Primarily on the basis of testimony by an undercover police officer, a jury found appellant guilty of unlawful distribution of cocaine (D.C.Code § 33-541(a)(l) (1988)). On appeal he contends that the trial judge erred in refusing to allow him to question the officer about the fact that, on this occasion and others, he earned overtime pay for testifying in court. Appellant further contends that the evidence that he distributed a usable amount of narcotics was insufficient as a matter of law. We reject both contentions and affirm.

I.

Bruce Faison, a member of the Narcotics Task Force of the Metropolitan Police, was operating undercover with his partner Officer Bush when they approached appellant in the Potomac Gardens area of Southeast Washington, D.C., at around 5:00 p.m. on August 31, 1990. Faison’s attention was drawn to appellant because the latter was standing near Building 714 in Potomac Gardens exclaiming “shake, shake, got that shake,” which Faison understood to be a street name for a powder form of cocaine. Faison asked appellant, “Who got that shake?”, and appellant replied, “Come on, Shorty, I will take you.” The two men walked toward the building where a third man, Willis Baum, joined them. Appellant told Baum what Faison wanted, and Baum asked, “How many?”, to which Faison replied, “Two dimes,” meaning two ten-dollar packets of white powder. As appellant stood beside the men, Baum gave Faison *903two plastic bags containing white powder from a bundle secured around his finger by a rubber band. Faison in turn gave Baum twenty dollars in pre-recorded police department money.

After the sale, Faison and Officer Bush, who had remained in a nearby doorway during the transaction, returned to their unmarked police vehicle, where Faison radioed descriptions of appellant and Baum to a waiting arrest team. He described appellant as a black male, tall, wearing a black baseball-style cap, a green short-sleeve shirt, blue jeans, and black tennis shoes.1 When appellant and Baum were stopped approximately four minutes later, Faison drove by and positively identified both men as the sellers. Identifying appellant in court as well, Faison was “very sure” of his recollection of the events, remembering the incident particularly because appellant had been advertising “shake” aloud.2 A search of Baum on the scene yielded the twenty dollars in pre-recorded funds used to buy the cocaine.

A forensic chemist from the Drug Enforcement Administration testified that the two plastic bags contained a total of 180 milligrams of powder, of which seventeen percent, or 30 milligrams, was cocaine. Detective Joseph Brenner testified that this was a usable amount of cocaine because it could be ingested into the body in the way powder cocaine normally is used, i.e., by snorting or injection.

The day before appellant’s trial began, codefendant Baum pleaded guilty to the charge of distributing cocaine on August 31, 1990. He took the stand on behalf of appellant and explained (in limited testimony) that he had had no arrangement with appellant to sell cocaine on the day in question, he had never met appellant before they were arrested, and appellant had come no closer to him than 40 to 50 feet before the arrest.3

II.

Appellant contends that the trial judge deprived him of his Sixth Amendment right to establish bias on the part of Officer Faison by precluding cross-examination about whether the officer received overtime pay for his testimony in court. When appellant’s counsel asked Faison, “[A]re you on overtime now?”, the government objected on grounds of relevancy, and a lengthy discussion ensued out of the presence of the jury. Ultimately the trial judge sustained the objection on the ground that the proposed questioning lacked probative value and could only distract and confuse the jury. We uphold the trial court’s ruling.

Appellant’s claim of bias was anything but straightforward. As he concedes on appeal, the financial incentive he sought to establish bore only indirectly on Officer Faison’s veracity at trial, for if — as defense counsel proffered — the officer was receiving time and a half pay for his appearance in court, he would earn that pay for all overtime work whether it involved testifying in court or not. Moreover, he would receive it whether or not his courtroom testimony incriminated appellant. What counsel sought to establish was that there was a self-interested, indeed a corrupt, link between the fact of appellant’s arrest and Faison’s testimony. He maintained that Faison was motivated from the beginning to arrest appellant and other persons who were either “marginally” involved in drug sales or not involved at all, in order to secure for himself time in court as a witness and overtime pay in consequence. Counsel proffered (and expressed confidence he could elicit from the government’s own police expert in the case) that police *904officers, particularly those from the Narcotics Task Force such as Faison, could earn “thousands of dollars a year” in overtime from their courtroom testimony. And this interest, counsel alleged, was linked directly to the potential innocence of defendants such as appellant, because among all persons arrested on drug charges, those most likely to put the government to its proof at trial were defendants believing in their innocence, whereas factually guilty defendants were more likely to enter guilty pleas to lesser drug charges before trial in order to avoid mandatory minimum sentencing.

This theory of bias, as the trial judge recognized, rested on a series of assumptions unsupported by any evidentiary proffer. The ultimate assumption, of course, was that an officer like Faison would jeopardize his career (and risk civil if not criminal liability) by falsely arresting innocent persons for the future gain of overtime pay. The underlying factual assumption was that most “guilty” drug defendants would plea bargain (hence offer little promise of remuneration for Faison) while defendants innocent in fact would commonly insist on their right to trial.4 Appellant proffered no evidence at all about the frequency of guilty pleas in drug cases, nor about the reasons (assuming these could be established empirically) why individual defendants go to trial — i.e., because they truly believe themselves innocent; or, innocence aside, because they are confident they can beat the government’s case, perhaps by a motion to suppress; or simply because they have not been offered a plea to a reduced charge because of recidivism or other reasons subsumed under prosecu-torial discretion. In particular, as the trial judge noted and defense counsel agreed, the government commonly “wired” plea offers in the case of jointly arrested defendants such as appellant and Baum, so much so that — in this case — until the very day before trial when Baum pled guilty to the charged offense, Officer Faison’s attendance at trial was not contingent on appellant’s decision to stand trial. Yet the theory of bias was that the officer had to cast his net broadly, arresting the guilty and innocent alike, to insure that at least one among codefendants would proceed to trial.

Aside from the lack of an evidentiary proffer, the flaw the trial judge perceived in appellant’s theory was that between the putative motive to arrest innocent or “marginal” defendants5 to earn extra courtroom pay and its ultimate realization lay numerous steps in the criminal justice process over which the officer had little or no control, and which in the aggregate made the imputed motive imaginary. Assuming, for example, that the officer could predict how prosecutors would exercise their charging and plea discretion in individual cases, there remained the grand jury as an obstacle before a putatively corrupt officer could have any confidence that innocently arrested defendants would be forced to stand trial.6 Without detailed explanation of how these factors operate in the process, a jury could not intelligently begin to evaluate the theory of bias. And even then, as the trial judge concluded, the result of the inquiry would be “so speculative and so interdependent upon other information that otherwise would have no place in the trial” that jury confusion would outweigh any possible utility on the issue of bias.

We find no reason to disturb this ruling. There is, of course, “no constitutional right to present irrelevant evidence.” Roundtree v. United States, 581 A.2d 315, 321 *905(D.C. 1990) (quoting Gibson v. United States, 536 A.2d 78, 82 (D.C. 1987)). “Although possible bias of a principal government witness is always a proper subject for cross-examination,” a proposed line of bias questioning must satisfy standards of relevancy: “[t]he party posing the question must proffer to the court some facts which support a genuine belief that the witness is biased in the manner asserted, that there is a specific personal bias on the part of the witness, and that the proposed questions are probative of bias.” Porter v. United States, 561 A.2d 994, 996 (D.C. 1989) (citations omitted). Having received no foundational proffer other than that narcotics officers generally earn substantial overtime for court attendance, the trial judge did not err in finding Officer Faison’s pay status to be “[injadequately probative of the fact it [was injtended to establish,” i.e., his bias, and thus “[insufficiently relevant to be admissible.” Reavis v. United States, 395 A.2d 75, 78-79 (D.C. 1978); Punch v. United States, 377 A.2d 1353, 1358 (D.C. 1977), cert. denied, 435 U.S. 955, 98 S.Ct. 1586, 55 L.Ed.2d 806 (1978).

Appellant points out that in a seemingly analogous setting we held (or at least implied) that a proffer of facts suggesting the existence of a police policy linking the number of arrests to an officer’s job advancement might require allowing some inquiry into that possible motive for bias. Van Ness v. United States, 568 A.2d 1079, 1082 (D.C. 1990); id. at 1083-89 (Schwelb, J., concurring in part and dissenting in part). But in Van Ness it was the existence of a “quota” of arrests without more that provided the asserted motive to arrest indiscriminately. And the bias theory was supported by a specific factual proffer. Id. at 1087. Appellant’s theory, in vivid contrast, depended on a web of assumptions about who among the broad class of arrestees for drug offenses would predictably go to trial and who would plead guilty — and it was supported only by a proffer that overtime for narcotics officers is commonplace. Both the speculativeness of these assumptions and the danger that the jury would ruminate on them despite any answers the officer might give to the questioning justified its exclusion. As was stated in Van Ness:

Once the question whether officers are promoted, in part, on the basis of the number of arrests they make has been posed, it may be difficult even for the most conscientious juror not to muse on the subject, no matter what the officer’s answer may be. Under these circumstances, trial judges should surely be reasonably exacting in requiring counsel to establish a good faith basis for pursuing such a line of inquiry.

Id. at 1087 (Schwelb, J., concurring in part and dissenting in part). While we need not hold that a sufficient predicate could never be laid for the proposed bias questioning here, appellant’s proffer fell far short of satisfying the threshold standard of relevance. Porter v. United States, supra.7

III.

Appellant also contends that the government failed to prove that the substance purchased by Officer Faison, although admittedly containing 30 milligrams of cocaine (one-sixth of the total substance), “could produce a narcotic [or pharmacological] effect upon the user” and so contained a usable amount of cocaine. The argument *906that the government must adduce testimony specifically about narcotic effect in a case such as this has been laid to rest by our recent decisions. In Judge v. United States, 599 A.2d 417 (D.C. 1991), we explained:

Appellant argues, nonetheless, that proof of usability requires proof that the amount was sufficient “to have a pharmacological effect on the user,” citing Singley v. United States, 533 A.2d 245, 247 (D.C. 1987). However, as we noted in Davis v. United States, [590 A.2d 1036 (D.C. 1991) ], distinguishing Singley, that case involved a chemical analysis which stated only that the material analyzed contained a “small amount” of heroin. The expert had testified that in his opinion a powder containing only “trace amounts” would not be usable, and that he did not know whether in a report of chemical analysis, “small amounts” differed from “trace amounts.” In Singley, we cited Edelin v. United States, 227 A.2d 395, 399 (D.C. 1967), which also involved traces and where we noted that there was “no additional proof” of its usability as a narcotic. Likewise, in Singley, we concluded that since “a trace amount is insufficient to convict whenever it cannot produce a narcotic effect in any form,” the bare record presented could not support a jury finding of usability. There is no warrant to read Sing-ley as imposing a requirement of proof of narcotic effect regardless of the quantity of the controlled substance and other proof of its usability. Davis v. United States, supra.

Id. at 420 (footnote omitted). In other decisions as well, we have narrowly construed Edelin’s requirement of proof of utility as a narcotic. Thus in Wishop v. United States, 531 A.2d 1005 (D.C. 1987), we stated: “Edelin holds only that if the quantity of a drug is too small to be capable of quantitative analysis, there must be ‘additional proof of its usability as a narcotic’ in order to sustain a conviction.” Id. at 1008 (emphasis added) (quoting Edelin, 227 A.2d at 399). Most recently we said that Edelin “goes no further than to recognize that the common sense application of the narcotics laws does not reach a seized amount ‘so inconsiderable as to make it of no utility to a user and unmarketable....’” Gray v. United States, 600 A.2d 367, 369 (D.C. 1991) (quoting Wishop, 531 A.2d at 1008). And in Wishop we expressly stated that “the fact that a drug is measurable — i.e., capable of quantitative analysis — will usually suffice to prove that it is usable.” 531 A.2d at 1008. See also Johnson v. United States, 611 A.2d 41, 43 (D.C. 1992); D.C.Law 8-138 (1990) (amending D.C.Code § 33-516’s placement of cocaine in list of controlled substances).

The essence of our decisions, therefore, is that the government’s proof of usable amount will not fail unless “there is only a trace of a substance, a chemical constituent not quantitatively determined because of minuteness, and there is no additional proof of its usability as a narcot-ic_” Edelin, 227 A.2d at 399 (emphasis added). Here the testimony plainly established that the substance recovered contained more than a trace amount of cocaine. As the DEA chemist explained, the thin-layer chromatography test demonstrated “a fingerprint^] ... a pattern of cocaine”; “[c]ocaine [was] heavily suggested.” Even as to the residue of cocaine (20 milligrams) remaining after the test, the chemist noted that “you can see it ... you can weigh it also.” In his opinion, “[n]o other substance [than cocaine] would have yielded [the] results” obtained from the testing. Detective Brenner, also called as an expert, testified that 30 milligrams of cocaine was a usable amount because it could be ingested into the body in the way powder cocaine is normally used, i.e., by snorting or injection. Finally, the evidence showed that the cocaine purchased by Fai-son was packaged in two small plastic bag containers similar to those in which cocaine is usually sold on the street. All told, these facts were easily sufficient to satisfy the usable amount requirement.

The judgment of the Superior Court is

Affirmed.

dissenting:

Because I conclude that the trial judge erred by denying, without voir dire after a *907defense proffer, any cross-examination of an undercover officer regarding his possible financial bias, I respectfully dissent.

The court has long made clear that:

[t]he permissible scope of cross-examination “must be limited with the utmost caution and solicitude for the defendant’s Sixth Amendment rights,” United States v. Houghton [, 554 F.2d 1219,] 1225 [ (1st Cir. 1977)], and “this discretionary authority to limit cross-examination comes into play after there has been permitted as a matter of right sufficient cross-examination to satisfy the Sixth Amendment.” United States v. Bass, 490 F.2d 846, 957-58 n. 12 (5th Cir. 1974). The trial court’s “wide latitude in the control of cross-examination ... ‘cannot ... justify a curtailment which keeps from the jury relevant and important facts bearing on the trustworthiness of crucial testimony.’ ” United States v. Harris, 501 F.2d 1, 8 (9th Cir. 1974), quoting Gordon v. United States, 344 U.S. 414, 423, 73 S.Ct. 369 [375], 97 L.Ed. 447 (1953).

Goldman v. United States, 473 A.2d 852, 856 (D.C. 1984) (quoting Springer v. United States, 388 A.2d 846, 855 (D.C. 1978)). My concern is that the trial judge was insufficiently solicitious of appellant’s right to show bias of the key government witness.1 The majority appears to suggest that a witness’ bias arising from financial interest can be shown only where there is no chance that a witness’ expectations of financial gain may be defeated. See majority opinion at 904. Of course, the prosecutor might decline to present the case and the grand jury might refuse to indict. But, in view of the “war on drugs,” it cannot be gainsaid that it would be unreasonable for a police officer to assume that a person caught red-handed with marked police money is more likely to plead guilty than a person who is merely in the area at the time of the drug sale. The officer could therefore conclude that, having purchased drugs from one person, an additional arrest of someone less likely to enter a guilty plea would enhance the likelihood that the officer will put in court time and, thereby, earn overtime pay. Furthermore, a defendant’s right to cross-examine a witness to show bias is not dependent on proof beyond a reasonable doubt of bias in fact.

It is the essence of a fair trial that reasonable latitude be given the cross-examiner, even though he [or she] is unable to state to the court what facts a reasonable cross-examination might develop.... To say that prejudice can be established only by showing that the cross-examination, if pursued, would necessarily have brought out facts tending to discredit the testimony in chief, is to deny a substantial right and withdraw one of the safeguards essential to a fair trial.

Best v. United States, 328 A.2d 378, 382 n. 3 (D.C. 1974) (quoting Alford v. United States, 282 U.S. 687, 692, 51 S.Ct. 218, 219, 75 L.Ed. 624 (1931)).

If a police officer were arresting persons associated in a noncriminal way with drug distributors for the purpose of making extra money through overtime court appearances, cross-examination to show financial bias would be proper since financial motives can skew testimony. See Hayward v. United States, 612 A.2d 224, 227 (D.C. *9081992) (“[b]ias relates to a witness’s ‘testimonial motivation,’ i.e., whether the witness has an interest in the parties or proceedings that may color the witness’ testimony” (quoting Coligan v. United States, 434 A.2d 483, 485 (D.C. 1981), and citing Davis v. Alaska, supra note 1, 415 U.S. at 316, 94 S.Ct. 1105 at 1110). Although such bias is not precisely the same as that of a witness who receives a fee for testifying for one party, a reasonable inference could be drawn, under some circumstances, that there was a financial motive inconsistent with the officer’s official duties behind an officer’s arrest of a defendant. Such an inference would call into question the truthfulness of the officer’s written reports and trial testimony.2 Cf. Van Ness v. United States, 568 A.2d 1079, 1082 (D.C. 1990) (where cross-examination designed to show that police officer in Operation Clean Sweep was supposed to make drug arrests on a quota, through deception, without revealing his identity or purpose, “the trial court might have permitted a limited further examination on the subject”); Mitchell v. United States, 408 A.2d 1213, 1215 (D.C. 1979) (cross-examination on bias especially important where key witness is informant, accomplice, or where witness has been given immunity from prosecution).3

Defense counsel’s theory was that Officer Faison had made an unnecessary arrest (of appellant) in the interest of receiving overtime pay for appearing in court. Once in court, defense counsel theorized, Officer Faison would fabricate testimony about appellant’s involvement in order to assure that the case would come to trial and that he would be paid as a witness. The trial judge was properly concerned that such an inference would have serious and sensitive implications. Where the officer’s regular tour of duty occurs when the trial court is not in session, the officer is entitled to receive overtime pay for court appearances. See D.C.Code §§ 4-1104(d)(l)(A) & (B), 4-1104(e) (1988 Repl.).4 Hence, it could be misleading to suggest any ulterior motive from the fact that the officer is paid overtime for appearing in court at trials. However, defense counsel made a detailed proffer that met the three-prong test of Parker v. United States, 586 A.2d 720, 723 (D.C. 1991), and the judge misconstrued the relevance of the existence of appellant’s co-defendant. Consequently, the judge erred by cutting off any cross-examination about the officer’s financial status before examining the officer on voir dire to determine whether the proffer could be sustained. Because of weaknesses in the government’s case, the error cannot be deemed harmless.

I

Defense counsel met his burden under Parker, supra, 586 A.2d at 723, when he made a proffer to the trial judge, and also proposed alternative means of proving Officer Faison’s bias.

During cross-examination, defense counsel asked Officer Faison, a member of the narcotics task force, if he was “on overtime right now.” The officer repeated the question, and defense counsel affirmed that that was the question he was asking. The *909prosecutor then objected on the grounds of relevance. Defense counsel advised the trial judge that “I want to establish that this officer has a financial interest in having cases go to trial as opposed to having them plead. In other words, that he is making extra money by being here as a witness in this case.”

Defense counsel thereupon proffered to the trial judge that he sought to show that the officer had a financial interest in the trial because officers who work nights or evenings in drug details are paid overtime for being in court during the day for the trials.5 Defense counsel explained that the government always asks the jury why a police officer would lie, and argues that “the police were just doing their job and have no reason to lie.”6 Defense counsel continued:

By bringing someone into the case whose involvement is marginal or nothing he [the officer] makes it more likely that he will be in the trial of the case as opposed to the plea in the case. That I can establish I am sure through this officer’s own testimony that he has a financial stake in being a witness in trials as oppose[d] to having cases disposed of prior to trial. [Emphasis added]

Defense counsel also offered to present expert testimony that undercover police officers, especially narcotics task force members, “make a significant amount of money each year, each month, each week,” some even “double their salaries” because “they get paid overtime for being present here for trials.”

In explaining how his theory was consistent with the charge against Baum, the codefendant who pleaded guilty, defense counsel stated:

The theory is not that [the officer] is just going to willy nilly go out and arrest people who didn’t do anything. The theory is that having made a buy it is to his advantage to — to his financial advantage to involve as many people as possible in that buy. Because it is much more likely that a man like [appellant] who basically[,] if Baum is believed[,] did nothing, it is much more likely that he will go to trial as opposed to a fellow like Baum who is caught with marked money in his pocket. And the court knows that as well.[7]

A three-pronged test has evolved for determining whether a proffer indicates evidence from which a jury could reasonably infer the bias of a witness: the proponent of the cross-examination must proffer to the trial judge facts which support (1) a genuine belief that the witness is biased in the manner asserted, (2) proposed questions that are probative of that bias, and (3) a specific personal bias on the part of the witness. Parker, supra, 586 A.2d at 723; Porter v. United States, 561 A.2d 994, 996 (D.C. 1989); Jones, supra note 1, 516 A.2d at 517. Defense counsel proffered that he was certain that he could show the officer’s personal financial motive through the officer’s own testimony. There were critical discrepancies between the officer’s trial testimony and the written reports that he had prepared shortly after the drug buy from Baum.8 Combined with the general *910circumstances of appellant’s arrest— where, according to the officer, appellant was simply one of many persons within a perimeter area, some of whom were announcing they had drugs to sell — the trial judge was presented with evidence that the defense had a genuine belief that the officer was biased, since nothing else appeared to explain the discrepancies in his testimony, the first prong of the Parker test. With regard to the second prong of the Parker test, defense counsel informed the trial judge that he wanted, for example, to ask the officer whether he was receiving overtime pay for testifying at appellant’s trial and whether he had a financial stake in cases going to trial in order to show that the officer had a personal financial motivation for arresting appellant. As to the third prong, defense counsel informed the trial judge that he could show Officer Fai-son’s personal financial motivation by questioning him about his pay status; thus, the proffer was directly tied to Officer Faison.

In the absence of a voir dire indicating that the proffer could not be proven, therefore, cross-examination regarding the officer’s financial interest should not have been denied. See Best, supra, 328 A.2d at 382 (deficiency in a proffer “does not justify the complete curtailment of questioning,” particularly where “presentation of [a] proffer was hobbled by the court’s misconception of [its] relevancy_”). Cf. Porter, supra, 561 A.2d at 996 (failure to proffer that officer had actual knowledge of pending civil suit against police department fatal to establishing officer’s bias). The government did not dispute defense counsel’s claim that such overtime in repeated appearances at trials could result in a considerable amount of money, even doubling the officer’s regular salary. Nor did the government protest when defense counsel told the judge that cross-examination regarding the officer’s financial motive was important in order for defense counsel to respond to a closing argument inevitably made by prosecutors in response to any defense attack on the officer’s credibility, i.e., that the officer was simply doing his duty. Consequently, it is difficult to understand why questions about who was paying the officer’s salary, the hours of his normal tour of duty, and the fact that he received overtime pay for appearing at trials, were not permitted to be asked. See Best, supra, 328 A.2d at 382. Further, when counsel offered to provide further details to support the inference with respect to the particular officer who identified appellant, the judge would not permit it.9

[I]f you are suggesting that you have an expert to say that yes, some officers have made more [money] than judges.... I would not permit that sort of testimony anyway. It has nothing to do with this witness. This witness’ credibility can’t depend upon what some officers do in terms of their salary. We don’t know the circumstances of those officers....

*911Even if the judge would not permit the particular questions suggested by defense counsel, the judge was required, consistent with the solicitous approach that is required toward bias cross-examination, Goldman, supra, 473 A.2d at 856 (citations omitted), to consider counsel’s suggestions for a more limited cross-examination. See Best, supra, 328 A.2d at 382. When counsel indicated his willingness to pursue a more limited form of cross-examination, the judge again refused to allow the questioning. Instead, the trial judge emphasized the fact that had Baum not pleaded guilty, Officer Faison would have had to testify at Baum’s trial.10 The judge apparently did not fully grasp the implications of defense counsel’s proffer that, irrespective of what happened to Baum, who had been caught with marked money in his pocket, Officer Faison had (if Baum’s testimony was believed) arrested appellant, a marginally involved person, in order to ensure that the officer would have court time and, hence, receive additional compensation.

Under the majority’s approach, see opinion at 904-905, it remains unclear how a defendant could show financial motivation for an arrest based on payment for court time without proceeding as the defense proposed in the instant case. Furthermore, the trial judge’s concerns about the defense proffer of financial bias could have been resolved through voir dire. If on voir dire Officer Faison had responded affirmatively to questions about whether he was being paid overtime for the time he spent in court testifying at appellant’s trial or whether he had a financial stake in the fact that cases go to trial, then the judge could have provided the government with an opportunity to respond before ruling on whether to allow the jury to hear such cross-examination. As it was, in the absence of a voir dire, the judge lacked a basis on which to determine that the proffer could not be proven.11

II

Appellant’s right to present evidence indicative of bias on the part of the key government witness was of heightened importance because the defense case, consisting of co-defendant Baum’s testimony, disputed the key government witness’ version of events.12 The prosecutor told the jury in closing argument that “the real issue” was appellant’s role, and that the government’s only eyewitness was Officer Faison, whom the prosecutor stated on three occasions during argument was just “doing his job.” By depriving the defense of an opportunity to respond to the government’s argument that the officer was just doing his job, the judge left the jury with an incomplete version of the facts, and thus without “sufficient information from which to infer bias (should it so choose).” Reed v. United States, 452 A.2d 1173, 1177 (D.C. 1982); Beynum v. United States, 480 A.2d 698, 706-08 & n. 21 (D.C. 1984).

In Best, supra, 328 A.2d at 382 n. 4, the court distinguished Davis v. Alaska, supra note 1, 415 U.S. at 308, 94 S.Ct. at 1105, notwithstanding what it acknowledged was *912“strong language suggesting that any prevention of cross-examination intended to establish bias constitutes reversible error.” 13 No such distinction can be made here. Although Officer Faison’s testimony was the only government evidence tying appellant to the crime, there were a number of serious problems with his testimony, including a faulty identification.14 There was also no physical evidence connecting appellant to the crime. In addition, the defense presented a different version of events through the testimony of Baum, who, having pleaded guilty to the indictment was subject to a mandatory minimum sentence and, as the judge instructed the jury, had nothing to gain (and something to lose) in his own sentencing if he did not testify truthfully.15 Evidence of the officer’s bias resulting from a financial motivation might have explained the inconsistencies between his trial testimony and his contemporaneous written reports and the apparent deviation from police procedures for DEA pickups, and thus, could have tipped the balance for the jury in evaluating the officer’s testimony. Accordingly, the error in the curtailment of any questioning about the officer’s pay status, see Best, supra, 328 A.2d at 382, was not harmless beyond a reasonable doubt. Delaware v. Van Arsdall, 475 U.S. 673, 684, 106 S.Ct. 1431, 1437, 89 L.Ed.2d 674 (1986); District of Columbia v. Clawans, 300 U.S. 617, 632, 57 S.Ct. 660, 665, 81 L.Ed. 843 (1937) (“prevention of all inquiry ... particularly where the excluded questions have a bearing on credibility and on the commission by the accused of the acts relied upon for conviction, passes proper limits of discretion and is prejudicial error” (citing Alford, supra, 282 U.S. at 687, 51 S.Ct. at 218)); Brooks v. United States, 516 A.2d 913 (D.C. 1986).16