted to his coconspirators that few if any of the migrant workers would remain to complain or testify. He was almost correct. Many of these victims—as Martinez knew—were undocumented aliens. Their fear of contact with authorities was an article of faith. Rather than come forward as a victim-witness and risk deportation, they disappeared. Such characteristics made these poor human beings, particularly vulnerable, easy prey for a conscienceless brigand. Before committing these robberies, Martinez told three members of his gang it would be “easy” to rob alien farm workers. He had done so before. The victims would not report the crimes. He was almost correct. Only one brave victim, Olea, remained to testify.
The trial court dismissed eight counts against Martinez because the victims could not be located. The majority opinion now reverses Martinez’ conviction on the remaining two counts (and by this decision will free him). The basic reason assigned? There is not enough evidence as a matter of law to convict him.
Several legally and factually flawed predicates are assumed by the majority in order to reach this patently unjust result.
Discussion
I
First, the majority err in holding that, as a matter of law, Heath was Martinez’ accomplice. Whether a person is an accomplice is a question of fact for the jury unless there is no dispute as to either the facts or the inferences to be drawn therefrom. (People v. Tewksbury (1976) 15 *138Cal.3d 953, 960 [127 Cal.Rptr. 135, 544 P.2d 1335].) Here, there is a dispute as to the inferences to be drawn from Heath’s testimony. As the following discussion makes clear, a reasonable trier of fact could have found (had the issue been presented to it) that Heath was not an accomplice as to the two robberies of which Martinez was convicted.1 Therefore, the majority err in holding he was an accomplice as a matter of law.
The fact that Heath was prosecuted for the same offense as Martinez does not by itself establish him to be an accomplice as a matter of law. (People v. Gordon (1973) 10 Cal.3d 460, 467 [110 Cal.Rptr. 906, 516 P.2d 298].) Nor is Heath necessarily an accomplice merely because he was held to answer for the same crimes as Martinez and then granted immunity.
Nor did Heath necessarily become an accomplice if it be deemed that he aided in the commission of the crimes. “Criminal liability as a principal attaches to those who aid in the commission of a crime only if they also share in the criminal intent ... or, in the language of section 31, abet the crime.” (People v. Tewksbury, supra, 15 Cal.3d at p. 960.) Heath was thus an accomplice only if at the time he acted he had guilty knowledge and intent with regard to the commission of the crime. (People v. Duncan (1960) 53 Cal.2d 803, 816 [3 Cal.Rptr. 351, 350 P.2d 103].) The burden of proof of the issue is on the defendant. (People v. Tewksbury, supra, at p. 968; People v. Belton (1979) 23 Cal.3d 516, 523 [153 Cal.Rptr. 195, 591 P.2d 485].)
The undisputed facts disclose Heath was Martinez’ accomplice as to the first two of the series of robberies—those on which Martinez was acquitted—the charges were dismissed for lack of witnesses. However, on the third night, Heath’s attitude, intent changed. After two nights of robberies, Heath “just wanted to leave the area.” He no longer wanted to commit any robberies, but the group wanted to commit one more. With this state of mind, Heath drove the trio to the area where they had committed robberies the first night. Heath was aware of the crimi*139nal intent of Martinez and his cohorts to return again to a ranch near Vista and rob helpless aliens. As in the earlier robberies, his function in that illegal scheme was to be “wheelman,” the driver of the getaway car. However, as Heath awaited to perform that duty, fate intervened in the form of Officer Burke, who stopped and questioned him. Heath then left the area and returned to Los. Angeles. He did not wait to pick up the robbing trio; he did not aid in the asportation of the loot. He did in fact abandon his agreed upon unlawful duties as getaway driver. He left the three robbers to return to the motel or to Los Angeles as best they could.2 Thus there is uncontroverted evidence of Heath’s in fact physical withdrawal from actual participation in the crime. The effect of Heath’s withdrawal from participation was to impede, by denial of the means of escape, the fulfillment of the crime of robbery. From these facts, the trial judge could, if he believed Heath’s testimony, reasonably conclude Heath had withdrawn from participation in the crime and therefore did not have the requisite intent to abet a robbery by driving the getaway car.
It was said in an early California case dealing with the law of withdrawal from criminal activity and avoidance thereby of criminal responsibility “[Responsibility of an accessory ‘does not cease simply, because after starting the ball, he changes his mind, and tries, when too late, to stop it. To emancipate him from the consequences, not only must he have acted in time, and done everything practicable to prevent the consummation, but the consummation, if it takes place, must be imputable to some independent cause.’ [Citation.]” (People v. Ortiz (1923) 63 Cal.App. 622, 670 [219 P. 1024]; see People v. King (1938) 30 Cal.App.2d 185, 204 [85 P.2d 928]; People v. Jones (1961) 197 Cal.App.2d 503, 508 [17 Cal.Rptr. 252].) However, the California Supreme Court in People v. Zamora (1980) 18 Cal.3d 538, 558 [167 Cal.Rptr. 573, 615 P.2d 1361], criticized this rule, saying the elimination of the withdrawal defense is “certainly not calculated to encourage the abandonment of criminal associations or goals.”
If we accept without question the Ortiz rule (and give no heed to the People v. Zamora criticism of the rule), then there is substantial evidence to warrant the trier of fact finding that Heath, within time, acted *140in such fashion as to indicate an intent to withdraw. He did “everything practicable to prevent the consummation,” i.e., the escape, of the robbers from the crime scene with the loot. Heath agreed (conspired) to take part in a criminal group activity; he was not to participate in the act of robbing but rather his function was to aid and abet by driving the getaway car. However, Heath did not in fact carry out his agreed upon duty. His departure for Los Angeles immediately hindered the robbers in the fulfillment of the criminal act—escape from the scene. Heath could have done no more to prevent the successful culmination of the robbery other than to confess the real purpose for his presence to the questioning officer. This latter act is not required under People v. Zamora, supra, 18 Cal.3d at pages 557-559. Heath’s departure, his absence when needed most, was a most forceful communication—by action rather than words—to Martinez and his band of robbers of an intent to withdraw from participation in the crime.
Sound public policy requires the encouragement of withdrawal from participation in crime at any stage of completion of the misdeed. Such policy should apply with greater force where the crime contemplated— as here—is yet in the conspiracy stage, where the substantive crime— the object of the conspiracy—has not yet commenced its evil course. As a matter of sound policy, Heath’s acts with his declared intent—if believed—should be sufficient to establish the defense of withdrawal from participation in the crime of robbery. The law encourages the abandonment of criminal objectives at an early stage by cutting off liability at the point of withdrawal. In contrast, the majority’s “in for a penny, in for a pound” approach only encourages the continuation to consummation of criminal activity once started. Here, the trial court, had the question been presented to it, could reasonably have, on the uncontradicted facts, found Heath was not an accomplice because he withdrew from participation before commencement or completion of the projected crime by use of the most graphic and most effective means possible to notify Martinez of his intent to withdraw. On the other hand, the court could have found Heath was an accomplice (if it disbelieved Heath’s testimony as to his acts and intent), concluding he did not intend to withdraw from participating in the crime. The critical point is: the resolution of this fact question was for the trial court, not this court. (People v. Tewksbury, supra, 15 Cal.3d 953.) The majority’s first error is in deciding Heath was an accomplice as a matter of law and refusal to imply findings in support of the judgment.
*141II
If we pretend there was no evidence to support Heath’s acts and claim of an intent to withdraw from participation, no evidence that he in fact returned tp Los Angeles, no evidence that he deliberately refused to act as the driver of a getaway car and assume, arguendo, that Heath was Martinez’ accomplice, yet Heath’s testimony was adequately corroborated as required by law. The majority wrongly conclude it was not, however, by the process of restricting their examination of the record to the evidence presented by the prosecution in its case-in-chief. This is a fallacious conclusion because it rests upon an unsound legal foundation. It is grounded upon the majority’s conclusion that Martinez effectively moved for acquittal under Penal Code section 1118 at the close of the People’s case, thus imposing on the trial judge the burden of ferreting out and deciding the accomplice and corroboration issues at that point in the trial. This factual assumption is wrong, belied by the cold record. Martinez’ motion made at the conclusion of the People’s case-in-chief was not brought under section 1118. Consequently, the question of the sufficiency of the evidence to corroborate Heath’s testimony was not presented to the trial court at the conclusion of the People’s case-in-chief. This court is therefore required to examine the entire trial record in determining whether Heath (assuming he was an accomplice) was corroborated. When this is done, it becomes abundantly clear that Heath’s testimony was sufficiently corroborated.
As the majority opinion openly confesses, Martinez’ motion did not expressly ask for acquittal under section 1118 for prosecutorial failure to present a prima facie case. Nor did it do so by implication. Neither the form nor the substance of a section 1118 motion can be intuited from the words used or the court’s response. Rather, Martinez moved to dismiss “on the grounds there is no direct evidence that he committed any robberies at all. At best there is circumstantial evidence that he committed such robberies . . .. ” (Italics added.)
When purified of all sophistic content, the majority come to this remarkable conclusion: If a criminal defendant, at the close of the prosecution’s case, moves to dismiss on the ground that the only evidence of his guilt is circumstantial (a ground that as a matter of law is unmeritorious on its face and concedes lack of merit to a § 1118 motion if made), the trial court must intuit or divine that the defendant moved instead for acquittal under Penal Code section 1118 (for prosecutorial failure to prove a prima facie case) and must scour the record for de*142fects in the prosecution’s case which the movant’s very words have first assured the court are not present.
Here, the majority reverse the conviction because the trial court did not do so (without request to do so) and sua sponte determine (1) Heath was an accomplice and (2) the corroboration of his testimony was insufficient. The majority hold this to be trial court error even though the defendant (whether tactically or inadvertently we do not know) diverted the court’s attention from the precise questions that would be raised by a Penal Code section 1118 motion by conceding the validity of Heath’s testimony and not raising the question of Heath’s status as an accomplice. Mattinez’ concession that “[a]t best there is circumstantial evidence” and circumstantial evidence pointed to Martinez as one of the robbers invited the trial court not to search the record as it should have, had it been informed that a Penal Code section 1118 motion was being made.
The majority decision invites gamesplaying at least, a cynical sandbagging of the trial judge at most. Error if any was clearly invited. This fact further critically distinguishes this case from the one on which the majority rely, People v. Belton (1979) 23 Cal.3d 516 [153 Cal.Rptr. 195, 591 P.2d 485]. Belton held a defendant who moves for acquittal under Penal Code section 1118 need not specify his grounds. The Supreme Court did not by any stretch of imagination sanction a kind of gamesmanship/sandbagging the majority opinion would invite.
The majority’s approach imposes an unfair and unrealistic burden on trial judges. The law requires the trial judge to determine an accomplice’s status and the sufficiency of the corroboration of his testimony when made aware by some intelligible signal, a proper section 1118 motion is brought. No statute or decision has as yet been found that requires him to do so when no section 1118 motion is brought at all and his attention is, in fact, affirmatively diverted from a Penal Code section 1118 quest by the defendant’s representations. Trial judges are not required to assume the role of defense counsel and to make all appropriate motions on behalf of a defendant. The majority here comes dangerously close to imposing a Pope-type duty on the trial court to conduct an effective defense for the criminal defendant. The majority has created a new, an unwise and unneeded procedural rule to reach a socially abhorrent result.3
*143III
The record here compels this conclusion: The corroboration issue was not placed before the trial court at the close of the prosecution’s casein-chief. Therefore, the question of the sufficiency of the corroboration of Heath’s testimony must be determined by examining the entire trial record. Here, in addition to the corroborative testimony of the victim and the police (which it is argued is insufficient under the law as it now stands—but see discussion infra) the rational inferences from Martinez’ own trial testimony independently connect him with the crime and corroborate Heath’s testimony.
A defendant’s own testimony, statements and admissions may afford corroboration, proof, sufficient to sustain a verdict. (People v. Wade (1959) 53 Cal.2d 322, 329 [1 Cal.Rptr. 683, 348 P.2d 116]; 7 Wigmore on Evidence (Chadbourn rev. ed. 1978) § 2059, p. 421; People v. Rissman (1956) 154 Cal.App.2d 265, 277, 278 [316 P.2d 60].) In his direct testimony Martinez was able to recall that in the late evening hours of December 13, 1978, he was in Los Angeles with a girlfriend (who could not be located). Yet when asked if he was in the company of Mr. Heath on the 11th, 12th or 13th of December 1978, he replied “not to my recollection.” Did he rob anyone near Vista in the early morning hours of December 14, 1978? He answered “not to my recollection.” Did he rob anyone on the evening of December 13, 1978? He again replied “not to my recollection.” Martinez’ remarkable lack of memory followed closely upon a most detailed and memory-refreshing account by Heath and Olea of the not-too-forgettable events of December 11 through 14, 1978.
3A Wigmore on Evidence (Chadbourn rev. ed. 1970) section 1042, page 1056 sets forth this premise: “A failure to assert a fact, when it would have been natural to assert it, amounts in effect to an as*144sertion of the non-existence of the fact. This is conceded as a general principle of evidence (§ 1071 infra). There may be explanations, indicating that the person had in truth no belief of that tenor; but the conduct is ‘prima facie’ an inconsistency.” And Wigmore further observes an “unwilling witness often takes refuge in a failure to remember .. ..” (Id., p. 1061; fn. omitted.) Thus Martinez’ repeated refrain non mi ricordo in contexts where a direct affirmative or negative was patently called for warrants an inference not favorable to his assertion of innocence.
The rule is well established that “[w]here a statement is made within the presence and hearing of an accused, which is incriminatory or accusatory in character, and such statement is not denied by him, the statement and the fact of his failure to deny are admissible as an admission of the statement’s truth.” (3 Wharton, Criminal Evidence (1973) § 700, pp. 501-502; fn. omitted.) And “[a]n evasive or unresponsive reply to an incriminatory or accusatory statement is tantamount to silence, in which case the statement and the reply are admissible as an admission of the statement’s truth.” (Wharton, supra, p. 505; fn. omitted.) If such conduct—evasive, nonresponsive answers—given in an extrajudicial setting may under certain circumstances constitute an admission of the accusatory statement’s truth then certainly that same conclusion follows where the non mi recordo responses are given in. court.
Under California law, inferences from the testimony of the defendant may be considered in corroboration of an accomplice’s testimony (People v. Allen (1951) 104 Cal.App.2d 402, 411 [231 P.2d 896]; People v. Malone (1947) 82 Cal.App.2d 54, 63 [185 P.2d 870]) and the making of false or evasive or contradictory statements may constitute an admission. (People v. Wayne (1953) 41 Cal.2d 814, 822-823 [264 P.2d 547]; People v. Wright (1949) 94 Cal.App.2d 70, 80 [210 P.2d 263]; People v. Simmons (1946) 28 Cal.2d 699, 712 [172 P.2d 18].) The trial court could and did see Martinez’ evasive, nonresponsive answers. The trial court was clearly entitled to draw as inference of deception, falsity, and admission of the truth of the accusatory facts implicit in Martinez’ repeated non mi ricordo responses. Martinez’ own evasive, nonresponsive answers thus supplied the independent corroboration of Heath’s testimony required by Penal Code section 1111. (Cf. People v. Sullivan (1904) 144 Cal. 471, 473 [77 P. 1000].)
*145The fundamental rule is: Corroborative evidence, direct or circumstantial, is sufficient if it tends to connect the defendant with the crime, even though it is slight and entitled, if standing alone, to little weight. (People v. Wade (1959) 53 Cal.2d 322 [1 Cal.Rptr. 683, 348 P.2d 116]; People v. Negra (1929) 208 Cal. 64, 69, 70 [280 P. 354].) Martinez’ guilt-admitting responses on the witness stand were more than sufficient to corroborate Heath—if corroboration was necessary.
IV
In addition to the inferences derived from Martinez’ evasive nonresponses to questioning, a mass of evidence affirms the truthfulness of Heath’s testimony. At least 10 important facts of the crimes testified to by the victim Olea and Officer Burke mirrored in exact detail those provided by Heath. Heath’s testimony as to the number of the robbers, their race, their sex, the time of the robbery, the place of the robbery, the robbers’ use of two handguns, one larger than the other, the robbers’ use of a pillowcase stolen from a hotel, the robbers’ use of a single flashlight, and the green Kawasaki motorcycle was confirmed with exactitude by the testimony of Olea and Burke.
The large number of points of agreement between Heath and the nonaccomplice witnesses, and the complete absence of any contradictory points, renders virtually inescapable the conclusion that Heath is a believable witness and was telling the truth about Martinez’ involvement in the robberies. Nevertheless, as the law presently stands, does this evidence by itself sufficiently corroborate Heath’s testimony so as to support the verdict? Does it independently connect Martinez to the robberies? ‘“To corroborate the testimony of an accomplice, the prosecution must produce independent evidence which, without aid or assistance from the testimony of the accomplice, tends to connect the defendant with the crime charged.’” (People v. Szeto (1981) 29 Cal.3d 20, 27 [171 Cal.Rptr. 652, 623 P.2d 213].) Martinez argues the testimony of Olea and Burke, despite its obvious probative force in making Heath’s testimony absolutely believable, connects Martinez with the robberies only when read in conjunction with Heath’s testimony. Martinez argues it is therefore to be deemed valueless.
Counsel for Martinez conceded in open court “circumstantial” evidence connected Martinez with the Olea robberies. The majority opinion bursts with the admission there is no doubt that Martinez is *146guilty as charged. Why such concessions from the majority as well as defense counsel?
First, the testimony of Heath has such realism, such color, such detail as would require the word skills, the imagination of a Shakespeare or a Faulkner to fabricate such a story. Secondly Heath’s story has at least ten points of absolute congruence with the testimony of two totally believable witnesses.
A single thread from the white pillow case found on Martinez or a key in his pocket that would fit the door of the bunkhouse, or a shoe print at the robbery site of the same size, make as Martinez’ would meet the ritualistic requirement of tending to connect the defendant with the crime. We have no such talismanic bits of evidence here. However, the 10 points of agreement are analogous to the requirement of a minimum number of points of similarity required before a “scientific” opinion can be given as to the identity of a depositor of a questioned fingerprint. The same unanswerable reasoning (an astronomical degree of mathematic improbability of error is present where there are 10 or more points of congruence found on a questioned fingerprint) assures us that Heath is telling the truth. Martinez was one of the three robbers identified by Olea.
The code requires “such other evidence as shall tend to connect the defendant with the commission of the offense ....” As to the matter of that evidence requirement, the Supreme Court in People v. Henderson (1949) 34 Cal.2d 340, 342-343 [209 P.2d 785], said: “[Corroborating evidence ... is sufficient if it connects] the defendant with the commission of the crime in such a way as reasonably to satisfy the fact finding body that the accomplice is telling the truth.” For the same rule see People v. Medina (1974) 41 Cal.App.3d 438, 460 [116 Cal.Rptr. 133] (hg. den. by Supreme Ct.); People v. Randono (1973) 32 Cal.App.3d 164, 173 [108 Cal.Rptr. 326] (hg. den. by Supreme Ct.).
In Henderson, supra, the Supreme Court in discussing the mode of proof, the nature of the evidence that would constitute corroborative evidence said: “The testimony of Roberts that he held the black pistol and that the defendant held the shotgun at the time of its discharge has some corroboration in the testimony of the victims in the cafe who stated that one of the culprits had a shotgun and the other a black pistol.
*147“The testimony of the victims in the cafe that both men wore flesh-colored stockings as hoods with slits cut in them for eye holes shows circumstances in substantiation of the testimony of Roberts that he and the defendant wore hoods so described by him at the time of the commission of the crime.
“We are satisfied that the cumulative effect of the testimony of the women companions of the two men, the testimony of Roberts’ sister and the testimony of the victims in the cafe, none of whom could be deemed accomplices, point sufficiently to the establishment of the fact that the defendant was the other participant in the crime, and that the code requirements of evidence tending to connect the defendant with the commission of the crime have been sufficiently met.” (Id., at p. 346.) (See also People v. Sullivan, supra, 144 Cal. at p. 472, where similar “mirror image” testimony was held corroborative.) Thus the evidence received from Olea and Officer Burke was of the exact same type as approved by the Supreme Court. This specie of evidence patently fulfills the primary reason for the requirement of corroboration—distrust of accomplice testimony. (People v. Robinson (1954) 43 Cal.2d 132, 141 [271 P.2d 865].) Olea and Burke’s evidence made Heath’s believable. If corroboration was necessary, it was present in the required quantum in the People’s case-in-chief.
V
At common law, the uncorroborated testimony of an accomplice could sustain a conviction. (See 7 Wigmore (Chadbourn ed. 1978) Evidence, § 2056.) A practice gradually developed, however, of having the trial judge comment to the jury upon the weight and credibility of the accomplice’s testimony and the desirability of corroboration, and in roughly half of the jurisdictions in the United States this customary practice has become a fixed rule of law by statute.
A debate arose early as to the nature of the corroborative evidence required. Some English jurists took the position corroboration of the facts of the crime which did not independently implicate the defendant would not do. As stated by Lord Abinger, C.B., in R. v. Fawler (1837) 8 Car. & P. 106, 108: “A man who has been guilty of a crime himself will always be able to relate the facts of the case, and if the confirmation be only on the truth of that history, without identifying the person, that is really no corroboration at all.” Chief Baron Joy, however, was of the opposite view: “The correct and accurate manner in which an ac*148complice details the circumstances of the transaction shows that he was cool and collected, that he possessed observation, that his recollection is fresh, that he was an observer, not an inventor of facts and incidents; and if we find that in every point in which the evidence of other witnesses can be brought into contact with his, they fit into one another and correspond exactly, it is good ground for presuming that his entire narrative is correct. ... The accomplice, who must be supposed to know the whole details, is expected to relate them, and is thus exposed to detection in a variety of ways.” (Joy, Evidence of Accomplices (1844) p. 10; italics added.)
In California, the debate was resolved by statute. Penal Code section 1111, as originally enacted, provided: “A conviction cannot be had on the testimony of an accomplice, unless he is corroborated by other evidence which in itself, and without the aid of the testimony of the accomplice, tends to connect the defendant with the commission of the offense; and the corroboration is not sufficient, if it merely shows the commission of the offense, or the circumstances thereof.” Accordingly, the early California cases, without exception, and in conformity with this express legislative provision, held that corroborating evidence, to be sufficient, must tend to connect the defendant to the crime without the aid of the testimony of the accomplice. (See, e.g., People v. Morton (1903) 139 Cal. 719, 727 [73 P. 609].)
In 1911, however, the Legislature amended the statute to delete the phrase “which in itself, and without the aid of the testimony of the accomplice.” The section now reads: “A conviction can not be had upon the testimony of an accomplice unless it be corroborated by such other evidence as shall tend to connect the defendant with the commission of the offense; and the corroboration is not sufficient if it merely shows the commission of the offense or the circumstances thereof.”
In 1915, the Supreme Court had occasion to rule on the meaning of the section as amended. In a decision which can fairly be described as surprising, the court held (but not without dissent) that the amendment did not change the meaning of the section “at all.” (People v. Robbins (1915) 171 Cal. 466, 473-474 [154 P. 317].) The court cited none of the familiar canons of statutory construction, nor did it cite any evidence of legislative intent.4
*149The Robbins rule—the inclusion of a mode of proof excluded by the Legislature—has been questioned. “In view of the more recent decisions of the Supreme Court we conclude the test suggested in People v. Morton, supra, [139 Cal. 719] and followed by People v. Reingold, supra, (1948) 87 Cal.App.2d 382 (197 P.2d 175)] is not an exclusive method of analyzing the sufficiency of evidence corroborative of the testimony of an accomplice.” (People v. Griffin (1950) 98 Cal.App.2d 1, 28 [219 P.2d 519]; People v. Allen (1951) 104 Cal.App.2d 402, 413 [231 P.2d 896], and cases cited.)
The rule requiring corroboration of an accomplice’s testimony is somewhat an anachronism in modern evidence law. As a general rule, a criminal conviction can rest on the uncorroborated testimony of witnesses whose veracity is suspect. Those who have an interest in the case need not be corroborated; nor do those who have been convicted of crime. Yet an accomplice must be. The rule requiring corroboration has been widely criticized (see 7 Wigmore, Evidence (Chadbourn rev. ed. 1978) § 2057) and roughly half the jurisdictions in the country have never adopted it. Moreover, some of those states that do have the rule have sought to repeal it. (See, e.g., N.Y. Com. on the Admin, of J., 3d Supplemental Rep., p. 16 (Leg. Dec. 1937, No. 77) quoted in Wigmore, supra, at p. 420. )5
*150The reason for the rule requiring corroboration of an accomplice’s testimony is said to be this: An accomplice’s veracity is suspect because he may have secured a promise of leniency in exchange for his testimony. (Wigmore, supra, at p. 417.) The rule is not based on the notion that one who was involved in a criminal enterprise is, by virtue of that fact, inherently untrustworthy. As Wigmore notes: “We have passed beyond the stage of thought in which [an accomplice’s] commission of crime, self-confessed, is deemed to render him radically a liar.... The extreme case of the wretch who fabricates merely for the malicious desire to drag others down in his own ruin can be no foundation for a general rule.” (Ibid.) The problem with the rule, of course, is that even the promise of immunity is not always present. If that element is lacking in a particular case, so is the element of distrust. And even where present, its influence will be highly variable. This has led Wigmore and others to conclude that the rule requiring corroboration of accomplice testimony in all cases is unsound.
The section’s legislative purpose, if it is to assure an accomplice’s veracity, would be amply served by allowing the requisite corroborative evidence to refer to the accomplice’s testimony in order to be meaningful. The fact that the accomplice has told the truth about some aspects of the crime is some evidence that he can be believed as to other aspects. And this is especially true where, as here, the accomplice’s testimony is highly detailed and verified in every respect. Olea’s and Officer Burke’s detailed verification of Heath’s testimony should be more than enough to satisfy the legislative purposes expressed in Penal Code section 1111.
The corroborative evidence admitted here does independently connect the defendant to the crime scene, it independently establishes the truthfulness of the accomplice “testimony.” (People v. Henderson, supra, 34 Cal.2d 340.)
Judicial recognition of the fact that accomplices are really no less reliable than convicted felons or interested witnesses is already evidenced by the rule that the required corroborative evidence, if in the correct form, will satisfy the statute even when “slight, and entitled to little consideration when standing alone.” (People v. Wade (1959) 53 Cal.2d 322, 329 [1 Cal.Rptr. 683, 348 P.2d 116].)
The majority remand case No. CR 48564 to the superior court for a redetermination of Martinez’ entitlement to presentence credit. This *151task should be delegated to the Department of Corrections and not be imposed on an already overburdened trial judge. (See People v. Sage (1980) 26 Cal.3d 498, 507, fn. 7 [165 Cal.Rptr. 280, 611 P.2d 874].)
I would affirm the trial court judgment.
Respondent’s petition in No. 11727 for a hearing by the Supreme Court was denied July 28, 1982. Richardson, J., was of the opinion that the petition should be granted.
We are required to imply findings of fact to support the judgment and the conclusion that Heath was not an accomplice and needed no corroboration; or if an accomplice, he was corroborated (In re Pratt (1980) 112 Cal.App.3d 795, 929 [170 Cal.Rptr. 80]; People v. Rosoto (1962) 58 Cal.2d 304, 329 [23 Cal.Rptr. 779, 373 P.2d 867]; People v. Henderson (1949) 34 Cal.2d 340, 343 [209 P.2d 785].) Substantial evidence supports either of these alternatives.
Heath met Martinez in Los Angeles the next day. Martinez was “upset” and threatened “to beat ... [Heath’s], head” because of Heath’s abandonment of the robbers, leaving them without any means of getaway excepting Martinez’ Kawasaki motorcycle which was some distance down the road.
The majority concede that their extension of Belton results in “injustice” in this case and gives rise to a “difficult problem.” Indeed it does. On a fundamental level, the *143problem with the majority’s unwarranted extension of Belton is that it injects yet more gameplaying in a criminal trial. The defendant is insulated from the consequences of ineffective strategy because if he errs, the trial judge must step in and conduct his defense for him. Ours is an adversary system; a criminal trial is not a game, with victory going to the most skillful player. It is a search for truth with the proviso that this interest does not outweigh the need to protect certain of our fundamental preferred values (e.g. due process, freedom from unreasonable searches and seizures, the privilege against self-incrimination, etc.) Extending Belton to cases such as this rewards game-playing by the defendant at the expense of the integrity of the factfinding process. As applied here, this approach operates not to safeguard the innocent, but to free the guilty, and it does so without advancing any fundamental interests.
It is hornbook law that the Legislature is presumed to know the decisions of appellate courts and to have them in mind when amending statutes which the courts have *149construed. It is equally well settled that a substantial change or deletion in the language of a statute is presumed to change its meaning. (See, e.g., Subsequent Injuries Fund v. Industrial Acc. Comm. (1963) 59 Cal.2d 842, 844 [31 Cal.Rptr. 477, 382 P.2d 597]; Clements v. T. R. Bechtel Co. (1954) 43 Cal.2d 227, 231 [273 P.2d 5]; Hoffman v. McNamara (1929) 102 Cal.App. 280, 285 [282 P. 990]; see generally 1A Sutherland (4th ed. 1972) Statutory Construction, § 22.30.)
As Chief Baron Joy notes: “Why the case of an accomplice should require a particular rule for itself; why it should not, like that of every other witness of whose credit there is an impeachment, be left to the unfettered discretion of the judge, to deal with it as the circumstances of each particular case may require, it seems difficult to explain. Why a fixed, unvarying rule should be applied to a subject which admits of such endless variety as the credit of witnesses, seems hardly reconcilable to the principles of reason. But, that a judge should come prepared to reject altogether the testimony of a competent witness as unworthy of credit, before he had ever seen that witness; before he had observed his look, his manner, his demeanour; before he had had an opportunity of considering the consistency and probability of his story; before he had known the nature of the crime of which he was to accuse himself, or the temptation which led to it, or the contrition with which it was followed;—that a judge, I say, should come prepared beforehand, to advise the jury to reject without consideration such evidence, even though judge and jury should be perfectly convinced of its truth, seems to be a violation of the principles of common sense, the dictates of morality, and the sanctity of a juror’s oath....” (Joy, Evidence of Accomplices (1844) 4. For a particularly scholarly Connecticut case exploring the foundations of the rule and criticizing its retention in the 20th century see State v. Carey (1904) 76 Conn. 342, 346-349 [56 A. 632].)