concurring:
I write separately to underscore the narrowness of the court’s opinion.
I.
A.
In citing Luce v. United States, 469 U.S. 38, 105 S.Ct. 460, 83 L.Ed.2d 443 (1984), the court today does not formally adopt it. Luce held that, in order “to raise and preserve for review the claim of improper impeachment with a prior conviction, a defendant must testify.” Id. at 43,105 S.Ct. at 464. Here, in contrast, in order to rebut Butler’s anticipated accusations of police bias, fabrication, and harassment, the government sought to introduce in evidence six of Butler’s prior “bad acts,” as reflected in three police “stops,” one formal arrest for an unadjudicated charge, an acquittal after trial, and one juvenile adjudication — none of which qualifies as a “prior conviction” under D.C.Code § 14r-305 (1995 Repl.).1 The trial court effectively ruled that the government could do so, strictly in rebuttal.
Apparently as a result of this ruling, Butler abandoned his bias, fabrication, and harassment defense. Thus, the question is: whether Butler has preserved for review, through proffered evidence, his challenge to the government’s proffered rebuttal, without having testified himself (or called other defense witnesses to testify or proceeded with cross-examination of the police officers on bias) in support of his abandoned defense.
*390B.
This case has much in common with Marshall v. United States, 628 A.2d 551 (D.C.1992), where we sustained the trial court’s in limine2 ruling that the government could cross-examine the defendant’s “character witness concerning a previous arrest of the appellant for possession of a controlled substance (PCP).” Id. at 555. There, the government, citing Luce, argued that the appellant should not be allowed to challenge the ruling because the claimant’s witness had not, in fact, testified. In sustaining the ruling on the merits as a matter of law, we said that “we do not need to reach the issue raised in Luce.” Id. at 556. Nor did we have to review the adequacy of appellant’s proffer at trial. We noted, however, that the failure of the defense to proffer particular details concerning the time periods the witness and the community knew the appellant posed troubling uncertainties.
Those [time] factors may, or may not, have affected the trial court’s ruling had [the court] known of them. But, this failure to present them in a more precise proffer, or by the witness’s actual testimony, prevented those uncertainties from being explored.
Id. at 556. We therefore indicated that, if we had been unable to sustain the court’s in limine ruling as a matter of law, we might well have held that the appellant’s challenge failed for lack of “a more precise proffer,” id. — as we rule in this case.
Marshall accordingly made clear that af-firmance in a case such as this can be premised on lack of an adequate proffer. Thus, we need not — and do not — rule here, as Lace would have it, that Butler is foreclosed from contesting the trial court’s ruling on appeal simply because he abandoned his intended bias/fabrication/harassment defense at trial.
Judge Reid’s opinion for the court makes clear that “a more precise proffer,” Marshall, 623 A.2d at 556, could have saved appellant’s right to appellate review of the trial court’s in limine ruling. See ante at 387. The court’s opinion states that “the principles enunciated in Luce are applicable” in this jurisdiction only when three criteria are satisfied: (1) the trial court’s in limine ruling is not final, (2) defendant does not testify as to the proffered but abandoned defense theory, and (3) “the trial court cannot reasonably rule on the in limine motion on the basis of the proffered testimony.” Id. (emphasis added). If, therefore, the in li-mine ruling is final, the defendant (or other defense witness) does not testify, but the trial court reasonably can decide the motion on the basis of proffered defense testimony, the Luce rule (“a defendant must testify”) does not apply; the appellate court can review the merits of the defense argument on the basis of the parties’ respective proffers. Indeed, in light of the two concurring opinions in this ease, I assume that Judge Reid would have had the court expressly adopt Luce if she had believed that an across-the-board rule against standing to appeal based on trial court proffer was necessary to the result in this case. As the following discussion makes clear, however, the case is decided on the limited basis of Butler’s inadequately proffered defense.
C.
The opinion for the court makes clear that, in this case, we are unable meaningfully to review the pretrial ruling on the admissibility of the government’s proffered evidence of Butler’s prior contacts with the police, because Butler himself failed to proffer sufficiently detailed evidence of the bias, fabrication, or harassment which the government’s testimony was intended to rebut. Thus, Butler’s failure to put on live testimony, instead of a proffer, to establish his defense would not preclude our addressing the merits of that defense if the proffer had been adequate.
In anticipation of Butler’s bias/fabrieation/harassment defense, the prosecutor told the court:
[PROSECUTOR]: I believe the defense in this case ... will be that Officer Spald-*391ing and that he’s been harassing him over a significant period of time, and that this whole thing was concocted by Officer Spalding to get back at the defendant. Of course, the Government’s position is that, that’s false. If we do get into those issues, Your Honor, the Government is prepared to present evidence to the Court of what the contacts have been with Mr. Butler in this case. And those have been criminal contacts, one that led to a juvenile arrest and conviction for possession with intent to distribute. has it out for the defendant in this case
The prosecutor then began to describe other such “criminal contacts” between Officer Spalding and Butler.
Before the prosecutor completed his list, defense counsel acknowledged that police bias, harassment, and fabrication would be a defense:
[TRIAL COUNSEL]: [C]ertainly I think certain motive to fabricate bias against the defendant is legitimate grounds for cross examination. I think no one will — will dispute that.
And so that will be — and—and this officer having stopped my client previously, driven by his house, harassed him in a number of unofficial ways, is also legitimate grounds for cross examination.
More specifically, in support of Butler’s bias/fabrication/harassment defense, Butler’s counsel elaborated that the arresting officers had a “grudge” against him. Counsel told the court:
[TRIAL COUNSEL]: The basis for the grudge would be that back in 1992 these two officers, Officers Baker and Spalding were — my client was a witness for the defense in a case in which Officers Baker and Spalding were the arresting officers and the two main Government witnesses.
* * * * * *
[TRIAL COUNSEL]: The defendant was acquitted in that case and the officers have been threatening, harassing my client and his family since that time.
[THE COURT]: Threatening the defendant and harassing his family?
[TRIAL COUNSEL]: Yes, Well, harassing — I would say threatening the defendant and harassing his family. I’m not going to allege threats against his family.
[THE COURT]: How were they threatening him?
[[Image here]]
[TRIAL COUNSEL]: They have threatened to get him and to make him pay and made those kinds of threats against him. They also — they also expressed that they had very, very great anger with my client, to the defendants’ lawyer in that case. But I don’t anticipate calling him as a witness.
ij: sj: # ‡
[THE COURT]: ... All right, [counsel], why don’t you just indicate what it is that — that is — that is the thrust of your defense here.
[TRIAL COUNSEL]: Wei, one of the — one of our — one of the points in our theory of defense, because that is not by any means the entire defense theory, but relevant to this issue is that the main reason that this officer has a grudge against my client is that my client was a defense witness for a defendant in a trial in 1992 in which Officers Baker and Spald-ing were the two chief Government witnesses.
That trial resulted in an acquittal of the defendant. And Officer Spalding and Officer Baker basically have harassed my client ever since then; have made their displeasure with him for testifying known; have expressed their dislike of him, and have harassed him and — and stopped members of his family and — and engaged in a multitude of harassing acts against my client. And that’s one of the bases for their fabrication of this charge.
Trial counsel then argued that evidence of such harassment did not open the door to admission of prior arrests evidence — even “legitimate stops” evidence — in rebuttal, because “those official stops don’t preclude harassment_ Arresting ... someone for a legitimate reason doesn’t necessarily mean you haven’t harassed the individual in the past.”
*392The prosecutor then began to describe the evidence of Butler’s “prior encounters” with the police that the prosecutor intended to use in rebuttal to show that many police contacts with Butler had been for legitimate law enforcement purposes and, as a result, tended to refute an impression of encounters characterized as harassment.3
Defense counsel then indicated she was not arguing that the particular incidents to which the prosecutor had referred were “harassing stops.” The court asked what counsel meant by “harassing.” Counsel replied:
[TRIAL COUNSEL]: That, for example, pulling — the defendant over, berating him and then letting him drive off for no— for no — no lawful purpose; stopping cars in which my client was a passenger; driving back and forth past my client’s house at night; stopping members of his family as they were coming in and out of their house. That — that, alleging that those events take place is not rebutted by — by the officers saying well he’s a drug dealer and I know it and I[’ve] seen it.
The prosecutor then said:
[PROSECUTOR]: [T]he officers ... have claimed to me that their encounters with the defendant have always been in an official capacity [and] that the Government’s position is that the defense are going to have witnesses to come up and lie.
* * * * H*
Now, you know, if they’re going to have these witnesses come up and say this we should be allowed to explain, no, that’s not what happened, ladies and gentlemen. These are the contacts that we’ve had with the defendant. And I think they can’t have their cake and eat it too. They can’t just say, oh, this is prejudicial. Well, their evidence is prejudicial to the Government’s case. And I think the whole issue is irrelevant to the facts of this case; that you had a legitimate traffic stop that led to an incident.
Soon thereafter, the trial judge ruled “preliminarily” that if the defense were to cross-examine the police officers by referring to alleged harassment, the government would be allowed “to bring forth the other contacts that they have had with the defendant, to bring those ... contacts in full context.”
At this point in the proceeding the government had specified three of the six police/Butler encounters on which it intended to rely. See supra note 3 (Nos. 3, 5, and 6). The government’s detailed proffer, therefore, had shifted the burden of evidentiary production to Butler — a burden he never met. Butler had specified no particulars about the alleged “multitude of harassing acts” other than these government-proffered encounters. Conspicuously missing from the entire colloquy among the prosecutor, defense counsel, and the court was any defense proffer of Butler family witnesses (other than Butler himself) who would testify about particular *393incidents of harassment. Nor was there any-proffered testimony by Butler that would specify incidents that clearly were other than those involving legitimate stops or formal arrests on which the government expressly proffered it would rely in rebuttal. See supra note 3.
Accordingly, the defense proffer — allegedly premised on incidents other than those cited by the government — was never given in enough detail (1) to demonstrate that the defense’s police/Butler contacts did in fact differ from those relied on by the government, and thus (2) to permit the trial court, and this court on review, to evaluate Butler’s contention that the government’s incidents were (a) irrelevant, inflammatoiy rebuttal evidence that, on balance, (b) would be more prejudicial than probative. In sum, the defense proffer for cross-examination of the police never matched the detail of the government’s proffered rebuttal by the police. See supra note 3. As a result, no court could evaluate whether the incidents contemplated by each side were the same or different — an evaluation critical to deciding whether the prosecution or the defense was correct in arguing, respectively, that particular Butler/poliee contacts should, or should not, be admitted in evidence. If Butler had proffered the “multitude of harassing acts” in the same detail that the government proffered its proposed rebuttal evidence, see supra note 3, we would be in a position to review — and, according to Judge Reid’s opinion, would have reviewed — the merits of Butler’s argument that the government’s rebuttal evidence would have to be excluded at trial.
As foreshadowed in Marshall, this court correctly heeds Luce’s advice that we cannot “ ‘rule on subtle evidentiary questions outside a factual context.’” Ante at 388 (quoting Luce, 469 U.S. at 41, 105 S.Ct. at 463). Here, the probative value/prejudicial impact of the government’s proffered rebuttal evidence is impossible to determine without the significantly more detailed context that would have emerged had Butler proffered particular evidence of police impropriety.
II.
To put the Luce discussion in perspective we should note, first, that the Supreme Court’s decision — however broadly it should be interpreted — is limited to evidentiary rulings in the federal courts. For a variety of reasons, quite a few state courts have expressly declined to follow it.4 This noncon-stitutional, evidentiary decision is not the kind of Supreme Court ruling to which state courts (including this court) owe any particular deference. See Robinson v. United States, 623 A.2d 1234, 1241 n. 9 (D.C.1993) (rejecting Supreme Court interpretation of federal rule of evidence as inconsistent with D.C. law); State v. Whitehead, 104 N.J. 353, 517 A.2d 373, 375 (1986) (rejecting Luce as nonbinding, nonconstitutional interpretation of federal evidentiary law).
Second, the literal holding of Luce, that a non-testifying defendant failed to preserve the right to challenge on appeal the trial court’s ruling in limine authorizing impeachment by prior conviction under Fed.R.Evid. 609(a)(1), is of limited, if any, relevance in this jurisdiction. In the Superior Court, pri- or convictions “shall be admitted” for impeachment, D.C.Code § 14r-305(b)(l), under specified circumstances, without discretionary weighing of “probative value” against “prejudicial effect,” as the federal rule requires. Thus, any controversy over the use of prior convictions will pose a legal question for appellate review not dependant on a trial record; Luce, as to prior convictions, is inapplicable here.5
*394The question, then, is whether the Luce rationale should extend to limiting the defendant’s right to appeal other types of in li-mine rulings when the defendant elects not to testify, or not to put on particular witnesses, as a result of the ruling.
Trial court rulings in limine, whether sought by the defendant or, as in this case, by the government, are useful because in deciding, early on, whether particular evidence is admissible — e.g., prior convictions, other “bad acts,” various statements — such rulings clarify how the parties plan to proceed, minimize guesswork and surprise, and thus generally help provide an orderly, fair trial for everyone concerned. Indeed, “[i]t is only after a ruling on the admissibility of a conviction” or other bad act, for example, “that the prosecutor and defense counsel can make an informed decision how to effectively try the case.” State v. McClure, 298 Or. 336, 692 P.2d 579, 583 (1984).
This court has recognized and encouraged the in limine procedure. For example, although we have not adopted the portion of Fed.R.Evid. 404(b) entitling a defendant to advance notice if the prosecutor intends to impeach with other crimes evidence, we have said that the trial court has discretion to require such notice — thereby inviting a challenge in limine — and that the government risks reversal for abuse of discretion in event of unfair surprise. See Johnson v. United States, 683 A.2d 1087, 1100 n. 17 (D.C.1996) (en banc); Ford v. United States, 647 A.2d 1181, 1185-86 (D.C.1994); cf. Hunter v. United States, 606 A.2d 139 (D.C.) (encouraging use of in limine motion by defendants to determine proper scope of argument by government during closing arguments), cert. denied, 506 U.S. 991, 113 S.Ct. 509, 121 L.Ed.2d 444 (1992). There may be occasions when a trial judge wishes to withhold a ruling on a motion to exclude evidence until relevant trial testimony is complete, in order to weigh probative value against prejudicial effect, but “this should be a rare occurrence.” McClure, 692 P.2d at 583. “[I]n most cases the trial judge should be able to make this determination from an offer of proof in a pretrial motion.” Id. A mini-trial is unlikely; the prosecutor and defense counsel — as officers of the court — can provide the judge with sufficient details about the evidence they expect to present, and the judge can rule accordingly. See id.; see also Settles v. State, 584 So.2d 1260, 1263-65 (Miss.1991); Whitehead, 517 A.2d at 376-77.
Ultimately, the reason for motions and rulings in limine, beyond the practicalities required for an orderly, fair trial, may be to prevent an impermissible burden on the defendant’s constitutional right to téstify in his or her own defense. See Apodaca v. People, 712 P.2d 467, 472-73 (Colo.1985); People v. Moore, 156 A.D.2d 394, 548 N.Y.S.2d 344, 346 (1989). A defendant, after all, has a right to know, within limits, what is coming, in order to structure the defense, including a decision whether or not to testify. See Apodaca, 712 P.2d at 472-73; McClure, 692 P.2d at 583; see also Settles, 584 So.2d at 1262-64; Whitehead, 517 A.2d at 376; cf. United States v. Pryce, 938 F.2d 1343, 291 U.S.App. D.C. 84 (1991) (refusing to extend Luce to situation where trial court indicated a certain line of cross-examination by defense counsel would not be permitted; government argued that defendant should be required to ask the questions on the record, in front of the jury, to preserve the issue on appeal), cert. denied, 503 U.S. 941, 112 S.Ct. 1488, 117 L.Ed.2d 629 (1992). Although a chill on the right to testify is not asserted in this case, Butler’s unquestioned intention to defend on grounds of police bias/harassmeni/fabrication, through cross-examination and, possibly, through third-party witnesses, at least indirectly implicates the right to testify, because the court’s in limine ruling on these proffered defense tactics, if adverse, would certainly bear on Butler’s only remaining alternative for presenting a defense: testifying himself. Thus, evidentiary rulings in limine are very powerful shapers of a trial, significantly affecting the fundamental right to present a defense.
Under the Luce theory, Butler would be required to present his proffered defense— cross-examination, witnesses, and perhaps his own testimony — as a condition of preserving his right to appeal the claimed trial court *395error that permitted “bad act” evidence against him. But Luce, as a practical matter, implies much more. If adopted as the law, it would mean that Butler could preserve his appeal right only by putting his worst foot forward at trial — a roundly impeached defense — because of the claimed erroneous ruling. Alternatively, of course, he could put a better foot forward under the circumstances — a diluted defense without the feared impeachment (as Butler did here)— but only by foregoing his right to appeal the claimed error affecting the defense he had preferred to offer. Absent a right of interlocutory appeal, those are the only two choices under Luce.
The Luce alternatives are harsh indeed. If the defendant elects to testify and guesses wrong about the propriety of the allowed impeachment, the defendant will have preserved the right of appeal but, as it turned out, will have put on a worst-choice defense. If, on the other hand, the defendant elects not to testify, the defendant will have put on a second-choice defense and lost the right of appeal. The defendant, accordingly, will never have had an opportunity to test on appeal the first-choice defense — unim-peached or unrebutted testimony — without having crippled that defense at trial, through impeachment or rebuttal, as the price of admission to the appellate court. That price is too high.
The purpose in allowing a motion in li-mine is to permit a witness to testify without threat of use of inadmissible evidence. If the threatened use of inadmissible evidence can prevent the defendant from testifying altogether and also deny her the opportunity to appeal an erroneous ruling on the admissibility of the evidence, the State would have multiple illegitimate opportunities to silence defendants, and the very purpose of the motion in limine would be lost.
State v. Lamb, 84 N.C.App. 569, 358 S.E.2d 857, 865 (1987), aff'd, 321 N.C. 633, 365 S.E.2d 600 (1988); but see State v. Hunt, 475 S.E.2d 722, 725-26 (N.C.App.1996), (following Luce and classifying above-quoted language as dicta).
Luce suggests two fundamental reasons for conditioning the right of appeal as it does (and thus offering the defendant two unsatisfactory choices). First, an in limine ruling is inherently tentative and not necessarily premised on reality; as the “case unfolds,” the trial judge may change the in limine ruling to .the defendant’s benefit, Luce, 469 U.S. at 41, 105 S.Ct. at 463, or the government may elect not to use the prior conviction to impeach if the government has other means for doing so, id. at 42, 105 S.Ct. at 463-164, or the defendant may choose not to testify for reasons other than impeachment with a prior conviction, id. Thus, according to Luce, the record underlying the ruling may very well become obsolete. Second, because the ruling at best is based on proffers, not on actual testimony, the record available for appeal — even if the court, prosecutor, and defense counsel do not change position — will be too imprecise and incomplete to permit appellate review of the ruling, including review for harmless error. See id. at 41-42, 105 S.Ct. at 463 — 464. These reasons are unpersuasive.
As to the first, I believe the possibilities for a changed trial court ruling, or for a party’s change of mind, are far more speculative — far less likely to occur — than the usual scenario when a trial judge rules adversely to the defense in limine. Typically, the judge will stick to the ruling, the prosecutor will stand ready to use the prior acts for impeachment if allowed to do so, and the defendant, as a result, will decide not to testify (or not to put on the impeachable witnesses). See Whitehead, 517 A.2d at 377; see also Settles, 584 So.2d at 1263. Furthermore, even if the Supreme Court is correct in surmising that some defendants will file pretrial motions in limine in order to “plant” reversible error when they have no intention of taking the stand at trial, Luce, 469 U.S. at 42, 105 S.Ct. at 463-464,1 believe this possibility is minimal, although an assumption either way is speculative. In any event, I see no reason to base a rule of law, as Luce does, on an unproved assumption that such “error-planting” would be common enough to override the more logical assumption that defense counsel file motions to establish results on which they intend to rely. But even if there *396is a likelihood that some defendants would try to “plant” reversible error without intending to rely on it, we should not address that risk by creating a disincentive for taking the stand. Finally, the potential for “error planting” abuse is a risk worth taking because the alternative — a policy that chills a defendant’s right to testify and fosters waiver of the appeal right — is contrary, in my judgment, to the fair administration of criminal justice.
This case provides a good example of why the first Luce concern will often, if not usually, be illusory. On this record, I am easily satisfied that Butler was committed to his bias/harassment/fabrication defense not merely because he said so but also because the prosecutor anticipated it by filing the motion in limine. The prosecutor’s motion makes clear, moreover, that the government would not have declined to introduce its “bad acts” evidence if Butler had put on the anticipated defense. Finally, had the defense proffer been detailed enough, I see no reason why the trial court’s ruling would not have become final, rather than preliminary, at trial based on a proffered record adequate for our review. In short, I am confident that the record of the ruling in limine is settled enough — the court’s and the parties’ positions are firm enough — to assure that we would not be reviewing a hypothetical case, and, in particular, to assure that the ruling affected the defense strategy.6 The Luce rule that would take away such appellate scrutiny on a case-by-case basis is altogether too rigid.
The Supreme Court’s other basic concern in Luce — a trial court record, based on proffers, that is too imprecise and incomplete for review — is also much overstated. The Court worried that an' appellate tribunal cannot adequately review the weighing of probative value against prejudicial effect without knowing “the precise nature of the defendant’s testimony” — information that “is unknowable when, as here, the defendant does not testify.” Luce, 469 U.S. at 41, 105 S.Ct. at 468. The Court added that harmless error review would be virtually impossible because the appellate court would be deprived of any way of evaluating how the erroneous impeachment would have impacted the defendant’s testimony and affected the outcome, see id at 42, 105 S.Ct. at 463-464; thus, any error that “presumptively kept the defendant from testifying” would likely “result in the windfall of automatic reversal.” Id.
This analysis overlooks two considerations. First, if in limine rulings are desirable, indeed necessary on occasion, to achieve an orderly, fair trial for all concerned (recall that the government made the motion here), then they should be encouraged. Because an interlocutory appeal of such a ruling is not allowed, however, an incorrect ruling will continue to shape the trial. It is therefore not correct to say the defendant would receive a “windfall” if an appellate court were to hold, on review after trial, that the judge had erred and the defendant was entitled to a new trial free of the proscribed impeachment. The very nature and utility of the in limine procedure, which calls for a trial court decision based on proffers, would justify the reversal and remand precisely because the ruling presumably will have chilled the defendant’s exercise of the fundamental right to present a defense. The very nature of the review focuses on an issue in a record much more circumscribed than the entire trial universe. The very price of encouraging in limine rulings without right of interlocutory appeal, therefore, is a narrower review than the usual error/harmless error analysis based on a complete trial record. If one is con*397cerned that an award of a new trial based on a ruling derived from proffers would be a “windfall,” then an interlocutory appeal on an expedited basis is the answer, not compelled testimony, infected by unlawful impeachment, as the price of a later appeal.
The foregoing analysis proceeds from a premise that the record based on proffers is complete enough for appellate review. There is no reason why in many if not most cases the defense cannot supply a proffer, or on occasion even testimony out of the presence of the jury,7 that could provide not only the trial court but also the reviewing court with a record detailed enough to ascertain whether the ruling in limine was erroneous. See Settles, 584 So.2d at 1263-64; Whitehead, 517 A.2d at 376; McClure, 692 P.2d at 582-83. This is the second reason why Luce’s concern about an unreviewable trial record is unconvincing.
In the present case, the trial court ruled in limine that the government could use Butler’s prior “bad acts” to rebut the defense case presented through cross-examination of police officers and, possibly, through defense witnesses (or even through Butler’s own testimony). As Judge Reid points out, a sufficiently detailed proffer of the bias/harassment/fabrication defense could have provided not only the trial court but also this court with an adequate basis for evaluating the defense motion in light of the government’s proffered “bad acts” evidence. Had that been done, the result here might have been different. There would have been no need for Butler to put on his complete defense at trial, undermined by allegedly erroneous impeachment/rebuttal, if he had proffered the details of the defense to facilitate both a definitive tidal court ruling and this court’s appellate review.
The Supreme Court said in Luce that “[requiring a defendant to make a proffer of testimony is no answer; his [or her] trial testimony could, for any number of reasons, differ from the proffer.” Id. at 41 n. 5, 105 S.Ct. at 463 n. 5. True. But this overlooks the fact that the very purpose of proceeding in limine to resolve evidentiary questions in a criminal trial is the need to give both sides an opportunity to know, early on, what is coming and, as a consequence, the opportunity to make rational trial choices. Not every eventuality is predictable, of course; trials take unforeseen turns. But counsel, acting in the best interests of their clients — and as officers of the court — are, presumptively, not involved in prevarication when they seek evi-dentiary rulings. Presumably they are seeking answers on which to rely, for the reasons they give. See Whitehead, 517 A.2d at 377; Moore, 548 N.Y.S.2d at 346; McClure, 692 P.2d at 583, 584 n. 4. Indeed, if the defense were to prevail on a motion in limine based on proffered testimony that did not materialize as represented, the court obviously could change the ruling to permit withheld impeachment or rebuttal. Surprises, of course, will occur, but the likelihood of inaccurate proffers is substantially lessened by checks inherent in the system. In sum, the scope of an in limine ruling, based on detailed proffers, reflects an evidentiary universe with an imperfect but sufficient record for making the kind of decision needed at the time and later on appeal.
Whether the defense makes the motion or responds to it, the defense, of course, will have a burden of persuasion or production that requires a proffer detailed enough to satisfy both trial and appellate courts that the ruling should favor the defense. If the trial proffer — as in this case — is not sufficient to permit the reviewers to know whether the defense challenge was correct or not, then the review on the merits cannot go forward. But this is not to say — as Luce assumes — that a proffer never can suffice. It can. It can be detailed enough. But, even *398more importantly in answer to Luce — and this bears repeating — the record with the ruling in limine is the correct and adequate basis for review precisely because the criminal justice system says, in effect, that in limine procedure is a worthwhile, virtually inherent component of a trial, and thus that its integrity must be recognized by appellate review that acknowledges the trial record limitations. If traditional harmless error review is prevented by the absence of the very unimpeached testimony that trial court error kept from the trial, then so be it; let reversal for erroneously thwarting the first-choice defense be automatic — not because a defendant deserves a “windfall” but because this is the anticipated price of a justice system that properly says, all things considered: an interlocutory appeal is unavailable; therefore, we will not force a defendant to give impeached or rebutted testimony as the price of challenging the in limine ruling on appeal after trial.
Luce may make good sense theoretically, but in practice it unnecessarily forces defendants to make an unfair election between (1) putting on a last-choice defense — e.g., testifying and getting roundly impeached — in order to save the right to appeal the in limine ruling, and (2) putting on the best possible, though diluted, defense in light of that ruling while losing the right of appeal in event of conviction. There is a workable middle ground — use of the detailed proffer — that can resolve the Luce dilemma without the Luce rule.
I am satisfied that today we premise our opinion on that workable approach and correctly rule that Butler had standing to challenge on appeal the in limine ruling, but that his defense proffer at trial was inadequate for our review.
. Evidence of a juvenile adjudication cannot be introduced as a prior conviction to impeach the testimony of a defendant. See D.C.Code § 16-2318 (1989) (juvenile adjudication or disposition is not a “conviction of a crime”); Brown v. United States, 338 F.2d 543, 547-48, 119 U.S.App. D.C. 203, 207-08 (1964) (holding that a juvenile adjudication is not a "conviction” of a "criminal offense” for purposes of impeachment by prior conviction under D.C.Code § 14 — 305).
. Luce defines “in limine" as referring “to any motion, whether made before or during trial, to exclude anticipated prejudicial evidence before the evidence is actually offered." 469 U.S. at 40 n. 2, 105 S.Ct. at 462 n. 2.
. Eventually the prosecutor specified six such encounters, as summarized in Butler’s own brief:
(1) On August 24, 1992 Mr. Butler was stopped for driving without a license; the same day Mr. Butler was also allegedly observed handing money to a person named Jerome Waller. Mr. Waller was arrested for drug possession, Mr. Butler testified on Mr. Waller’s behalf at trial, and Mr. Waller was acquitted. Mr. Butler was never arrested in connection with either of these events.
(2) On August 28, 1992, Officer Spalding arrested Mr. Butler for receiving stolen property when Mr. Butler was stopped for operating a motorcycle with stolen license tags. Mr. Butler was charged in juvenile court and the charge was later dismissed as part of a plea.
(3) On September 4, 1992, Officer Spalding arrested Mr. Butler and charged him with possession with intent to distribute cocaine. Mr. Butler pled guilty and was adjudicated in juvenile court. The receiving stolen property charge was dismissed as part of the plea.
(4) On March 9, 1993, Officer Spalding allegedly observed Mr. Butler standing with several other persons, one of whom displayed a pistol. Although Mr. Butler was stopped, he was never arrested because there was no evidence that Mr. Butler ever had possession of a weapon.
(5) In February or March of 1993, Officer Spalding stopped Mr. Butler and allegedly recovered drugs from him, but made no arrest because the buyers were not caught.
(6) In September of 1994, Officer Spalding was one of the officers who participated in the arrest of Mr. Butler in connection with an armed robbery. Mr. Butler was acquitted at trial.
. See, e.g., Apodaca v. People, 712 P.2d 467 (Colo.1985); State v. Ford, 381 N.W.2d 30 (Minn.Ct.App.1986); Settles v. State, 584 So.2d 1260 (Miss.1991); State v. Whitehead, 104 N.J. 353, 517 A.2d 373 (1986); People v. Moore, 156 A.D.2d 394, 548 N.Y.S.2d 344 (1989); State v. McClure, 298 Or. 336, 692 P.2d 579 (1984); Commonwealth v. Richardson, 347 Pa.Super. 564, 500 A.2d 1200 (1985), appeal denied, 525 Pa. 644, 581 A.2d 571 (1990).
. From time to time there may be questions, for example, whether a particular prior conviction is admissible under § 14-305(b)(l) as involving, for example, "dishonesty or false statement.” That inquiry would be purely one of law not subject to the Luce rule. See post at 398 (Newman, J., concurring) (Luce rule inapplicable where "evi-dentiary issue ... turns on ‘legal and not factual considerations' ”) (quoting Luce, 469 U.S. at 44, 105 S.Ct. at 464 (Brennan, J., concurring)); Set-*394ties, 584 So.2d at 1263; Whitehead, 517 A.2d at 376; McClure, 692 P.2d at 583.
. It is true, of course, that a defendant will always be free to decline to take the stand or to refuse to pursue a particular line of questioning, regardless of any proffers or commitments offered by defense counsel. See McClure, 692 P.2d at 583 n. 2 (noting that "despite good faith representations by counsel, a defendant will always be entitled to change his or her mind about taking the stand"). New Jersey has taken this observation to its logical extreme and held that a defendant need not promise to testify or proffer an outline of her testimony in order to preserve the reviewability of the in limine ruling on appeal. See Whitehead, 517 A.2d at 377. While this may be an appropriate rule in the context of impeachment of a defendant by prior convictions, in a case such as this one a defendant necessarily must make a sufficiently detailed proffer on the record for the trial court (and the appellate court) to determine the admissibility of the evidence. See ante at 388-389.
. In rare circumstances where a defendant's proffer is deemed inadequate to provide a basis for the trial court’s ruling, the court could use a procedure similar to that used in suppression hearings. See In re F.G., 576 A.2d 724, 727-28 (D.C.1990) (en banc). Specifically, the defendant could testify at a pretrial hearing, out of the presence of the jury, to establish any necessary foundation for the court's ruling, without waiving her right not to testify at trial. See id. at 727. If the defendant then chose not to testify at trial, the hearing testimony would be inadmissible; if the defendant proceeded to testify at trial, she could be impeached with the hearing testimony as appropriate. See id. at 727-28 & n. 4.