On August 26, 1986, following the denial of his motion to suppress tangible evidence, appellant Marvin Brown entered a conditional plea of guilty to unlawful possession of PCP and of marijuana, in violation of D.C.Code § 33-541(d) (1988). He was placed on probation without judgment on each count for a period of six months, pursuant to the provisions of § 33-541(e). The judge ordered that the two probationary terms run concurrently. On appeal, Brown contends that the trial judge should have granted his motion to suppress because police lacked probable cause to arrest him or reasonable suspicion to stop him. We agree and reverse.
I
THE EVIDENCE
The government’s evidence at the suppression hearing consisted exclusively of the testimony of Officer Wayne B. Walker of the Metropolitan Police Department. Officer Walker introduced himself by stating that he had been a police officer for two years and had made or participated in approximately seventy “drug arrests.” He identified Brown and turned to the events that led to Brown’s arrest.
On the evening of June 27, 1986, Officer Walker was alone in a marked police cruiser, patrolling his assigned area of northwest Washington. He was wearing his police uniform. Shortly after midnight, he monitored a police “radio run” for an individual who was said to be selling drugs at the corner of 17th and Euclid Streets, N.W. The radio run was based on an anonymous telephone tip to the police. According to Officer Walker, the seller was described in the lookout as a black male, approximately 5'6" in height, wearing a white shirt with dark writing on the front and blue jeans. The writing on the shirt was not further identified, nor was there any information provided in relation to the seller’s age, build, facial hair, features, or other identifying detail.
Officer Walker testified that he was already in the immediate vicinity when he *1011heard the report over the police radio. He said that at that time, he observed approximately fifty people in the area. He quickly ruled out as suspects all but two, however, because only two potentially matched the broadcast description.1 Having so narrowed the field, the officer requested over the police radio that the description be rebroadcast. After hearing it for a second time, he eliminated one of the two remaining potential suspects because that individual was about 6'2" in height and was wearing a white shirt and white shorts.2 There was only one person left who, according to Officer Walker, matched the description which had been broadcast over the radio. That man was Marvin Brown.
Officer Walker testified that he walked up to Brown and called out “sir!” Brown turned and began to walk away at a crisp pace. The officer called out to Brown twice more, and finally established eye contact with him. Brown still did not respond. Officer Walker then stopped Brown, told him that he fit the description of somebody who was alleged to be selling drugs, and requested him to produce some identification. Brown stated that he did not have any.
After having made the stop, Officer Walker noticed an object in Brown’s right pocket. He stated that the object “looked to be about four inches [long],” and that it “just extruded from his pocket a little bit.” He asked Brown twice what the object was, but Brown did not reply. Officer Walker then conducted a patdown, holding on to the extruding object over Brown’s clothing. He testified that he asked Brown for a third time what the object was. When Brown again failed to respond, the officer seized it from Brown’s pocket. It turned out to be a film canister.
Officer Walker testified that in fifty or sixty of the seventy drug arrests in which he had participated,3 drugs had been discovered in a film canister. He stated that at one point, apparently while the canister was still in Brown’s pocket, he had thought it was a knife. Elaborating, he explained that in the past he had discovered a number of knives in lipstick containers and that, since he was alone, he was concerned about his safety. Officer Walker testified that he then opened the film canister, explaining that he did so because he did not know what was in it. From the canister, the officer extracted four tin-foil packets which later proved to contain POP and marijuana. He placed Brown under arrest.
One of the principal issues which arose during the suppression hearing was whether, and to what extent, Brown matched the broadcast description of the seller. Officer Walker was questioned on the subject by both counsel, but the results were inconclusive. The officer testified that, so far as he knew, there was no photograph taken of Brown in connection with his arrest. He said he could not recall what writing or design, if any, appeared on Brown’s shirt. He had written on his police report that Brown was wearing maroon pants; he explained the discrepancy from the blue jeans reported in the lookout by indicating that the pants might have looked blue from a distance. Officer Walker could not initially recall whether the trousers which Brown was wearing were long pants or shorts. He testified on redirect examination, however, that if they had been shorts, he would have so noted in his report.
Brown also testified briefly at the suppression hearing. He stated that he was 5'8" to 5'9" tall. He testified that he was wearing a tan shirt which he described as “just like an Ocean Pacific shirt”4 and that
*1012there was nothing at all written on it.5 According to Brown’s account, he did not match the broadcast description at all — his height, shirt, and shorts all differed from the alleged seller’s, and no other information had been provided in the tip.
II
THE TRIAL COURT’S DECISION
The trial judge explicitly disbelieved Officer Walker’s testimony that he suspected that Brown had a knife.6 The judge was also troubled by the officer’s inability to recollect what Brown was wearing on the night in question. Observing that in determining whether a citizen’s tip is trustworthy, the court “has to look how detailed the information is ... that the tip contains,” the judge expressed concern that the description of the seller was “somewhat lacking in specificity.” Nevertheless, he denied Brown’s motion to suppress because
the officer’s testimony that he observed two people who seemed to fit the description, but singled out this defendant due to his height more closely approximating the radio run than any other possible suspect, as well as his individual recollection that the other suspect wore white shorts, is a reason, in my mind, to credit his testimony, and find that he, in fact, arrested Marvin Brown with probable cause ... [t]o believe that he was the person described in the radio run.
The judge ruled against the government on its alternative theory, namely, that Officer Walker’s actions could be viewed as a proper investigatory stop pursuant to Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968), and that articulable suspicion ripened into probable cause as events subsequently unfolded. The judge based this ruling on his belief that
articulable suspicion has to come from the officer’s observations. And this officer said that he was there in the area, but had paid no attention at all to this man up until the time he got the radio run. And after that, all he saw was that the man somewhat fit the description, and walked away from him when he called him.
The somewhat paradoxical result of this analysis was that the judge found probable cause to arrest Brown but no articulable suspicion to stop him.
Ill
PROBABLE CAUSE, ARTICULABLE SUSPICION, AND TOTALITY OF THE CIRCUMSTANCES
An officer has probable cause to arrest an individual when he or she has reasonably trustworthy information at the moment of arrest “sufficient to warrant a reasonably prudent [person] in believing that the [suspect has] committed or [is] committing an offense.” Beck v. Ohio, 379 U.S. 89, 91, 96, 85 S.Ct. 223, 225, 228, 13 L.Ed.2d 142 (1979); see also Gerstein v. Pugh, 420 U.S. 103, 111, 95 S.Ct. 854, 861, 43 L.Ed.2d 54 (1975). “This standard ... represents a necessary accommodation between the individual’s right to liberty and the State’s duty to control crime.” Gerstein, supra, 420 U.S. at 112, 95 S.Ct. at 862. As the Supreme Court stated in Brinegar v. United States, 338 U.S. 160, 176, 69 S.Ct. 1302, 1311, 93 L.Ed. 1879 (1949),
[t]he rule of probable cause is a practical, nontechnical conception affording the best compromise that has been found for accommodating these often opposing interests. Requiring more would unduly *1013hamper law enforcement. To allow less would be to leave law-abiding citizens at the mercy of the officers’ whim or caprice.
The requirement of probable cause has roots that are deep in our history, for arrest on mere suspicion collides violently with the basic human right of liberty. Henry v. United States, 361 U.S. 98, 100, 101, 80 S.Ct. 168, 170, 4 L.Ed.2d 134 (1959). Thirteen years before the Declaration of Independence, Lord Chief Justice Pratt termed arrests on suspicion “totally subversive of the liberty of the subject.” Wilkes v. Wood, 19 How.St.Trials 1153, 1167, 98 Eng.Rep. 489, — (K.B.1763). The “forefathers” who wrote our Bill of Rights agreed; searches and seizures without probable cause “are the embryo of tyranny, and they well knew it.” Wrightson v. United States, 95 U.S.App.D.C. 390, 393, 222 F.2d 556, 559 (1955). The concept of probable cause is recognized as central to the protection of “the right to be left alone — the most comprehensive of rights and the most valued by civilized [people].” 1 W. LaFave, Search and Seizure, § 3.1, at 540 (2d ed. 1987), quoting Olmstead v. United States, 277 U.S. 438, 478, 48 S.Ct. 564, 572, 72 L.Ed. 944 (1928) (Brandeis, J., dissenting). It is the chief bulwark against investigative arrests proscribed by the Fourth Amendment. United States v. Short, 187 U.S.App.D.C. 142, 145, 570 F.2d 1051, 1054 (1978). Good faith on the part of the arresting officer is not enough, Henry supra, 361 U.S. at 102, 80 S.Ct. at 171, nor may probable cause be predicated on a hunch. Lathers v. United States, 396 F.2d 524, 531 (5th Cir.1968). A search is not to be made legal by what it turns up; it is good or bad when it starts and does not change character from its success, United States v. Di Re, 332 U.S. 581, 595, 68 S.Ct. 222, 228, 92 L.Ed. 210 (1948), or from evidence discovered subsequent to the arrest. Smith v. United States, 122 U.S.App.D.C. 300, 302 n. 1, 353 F.2d 838, 840 n. 1 (1965), cert. denied, 384 U.S. 910, 86 S.Ct. 1350, 16 L.Ed.2d 362 and cert. denied, 384 U.S. 974, 86 S.Ct. 1867, 16 L.Ed.2d 684 (1966). In sum, the probable cause requirement protects fundamental liberty interests which must be jealously guarded, lest their erosion bring a tear to the eye of the lady who stands guard in New York Harbor.
These considerations apply with particular force where, as here, the arrest was made without a warrant. “The informed and deliberate determinations of magistrates empowered to issue warrants as to what searches and seizures are permissible under the Constitution are to be preferred over the hurried action of officers and others who may happen to make arrests.” United States v. Lefkowitz, 285 U.S. 452, 464, 52 S.Ct. 420, 423, 76 L.Ed. 877 (1932). Absent a warrant, the burden of establishing probable cause is on the prosecution, for otherwise there would be little incentive for law enforcement agencies to bother with the formality of a warrant; moreover, the evidence comprising probable cause is peculiarly within the knowledge and control of the police. Malcolm v. United States, 332 A.2d 917, 918 (D.C.1975).
The Constitution, however, proscribes unreasonable searches and seizures. The italicized adjective imports a command of proportionality to Fourth Amendment jurisprudence. An arrest constitutes a substantial intrusion upon an individual’s liberty, and may not be lawfully effected unless the police can meet the comparatively exacting standard of probable cause. A lesser intrusion, on the other hand, requires a correspondingly lesser showing. At least since Terry, supra, it has been settled law that police may briefly detain or “stop” an individual and, if circumstances warrant, frisk him or her, even in the absence of probable cause, provided that they have an “articulable suspicion” of criminal conduct on the part of the individual sought to be detained. As the Court remarked in Adams v. Williams, 407 U.S. 143, 145, 92 S.Ct. 1921, 1923, 32 L.Ed.2d 612 (1972), “[t]he Fourth Amendment does not require a [police officer] who lacks the precise level of information necessary for probable cause to arrest to simply shrug his shoulders and allow a crime to occur or a criminal to escape.”
*1014Although the term eludes precise definition, “articulable suspicion” is and was intended to be substantially less than probable cause. As the Supreme Court recently reiterated in Alabama v. White, — U.S. —, 110 S.Ct. 2412, 2416, 110 L.Ed.2d 301 (1990), quoting United States v. Sokolow, 490 U.S. 1, 7, 109 S.Ct. 1581, 1585, 104 L.Ed.2d 1 (1989):
[t]he Fourth Amendment requires some minimal level of objective justification for making the stop.... That level of suspicion is considerably less than proof of wrongdoing by a preponderance of the evidence.
Id. 110 S.Ct. at 2416. (Citations and internal quotation marks omitted). The predicate for the stop must be “objective,” so that a “gut” feeling or “hunch” will not do. If objective justification exists, however, the threshold is not set unreasonably high.
In the present case, the prosecution predicated its claim of probable cause and artic-ulable suspicion primarily on information provided to the police by an anonymous tipster. Since Illinois v. Gates, 462 U.S. 213, 230, 103 S.Ct. 2317, 2328, 76 L.Ed.2d 527 (1983), the ordained methodology for determining probable cause in cases of this kind has been to assess the totality of the circumstances. More recently, in White, the Supreme Court, applying the approach which it adopted in Gates, stated that the existence of reasonable suspicion7 in anonymous informant cases likewise depends on the quantity and quality of the information available to the officer, and that these circumstances must be considered in the “totality of the circumstances — the whole picture.” 110 S.Ct. at 2416.
In light of White, we cannot agree with the trial judge’s apparent conception that, in deciding the question of articulable suspicion, he was restricted to a consideration of those of Brown’s activities which Officer Walker had personally observed, to the exclusion of the information contained in the radio communication which he had monitored. In Adams, supra, 407 U.S. at 147, 92 S.Ct. at 1924, the Court expressly “rejected] respondent’s contention that reasonable cause for a stop and frisk can only be based on the officer’s personal observation, rather than on information supplied by another person.” This court has previously upheld stops and frisks on the basis of informants’ tips, without requiring, as the judge evidently did here, that the arresting officer personally observe suspicious conduct. Allen v. United States, 496 A.2d 1046, 1048 (D.C.1985) (citing cases); see also United States v. Johnson, 540 A.2d 1090, 1092 (D.C.1988). Accordingly, we look to the same “totality of the circumstances” in assessing both the government’s claim that Officer Walker had probable cause to arrest Brown and its alternative contention that the facts before the officer warranted an articulable suspicion of unlawful conduct which justified an initial detention, and which ripened into probable cause as a result of events occurring after the stop.
IV
THE ANONYMOUS INFORMANT-HOW RELIABLE IS HIS REPORT?
In abandoning the “two-pronged test” of Aguilar v. Texas, 378 U.S. 108, 84 S.Ct. 1509, 12 L.Ed.2d 723 (1964) and Spinelli v. United States, 393 U.S. 410, 89 S.Ct. 584, 21 L.Ed.2d 637 (1969), to which it had subscribed for fifteen years, the Court made it clear in Gates that the factors that had been considered critical under Aguilar and Spinelli — an informant’s veracity or reliability, and the basis of his or her knowledge — remain “highly relevant” in determining the value of the informant’s report. Gates, supra, 462 U.S. at 230, 103 S.Ct. at 2328; see also White, supra, 110 S.Ct. at 2415. In evaluating the totality of the circumstances in this case, we must therefore consider the informant’s credibility and reliability and the basis for his or her knowledge. In the light of the particular facts before us, we must also address is*1015sues raised by Brown as to the sufficiency of the informant’s description and the degree to which Brown matched it, as well as Brown’s conduct on the scene and any inferences that may fairly be drawn from that. We deal with each of these topics in turn.
A. Credibility and Reliability.
In the present case, it is difficult to assess credibility or reliability because we know virtually nothing about the informant. As the Court observed in White, supra, 110 S.Ct. at 2415,
[t]he opinion in Gates recognized that an anonymous tip alone seldom demonstrates the informant’s basis of knowledge or veracity inasmuch as ordinary citizens generally do not provide extensive recitations of the basis of their everyday observations and given that the veracity of persons supplying anonymous tips is “by hypothesis largely unknown, and unknowable.”[8]
The government contends that the tipster was a “citizen” volunteering information rather than a “paid” informant, and that “a citizen is prima facie a more credible source than a paid police informant.” Allen, supra, 496 A.2d at 1048, quoting Rushing v. United States, 381 A.2d 252, 255 (D.C.1977). Although, technically, the government is right — Brown's counsel conceded in the trial court, after listening to the radio run, that the report was made by such a “citizen” — the tipster’s anonymity precludes any productive exploration of his or her arguably enhanced credibility.9
“The presumption in favor of the credibility of citizen informants is based on the assumed absence of ulterior motives.” Rushing, supra, 381 A.2d at 255 n. 3. The testimony presented by the government in the present case, however, revealed absolutely nothing about the tipster’s motivation. He or she may have been a public-spirited individual. Alternatively, the call might have come from somebody’s mendacious and mortal enemy. “Indeed, for all that this record tells us, the tipster may well have been another police officer who had a ‘hunch’....” White, supra, 110 S.Ct. at 2418 (Stevens, J. dissenting). Paid informants are supposed to be less reliable than ordinary citizens because they are generally drawn from the criminal milieu. Rushing, supra, 381 A.2d at 255 n. 3. We have no idea, however, whether the informant here was or was not a criminal (e.g., a drug dealer seeking to protect his terrain).
In light of these considerations, courts are properly wary of sustaining seizures on the basis of anonymous tips, and require a substantial measure of corroboration of information anonymously provided. An uncorroborated tip by an informer whose identity and reliability are both unknown does not constitute probable cause to make an arrest. Contee v. United States, 94 U.S.App.D.C. 297, 299, 215 F.2d 324, 326 (1954); see also Wrightson, supra, 95 U.S.App.D.C. at 392, 222 F.2d at 558. “[A] citizen who prefers to remain anonymous would seem less reliable than a citizen willing to accept personal responsibility for his accusations.” Rushing, supra, 381 A.2d at 255. Where its informant is known *1016to the police, the government has “a stronger case than obtains in the case of an anonymous telephone tip.” Adams, supra, 407 U.S. at 146, 92 S.Ct. at 1923; accord, Groves v. United States, 504 A.2d 602, 605 (D.C.1986) (informant’s credibility enhanced where he was willing to identify himself). “A named and identified person is not a ‘faceless informer’ whose inarticulated self-interest would render his information automatically suspect.” Commonwealth v. Atchue, 393 Mass. 343, 346-47, 471 N.E.2d 91, 94 (1984). A person who does not hide behind the cloak of anonymity, but who voluntarily comes forward and identifies himself or herself, is more likely to be telling the truth because he or she is presumably aware of the possibility of being arrested for making a false report. State v. Lindquist, 295 Minn. 398, 399-400, 205 N.W.2d 333, 335 (1973). See generally W. LaFave, supra, § 3.4(a), at 722-725.
In the present case, the tip was communicated to the police by telephone, rather than in person. “[Ajnonymity takes on even greater significance where there has not even been a face-to-face confrontation between the person giving the information and the police.” LaFave, supra, § 3.4(a) at 723. A tipster establishes his credibility as an interested citizen “by identifying himself and otherwise cooperating with law enforcement officials in a manner consistent with the best interest of society.” Rohrig v. State, 148 Ga.App. 869, 870-71, 253 S.E.2d 253, 255 (1979). By contrast, an anonymous telephone tip is of the “weakest reliability,” People v. Crea, 126 A.D.2d 556, 559-60, 510 N.Y.S.2d 876, 880 (2d Dept.1987), and an unverified one “do[es] not support or contribute to a probable cause determination.” Burks v. State, 293 Ark. 374, 378, 738 S.W.2d 399, 402 (1987).
This is not to say, of course, that information from a person not known to police may never be taken into account. LaFave, supra, § 3.4(a), at 724. Gates and White demonstrate the contrary; if the information has been substantially corroborated, the tip may be sufficient. Where the informant not only provides information as to facts or events observable by anyone (e.g., that a man in a white T-shirt and blue jeans is standing at a given corner) but also successfully predicts events that have not yet occurred at the time of the tip but which are thereafter verified by police, this may provide sufficient indicia of the informant’s credibility (as well as his or her basis of knowledge) to justify a stop. White, supra, 110 S.Ct. at 2415, see Gates, supra, 462 U.S. at 245, 103 S.Ct. at 2335. There is no such evidence in the present record, however, and to the still substantial extent that the informant’s credibility plays a role in the assessment of the totality of the circumstances, the anonymous telephone caller remains near the less reliable end of the spectrum.10
*1017 B. Basis of Knowledge.
An anonymous tip seldom demonstrates the informant’s basis of knowledge. White, supra, 110 S.Ct. at 2415. Here, as in Rushing, “there was no evidence indicating how the caller obtained [his or] her information or on what grounds [he or] she concluded that [the person described in the tip was] selling narcotics. Nor did the tip describe the criminal activity in sufficient detail to remedy this defect.” Rushing, supra, 381 A.2d at 256-57.
If a citizen “claims or appears to be a victim of a crime or an eyewitness to a crime, the reliability of his or her information is greatly enhanced.” Allen, supra, 496 A.2d at 1048. No such claim was made by the tipster here,11 and appearances are altogether ambiguous. The tip could have been based on first-hand information but, absent even a claim that it was, it is also possible that the tipster’s purported knowledge was second-hand or third-hand or supposition or rumor or the product of a grudge against a rival or vengeance against an enemy. “Basis of knowledge” might arguably be inferred from an accurate prediction of future events, White, supra, 110 S.Ct. at 2415, but there is no prediction, accurate or otherwise, in the present record.
C. The Adequacy of the Description of the Seller.
The trial judge was troubled by the fact that the informant’s description of the seller in this case was “somewhat lacking in specificity.” So are we.
According to Officer Walker, the radio run which he monitored described the seller as a black male, approximately 5'6", white shirt, blue jeans, with dark writing on the front of the shirt. If the informant was close enough to be able to discern that the suspect was selling drugs, one might ordinarily expect him or her to be able to identify the lettering on the shirt. Nevertheless, no information was provided as to the one fact that would make the tip distinctive — the words or letters on the suspect’s clothing — nor was anything said about the seller’s facial features or about any other distinguishing characteristics.
“As a general proposition, it may be said that the greater the number of ... identifying characteristics which are available, the inore likely it is that there will be grounds to arrest a person found with all or most of these characteristics.” 12 LaFave, supra, § 3.4(c), at 741. Descriptions applicable to large numbers of people will not support a finding of probable cause. Commonwealth v. Jackson, 459 Pa. 669, 673-74, 331 A.2d 189, 191 (1975).
The government argues that Officer Walker immediately ruled out all but two of the persons in the area as possible suspects, that one of those two was eliminated as soon as the lookout was rebroadcast, and that Brown was therefore the only person in the area who matched the description. But aside from the discrepancies between the lookout and Brown’s appearance and the paucity of identifying characteristics, the prosecutor failed to establish that the seller would probably still have been in the area at the time of Brown’s arrest. Although Officer Walker testified that he was more or less “on location” when he monitored the radio run, the government offered no evidence as to when the informant had called the police. We are therefore left to speculate as to how much time elapsed between the informant’s call and the sighting of Brown. Logically, one would suppose that the police would move with dispatch upon receipt of a tip of this kind. Supposition, however, is no sub*1018stitute for proof.13 Street scenes of the kind before us change rapidly; the seller can be gone in a jiffy.14
If the description in the radio transmission had been of a black male whose white shirt had O-C-E-A-N P-A-C-I-FI-C printed on it in dark lettering, and if the officer had spotted a black male wearing such a shirt in the immediate area who otherwise fit the description, then one could reasonably conclude, in spite of the lacuna in the government’s proof as to the time of the informant’s call, that the suspect was probably the individual described in the lookout. Cf. White, supra, 110 S.Ct. at 2417. The government’s inadequate showing could not be and was not cured, however, by “the police’s independent verification of the fact that there was a person in the area indicated wearing a white T-shirt,” Rushing, supra, 381 A.2d at 256, and, in this case, dark pants.
D. The Sufficiency of the Match.
“One question which causes the courts difficulties is whether probable cause is present when the police match up some but not all of the description with the person arrested.” LaFave, supra, § 3.4(c), at 743. Although less precision is required for ar-ticulable suspicion, similar problems arise in that context.
Not every discrepancy is fatal; “mistakes are irrelevant if there is sufficient particularized information to constitute probable cause.” Brown, supra, 125 U.S.App.D.C. at 46, 365 F.2d at 979. The fact that a part of the description does not fit is, however, obviously a negative factor. Mobley v. State, 270 Md. 76, 81-82, 310 A.2d 803, 807 (1973), cert. denied, 416 U.S. 975, 94 S.Ct. 2003, 40 L.Ed.2d 564 (1974). “[W]hat must be taken into account is the strength of the points of comparison which do match up and also whether the nature of those which did not match [is] such that an error could readily occur.” LaFave, supra, § 3.4(c), at 743.
As we have noted, the description in the lookout to which Officer Walker testified was remarkably skimpy, especially since the alleged writing on the seller’s shirt was not identified. The following chart summarizes the state of the record as to the match-up between the description of the black male who was alleged to be the seller and the actual appearance of Brown:
[[Image here]]
The trial judge did make a conclusory ultimate finding that Officer Walker arrested Brown with probable cause to believe that he was the person described in the radio run. The judge did not, however, resolve the remaining conflicts in the testimony.
A discrepancy as to the color of Brown’s pants may be explainable in view of the *1019hour of the night. “[Where] observations occurred in the early hours of the morning under artificial lighting conditions, a mistaken belief in the color of [the pants] is understandable.” People v. Washington, 89 Ill.App.3d 734, 738, 44 Ill.Dec. 925, 927, 412 N.E.2d 1, 3 (1980). Assuming that Brown knows his own height and that he is 5'8" tall, an error of two and a half or three inches in describing him is not conclusive in and of itself. The discrepancies with respect to height, lettering, and color take on greater significance, however, where no meaningful similarities have been positively established except that Brown, like the seller, is a black male.
E. Brown’s Conduct in the Officer’s Presence.
The government contends that additional support for the seizure was provided by what it describes as Brown’s attempt to “evade” the officer. On the particular facts of this case, we think that the significance of the evidence of “evasion” was marginal at most. To attach substantial importance to it would improperly penalize Brown for the exercise of his constitutional rights.
The allegedly evasive conduct to which the government alludes15 consisted of Brown’s attempt to walk away when Officer Walker spoke to him. The officer acknowledged that Brown did stop after the officer pressed the point. We recently observed in Smith v. United States, 558 A.2d 312, 316-17 (D.C.1989) (en banc) that “[topically, in those cases in which we have found that flight indicated a consciousness of guilt, the accused clearly knew that police were present and reacted by immediately running from the scene of the alleged crime.” (Emphasis added). That is not what occurred here. Brown did not run. There was no “flight”. Officer Walker’s testimony established that Brown was not eager to have a discussion with a policeman in uniform. The trial judge found no more16 and there was no more.
Citizens have no legal obligation to talk to the police. In re D.J., 532 A.2d 138, 141 (D.C.1987); see Cobb v. Standard Drug Co., 453 A.2d 110, 112 (D.C.1982). A police officer may approach a citizen on the street and speak to him if the citizen is willing to listen, but the latter has the right to proceed on his way without answering the officer’s question and without listening, and may not be detained solely for doing so. D.J., supra, 532 A.2d at 142.
[Departure ... from an imminent intrusion cannot bootstrap an illegal detention into one that is legal. To permit such justification would be effectively to create a duty to respond to the police, and would seriously intrude upon the liberty and privacy interests which the Fourth Amendment was designed to protect.
Id. (citations and internal quotation marks omitted).
We recently held en banc that
[l]eaving a scene hastily may be inspired by innocent fear, or by a legitimate desire to avoid contact with the police. A citizen has as much prerogative to avoid the police as he does to avoid any other person, and his efforts to do so, without more, may not justify his detention.
Smith, supra, 558 A.2d at 316; accord, Florida v. Royer, 460 U.S. 491, 498, 103 S.Ct. 1319, 1324, 75 L.Ed.2d 229 (1983) (plurality opinion). To say that a citizen is free to leave without responding to the officer’s questions, see Lawrence v. United States, 566 A.2d 57, 60-62 (D.C.1989), is meaningless if the exercise of that freedom generates authority for a seizure where none previously existed.
This is not to say that a suspect’s behavior on the scene is of no consequence. “Conduct by the individual which *1020suggests he is attempting to flee from a crime may also be taken into account with the available description.” LaFave, supra, § 3.4(c), at 748. Where a suspect reacts in an unusual way to police investigation short of detention, this may properly be included in the calculus. State v. Baldic, 131 N.H. 225, 226-27, 551 A.2d 977, 978 (1988) (opinion per Souter, J.). Brown’s brief attempt to exercise his right not to participate in an encounter with Officer Walker, however, did not constitute the kind of conduct on the scene that could significantly bolster the government’s showing of probable cause or articulable suspicion.17
V
THE PRINCIPLES APPLIED
A, The Scope of Review.
On appeal from the denial of a motion to suppress evidence, the scope of this court’s review is limited. Lawrence, supra, 566 A.2d at 60. Essentially, our role is to ensure that the trial court had a substantial basis for concluding that no constitutional violation occurred. Goldston v. United States, 562 A.2d 96, 98 (D.C.1989). We must accept the trial judge’s findings of evidentiary fact and his resolution of conflicting testimony. Lawrence, supra, 566 A.2d at 60.
Some courts have stated, perhaps oversimplifying the problem, that “our review of a district court’s determination of probable cause for arrest is limited by the clearly erroneous standard.” See, e.g., United States v. Purham, 725 F.2d 450, 455 (8th Cir.1984); United States v. Vanichromanee, 742 F.2d 340, 346 (7th Cir.1984). A more precise articulation was provided by the court in United States v. Campbell, 843 F.2d 1089, 1092 (8th Cir.1988):
Whether reasonable suspicion or probable cause exists to justify a seizure is a mixed question of fact and law. The findings with respect to the historical facts are reviewed under the clearly erroneous standard; the ultimate conclusion, however, is subject to de novo review. See United States v. Mendenhall, 446 U.S. 544, 551-52 n. 5 [100 S.Ct. 1870, 1875-76 n. 5, 64 L.Ed.2d 497] (1980).
See also United States v. Hoyos, 868 F.2d 1131, 1135 (9th Cir.1989) (determination of existence of probable cause is mixed question of law and fact in which legal issues predominate, and is therefore subject to de novo review; underlying facts are reviewed for clear error). We must therefore independently review the trial judge’s conclusion that probable cause existed, but that articulable suspicion did not, giving due deference to his underlying factual findings. Lawrence, supra, 566 A.2d at 60.
It is incumbent upon us, in this case as in any other, to eschew appellate fact-finding. Nevertheless, the absence of explicit findings on most of the disputed issues of fact reduces both our opportunity and our obligation to be deferential. As we have noted at page 1018, supra, the only conflict in the testimony which the judge specifically resolved was whether Brown was wearing shorts or long pants; he elected to credit Officer Walker on that point.18 The judge never addressed the other specific discrepancies, especially the dispute regarding the color of Brown’s *1021shirt and as to whether there was dark lettering on it.19
In any event, even if we treat the determination that probable cause existed as necessarily including, by implication, findings favorable to the prosecution on disputed issues which the judge did not explicitly address — e.g., whether Brown was wearing a white shirt with dark lettering— we would nevertheless be constrained to hold that the motion to suppress evidence ought to have been granted.
B. The Sufficiency of the Prosecution’s Evidence.
We now apply the principles described above to the record before us. Although the parties have concentrated their attention largely on the question of probable cause, we first assess the evidence to determine whether the facts before Officer Walker formed the basis for a reasonable suspicion justifying his initial st,op of Brown. We proceed in this order because, under our analysis, if there was no articula-ble suspicion, then a fortiori, in the light of our discussion at pages 1012-1014, supra, there was no probable cause.
During the pendency of this appeal, the Supreme Court has considered in Alabama v. White, supra, the question whether reasonable suspicion can be based on an anonymous tip. In that case, police received an anonymous telephone tip advising them that Vanessa White would be leaving a certain apartment at a particular time in a brown Plymouth station wagon with the right taillight broken, that she would be proceeding to Dobey’s Motel, and that she would be in possession of an ounce of cocaine in a brown attache case. The officers staked out the parking lot in front of the building in question and observed a brown Plymouth station wagon with a broken right taillight. A woman who later proved to be Vanessa White came out of the building at a time which the Supreme Court estimated to be within the time frame predicted by the caller.20 She was carrying nothing in her hand. Ms. White entered the station wagon and drove the most di*1022rect route toward Dobey’s Motel. This route included several turns. Officers stopped the vehicle when Ms. White was on the road on which Dobey’s is located, but before she reached the motel. With Ms. White’s consent, they searched her attache case. Finding marijuana inside, they placed her under arrest. A search incident to her arrest resulted in the discovery of cocaine in Ms. White’s purse.
Reversing the contrary holding of the Alabama Court of Criminal Appeals,21 the Supreme Court held, three justices dissenting, that the stop was lawful and that the evidence was sufficient to support a reasonable suspicion that Ms. White was in possession of contraband. The Court noted that reasonable suspicion is a substantially less demanding standard than probable cause, and that a Terry stop may therefore be based on information that is less reliable than that required to show probable cause. 110 S.Ct. at 2416. The Court held that, although not every detail mentioned by the tipster had been verified, “the anonymous tip had been sufficiently corroborated to furnish reasonable suspicion that respondent was engaged in criminal activity and that the investigative stop therefore did not violate the Fourth Amendment.” Id. Summarizing its holding, the Court stated that
[ajlthough it is a close case, we conclude that under the totality of the circumstances the anonymous tip, as corroborated, exhibited sufficient indicia of reliability to justify the investigatory stop of respondent’s car.
Id. at 2417 (emphasis added).
In reaching its decision, the Supreme Court effectively distinguished the situation in White from cases such as this one and made it plain that its holding depended on the presence of factors wholly lacking here. The Court emphasized that an uncorroborated anonymous tip, “standing alone, would not warrant a [person] of reasonable caution in the belief that a stop was appropriate.” 110 S.Ct. at 2416 (citations and internal quotation marks omitted). The Court therefore required “more than the tip itself,” id., and looked to the “totality of the circumstances,” id., to determine whether “the anonymous tip had been sufficiently corroborated to furnish reasonable suspicion that respondent was engaged in criminal activity and that the investigative stop therefore did not violate the Fourth Amendment.” Id. The Court found the required corroboration in the tipster’s predictions of future activity which the police could — and did — verify.22 The Court credited “the proposition that because an informant is shown to be right about some things, he is probably right about other facts that he has alleged, including the claim that the object of the tip is engaged in criminal activity.” 110 S.Ct. at 2417.
Critical to the determination that reasonable suspicion was established, therefore, was the anonymous tipster’s reference to future, not merely completed, activity. As the Court said in White,
[w]e think it also important that, as in Gates, “the anonymous [tip] contained a range of details relating not just to easily obtained facts and conditions existing at the time of the tip, but to future actions of third parties ordinarily not easily predicted.” Gates, 462 U.S. at 245 [103 S.Ct. at 2335-36]. The fact that the officers found a car precisely matching the caller’s description in front of the 235 building is an example of the former. Anyone could have “predicted” that fact because it was a condition presumably existing at the time of the call. What was important was the caller’s ability to *1023predict respondent’s future behavior, because it demonstrated inside information — a special familiarity with respondent’s affairs. The general public would have had no way of knowing that respondent would shortly leave the building, get in the described car, and drive the most direct route to Dobey’s Motel. Because only a small number of people are generally privy to an individual’s itinerary, it is reasonable for police to believe that a person with access to such information is likely also to have access to reliable information about that individual’s illegal activities. See Gates, supra, at 245 [103 S.Ct. at 2335]. When significant aspects of the caller’s predictions were verified, there was reason to believe not only that the caller was honest but also that he was well informed, at least well enough to justify the stop.
Id. (emphasis in original).
In the present case, at least prior to the seizure, the police acted only on static information — the tipster’s description of the seller — which afforded no basis in itself for a reasonable suspicion that Brown was selling drugs. The stop was thus based on the portion of the tip — appellant’s description— which gave “easily obtained facts ... existing at the time of the tip.” Id. Anyone could have observed the corner of 17th and Euclid, N.W., and phoned in appellant’s description. The tipster did not predict that the alleged seller would do anything except, perhaps implicitly, that he might still be selling drugs when police arrived, and there was no evidence that he was doing that. In the absence of any accurate prediction of future conduct, there is no reason to believe that the tipster had the sort of “special familiarity with [Brown’s] affairs” which the Court in White evidently viewed as essential to its conclusion that “the anonymous tip, as corroborated, exhibited sufficient indicia of reliability to justify the investigatory stop.... ” Id.23
We do not suggest that accurate prediction of future events has some talismanic quality or outweighs other constituent parts of the totality of the circumstances. The police might have had facts sufficient for reasonable suspicion if the tipster had not been anonymous, Adams, supra, 407 U.S. at 146, 92 S.Ct. at 1923 (tipster known to police officer), Groves v. United States, 504 A.2d 602, 605 (D.C.1986) (tip reliable because informant called twice and identified himself), or if the informant had provided reliable information in the past, Adams, supra, 407 U.S. at 146, 92 S.Ct. at 1923; Allen v. United States, 496 A.2d 1046, 1049 (D.C.1985) (tipster known to police and previous calls led to seizures of narcotics), or if the tip had accurately demonstrated some special familiarity with Brown’s affairs, see White, supra, 110 S.Ct. at 2417, or if the officer’s own observations before the stop had supported a reasonable suspicion of criminal activity, see Lawson v. United States, 360 A.2d 38, 39, 40 (D.C.1976) (suspect’s noticeable effort “to hide something” as police arrived was consistent with possessing a concealed weapon), or if the tipster had reported that the suspect had a pistol, see Johnson, supra, 540 A.2d at 1091-92 (emphasizing need for prompt police investigation of tips involving firearms), but none of these factors was present here.
White stands for the proposition that, without more, an uncorroborated anonymous tip that a correctly described individual is engaged in narcotics activity falls short of providing the “minimal level of objective justification,” 110 S.Ct. at 2416, which is required for a Terry stop. In the present case, there was no corroboration of any prediction of future conduct; moreover, the description of the seller was scanty and the match-up between that description and Brown’s appearance was in some measure flawed. If White was a “close case,” as the Supreme Court said it was, then the government’s evidence here was surely insufficient as a matter of law to show the existence of articulable suspicion. If there was no articulable suspicion, *1024then necessarily there was no probable cause.24
VI
CONCLUSION
For the foregoing reasons, the order placing Brown on probation without judgment is reversed, and the case is remanded with instructions to the trial court to grant Brown’s motion to suppress tangible evidence.
So ordered.
. The officer stated that the people whom he promptly eliminated from consideration included a number of young ladies, as well as men wearing shorts, windbreakers or sweat suits.
.The officer was not asked, and did not explain, why an individual who was eight inches taller than the person described in the tip, and who was wearing white shorts rather than blue jeans, remained a viable candidate even after most of those present had been eliminated.
. This translates roughly to seventy to eighty-five percent.
. The subject of this remarkable (or perhaps unremarkable) shirt generated the following vignette:
THE COURT: Was it an OP shirt?
BY [THE PROSECUTOR]:
Q: What is an OP shirt?
A: That's Ocean Pacific.
*1012Q: Oh, Ocean Pacific?
THE COURT: Come on, Ms. [Prosecutor], get with it! New wave stuff!
. In response to a question whether there was a cartoon on the shirt, Brown states that "[i]t’s not like a cartoon. Just like pictures, like — not even pictures. Just colors, and it goes all the way around."
. The judge stated, among other things, that "I really don’t think the officer thought a knife was in the film canister.” He remarked that "I don't buy that angle of his testimony.” He described Officer Walker’s credibility as "high and low,” but gave a charitable interpretation to that finding: "Not that he intentionally tried to deceive the court, but after all, when you are making a lot of arrests, you tend to get confused sometimes....”
. We assume that, for present purposes, "reasonable suspicion” as used in White is identical to "articulable suspicion” as used in Terry. The Court in White referred to Terry as a "reasonable suspicion" case. 110 S.Ct. at 2415.
. The quoted words from Gates appear at 462 U.S. 237, at 103 S.Ct. 2332.
. See In re Betrand, 451 Pa. 381, 387, 303 A.2d 486, 489 (1973):
The Commonwealth urges that probable cause was established here because the anonymous telephone call contained significant inner indicia of reliability, important details of which were corroborated by the arresting officers’ own knowledge. Apparently the Commonwealth is contending that the informant was inherently reliable since he was an anonymous citizen-informant who was too scared to reveal his identity to the police. Such an informant, argues the Commonwealth, is more reliable than the typical stoolie. However, the Commonwealth here engages in sheer conjecture, as it admittedly has no evidence that the unknown informant was either a typical stoolie or a citizen-informant. It thus has absolutely no indication of the informant’s reliability.
(Emphasis in original). Professor LaFave has suggested that "[pjolice claims that their information was received from an average citizen who was a victim of or, more likely, a witness to criminal activity should be viewed with healthy skepticism when the nature of the criminal conduct alleged and the relationship of the ‘citizen’ to that activity is more typical of that found when informants from the criminal milieu are utilized.” W. LaFave, supra, § 3.4(a), at 727.
. According to the government, “it is well settled that an anonymous citizen’s report describing the perpetrator of a crime may be sufficient, standing alone, to establish probable cause for arrest.” The cases cited by the government do not support that proposition. In Allen, supra, the citizen was anonymous only in the sense that police did not know her name; a community campaigner against drug traffic, she had frequently given police accurate information in the past, leading to the seizure of narcotics. 496 A.2d at 1049. In Carey v. United States, 377 A.2d 40 (D.C.1977), the lookout was not based on information from an anonymous tipster, but from a victim of an armed robbery; the report was "demonstrably based on personal observation.” Id. at 45. Galloway v. United States, 326 A.2d 803 (D.C.1974), cert. denied, 421 U.S. 979, 95 S.Ct. 1981, 44 L.Ed.2d 471 (1975) likewise did not involve an anonymous tip in the same sense that the present case does; a citizen told a police officer that the occupants of a specifically described vehicle had a pistol, and the police apparently could have and should have obtained the informant's name, but failed to do so. Id at 805. In Carter v. United States, 244 A.2d 483 (D.C.1968), police received information that Carter had a pistol from an unknown eye-witness; although the opinion contains little detail, there was no indication that the witness had insisted on anonymity. In Brown v. United States, 125 U.S.App.D.C. 43, 365 F.2d 976 (1966), the description of a robber which the police monitored came from a victim of the robbery, albeit one whose name the police did not know; the court stressed that it had "the virtue of being based on personal observation.” Id. at 46, 365 F.2d at 979.
Moreover, most of these cases involved violent crimes or firearms. Where this is true, as we held in Rushing, supra, 381 A.2d at 256, the need for immediate action on the part of the *1017police in the interest of public safety is substantially more urgent.
. During the course of the argument before the trial judge, the prosecutor stated that "we have an anonymous tipster, who calls in and says that she has seen this defendant engaged in the selling of PCP, and gives the description.” (Emphasis added). We have found no record basis whatever for this representation regarding what the tipster saw. Officer Walker testified only that at "a minute after midnight, I received a radio run for that location, for a subject selling drugs.” No other testimony relevant to the prosecutor's claim was adduced.
. In the present case, there is a substantial question whether Brown had all or most or even any of these characteristics. See pages 1018-1019, infra.
. See Rushing, supra, 381 A.2d at 258 (Gallagher, J. dissenting) and Galloway, supra, 326 A.2d at 806-07 (Gallagher, J. dissenting), discussing the need for the government to put its best foot forward at the suppression hearing.
. The judge based his finding of probable cause substantially on the fact that the officer eliminated “the other suspect [who] wore white shorts.” He was thus apparently operating on the assumption — not necessarily a sound one— that the suspect must still have been on the scene when Officer Walker began to look for him.
. Brown’s actions after the stop — his claim that he had no identification and his refusal to disclose what was in his pocket — are discussed at pages 1024-1025 & n. 24.
. The judge concluded that no articulable suspicion arose from Brown’s conduct in Officer Walker's presence because "all [the officer] saw was that [Brown] somewhat fit the description and walked away from [the officer] when he called him.” The judge did remark that the officer probably “factored in,” perhaps unconsciously, the fact that Brown, like numerous other persons, responded to the presence of a policeman in uniform by walking away.
. The government also contends that the police officer's expertise, and the fact that the area in which Brown was apprehended was one of high drug activity, support a finding of probable cause. With respect to this "familiar talismanic litany,” Curtis v. United States, 349 A.2d 469, 472 (D.C.1975), and see Smith, supra, 558 A.2d at 316, we fail to see how Officer Walker’s claimed expertise could have measurably assisted him in determining whether Brown matched the broadcast description of the seller. Moreover, while the fact that an otherwise ambiguous transaction occurred in a "high drug” area may make it more likely that it entailed narcotics, the character of the area is irrelevant to the questions whether Brown matched a lookout and whether it was he, rather than someone else, who allegedly sold drugs at a given location.
. Even as to the length of the pants, the judge did not mention Brown's contrary testimony or expressly weigh it against Officer Walker’s admittedly sketchy recollection.
. Moreover, the judge expressed reservations regarding Officer Walker's truthfulness in connection with the officer’s professed concern about the supposed possibility that there was a knife in the film canister. The judge was also troubled by the officer’s scanty recollection of the clothes which Brown was wearing. Despite his less than overwhelming confidence in Officer Walker’s credibility, the judge made no allusion at all in his findings to Brown’s testimony. Cf. Criminal Jury Instructions of the District of Columbia, No. 2.11 (3d ed. 1978): "In determining whether the government has established the charge against the defendant beyond a reasonable doubt, you must consider and weigh the testimony of all the witnesses who have appeared before you.” .
We do not suggest that the trial judge did not consider Brown’s testimony simply because he did not mention it, and we recognize that Brown had an obvious interest in the outcome of the proceedings. See Instruction No. 2.27. Nevertheless, in a close case potentially affecting the liberty of the citizen, we think it preferable for the court to make explicit findings on disputed points and to articulate the basis for them. Cf. Eilers v. District of Columbia Bureau of Motor Vehicle Servs., 583 A.2d 677, 684-86 (D.C.1990) (fact-finding by hearing examiner in proceedings to revoke driver’s license).
In United States v. Hoffa, 349 F.2d 20, 52-53 (6th Cir.1965), aff’d, 385 U.S. 293, 87 S.Ct. 408, 17 L.Ed.2d 374 (1966), the trial judge had instructed the jury that
there was a presumption that each witness, including the parties, has sworn to the truth and it was the duty of the jury to reconcile conflicting statements if they could, but if they could not, they were the exclusive judges of the credibility of witnesses and the weight to be given to their testimony. The Court instructed the jury that the presumption was rebuttable and may be outweighed by the manner in which the witness testified, by the character of the testimony given or by contradictory evidence.
The appellate court sustained the conviction notwithstanding a challenge to the instruction, and we think that the quoted language provides a useful framework for fact-finding for jurors and judges alike. See also United States v. Botts, 110 Daily Wash L.Rptr. 1257, 1260 n. 4 (Super.Ct.D.C. March 4, 1982).
As a matter of academic interest, we note that Brown's counsel represented at sentencing that his client had told him that he had purchased the drugs from someone who generally matched the lookout. Brown was not asked about this during his brief testimony at the suppression hearing, and there was thus no cross-examination about the film canister. The sentence eventually imposed was six months of probation without judgment — an unremarkable sentence for a purchaser.
. In White, as in this case, there was a gap in the police testimony regarding the relevant times. 110 S.Ct. at 2417.
. State v. White, 550 So.2d 1074 (Ala.Cr.App.), cert. denied, 550 So.2d 1081 (1989).
. The White tip included specific, factual data from which " 'veracity', ‘reliability’ and basis of knowledge’”, id., 110 S.Ct. at 2415, could be inferred. More specifically, the White tipster provided police not only with the suspect’s name and address, as well as salient details about her car (brown Plymouth station wagon with broken taillight), but also with the suspect’s time of departure and destination (Do-bey’s Motel) plus the location of illicit drugs (a brown attache case). Id., 110 S.Ct. at 2414. In the present case, the tipster provided only a minimal description and no predictive details aside from reported drug selling, which Officer Walker did not verify before the seizure.
. Familiarity with the suspect was acutely lacking here; Brown’s actual appearance failed to match the few details which the tipster did provide. See pages 1018-1019, supra.
. Judge Gallagher urges in his dissent that in determining whether reasonable suspicion exists, courts should look to the totality of the circumstances. We agree. He also contends that when the officer saw the bulge in Brown’s pocket, a stop and a protective frisk were justified. By the time the officer became aware of the bulge, however, Brown had already been stopped and a Fourth Amendment seizure had occurred.
Judge Gallagher quotes at length from Officer Walker’s testimony, but the key facts emerged very early in the officer’s account. According to Walker, Brown tried to avoid contact with him and to depart, but the officer stopped Brown and prevented him from doing so:
Q. Now, as you walked to the defendant, what, if anything, did he do?
A. Well, first time, I — I called him, he — he walked away.
Q. And what did you do?
A. As to — in order to acknowledge, I called him again, and he walked away. So I walked up to him pretty quick. I stopped him at that point.
Q. And what, if anything, did you do once you stopped him?
A. I advised him why I was stopping him, and told him that he fit the description of somebody selling drugs. And I asked him, did he have any identification. At that point, he stated, no.
Judge Gallagher apparently believes that a reasonable person in Brown's position would have considered himself free to leave even though, when he had attempted to do so, he had been stopped by a uniformed police officer and had been told that he matched the description of someone who had been selling drugs. We are of the opinion, however, that the officer’s frustration of Brown's efforts to depart, followed by notification that he was suspected of a crime, would have made any further attempt to walk away ill-advised as well as futile.
In any event, the trial judge remarked that when Officer Walker approached Brown after having listened to a rebroadcast of the lookout,
I believe that the defendant was probably not free to go at that point.
Noting that the officer had not been asked what he would have done if Brown had tried to leave, but that it would probably have served no purpose to ask him, the judge continued:
It’s clear to me, from hearing the testimony, that he went up there to arrest him.
Understandably, in light of the foregoing, the prosecution never contended, either on appeal or in the trial court, that it could rely on facts discovered after Officer Walker had stopped Brown to support a finding of articulable suspicion. On the contrary, the government capsulized its position in its brief to this court as follows:
In sum, the citizen’s tip, without more, warranted an investigatory stop.
In the trial court, the prosecutor insisted that the defendant’s "flight" supported probable cause, a notion directly at odds with the idea that Brown was at liberty after Officer Walker had stopped him. Our dissenting colleague’s position that the seizure occurred at some later time is thus solely his own; the prosecution has not pressed or even mentioned it, and the trial court’s views were to the contrary.
We therefore hold that no reasonable suspicion existed at the time of the seizure, which occurred when the officer, by his own account, "stopped Brown after he started to walk away,” and told him he fit the description of a suspected drug dealer. It was after, and indeed as a result of, the stop that the officer became aware of the bulge. We think that the cases on which Judge Gallagher relies, Lawrence v. United States, 566 A.2d 57 (D.C.1989) and United States v. Barnes, 496 A.2d 1040 (D.C.1985), are distinguishable in that, in each instance, the discovery of the suspicious circumstance — the drugs in Lawrence, 566 A.2d at 58-59, and a bulge in Barnes, 496 A.2d at 1041-45 — occurred at a time when the defendant was free to leave, and thus preceded the seizure. Cf. California v. Hodari D, — U.S. —, —, 111 S.Ct. 1547, 1549-51, 113 L.Ed.2d 690 (1991). The officer’s own testimony in the present case, reinforced by the judge’s remarks, renders this analysis inapplicable here.