dissenting.
*172I respectfully dissent. In my view, the court errs today in two significant respects. First, I disagree that Captain Sullivan was in hot pursuit — -not after instructing Officer Carolyn to get the car which, was parked a quarter of the way up Gratz Street, waiting approximately a minute to get picked up, backing out onto 66th Avenue, and only then proceeding in the direction that the two men he had briefly glimpsed were running. Second, and even more important, the tip here does not involve the relating of a simple fact observed by the van driver, but rather his statement that the men Captain Sullivan was looking for got on the SEPTA bus. This is simply a conclusion unsupported by any basis, at best an “inchoate and unparticularized suspicion or ‘hunch,’ ” Terry v. Ohio, 392 U.S. 1, 27, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968), insufficient to justify a stop. I would reverse.
I.
With respect to the court’s hot pursuit discussion, there is serious question as to whether the total period of time, from Officer Carolyn obtaining the car to the boarding of the bus, was actually two to three minutes. Captain Sullivan first testified on cross-examination that it took approximately one minute for Officer Carolyn to get the car. In fact, by the time the officers emerged onto 66th Street, so much time had passed that Captain Sullivan testified that he believed he’d lost the suspects. Then the van driver entered the picture, and Captain Sullivan testified further that approximately two to three minutes had passed from the time the car was retrieved to the time he encountered the van driver. That brings the total time from the spotting of the running men to the conversation with the motorist to three to four minutes. While Captain Sullivan testified still later that the total time from obtaining the car to pulling over the bus was two to three minutes, his earlier testimony calls that estimation into doubt. Whatever the exact time period, I conclude there was no hot pursuit in this case, but rather a trail gone cold, if not frigid or even frozen.
I further disagree with the application of any sort of hot pursuit analysis because the hot pursuit exception to the warrant requirement requires police to have probable cause. United States v. Santana, 427 U.S. 38, 96 S.Ct. 2406, 49 L.Ed.2d 300 (1976). In fact, at oral argument the Government explicitly denied that it was asking this court to extend the hot pursuit exception to Terry stops based on reasonable suspicion. Accordingly, because Captain Sullivan relied on an unknown and unidentified van driver’s tip in stopping the bus, we should be primarily guided by cases involving stops based on anonymous tips. But see supra at 169 (“Robertson characterizes the van driver as an ‘anonymous informant.’ We disagree. We view this as essentially a hot pursuit case, aided by a bystander’s informative tip.”).
II.
This court recently considered anonymous tips and the role they play in creating reasonable suspicion in United States v. Roberson, 90 F.3d 75 (3d Cir.1996). The tip in Roberson was that a “heavy-set, black man wearing green pants, a brown leather jacket, and a white hooded sweatshirt was selling drugs on the 2100 block of Chelten Avenue” in Philadelphia. Id. at 79. Despite the fact that officers found a man fitting that description when they arrived at the scene, this court concluded that reasonable suspicion was lacking for a stop because “the police had no basis for assessing either the reliability of the informant or the grounds on which the informant believed that a crime was being com*173mitted.” Id. at 80 (emphasis added). In his opinion for the court, Judge (now Chief Judge) Becker set out the principles used in analyzing whether such tips are a sufficient basis for police action. Id. at 77.
Judge Becker first discussed the two-factor Aguilar/Spinelli test, formerly used to evaluate an informant’s tip in the probable cause context. Id. (citing Spinelli v. United States, 393 U.S. 410, 89 S.Ct. 584, 21 L.Ed.2d 637 (1969); Aguilar v. Texas, 378 U.S. 108, 84 S.Ct. 1509, 12 L.Ed.2d 723 (1964)). That test required that an informant’s tip, to be credible, “had to indicate both the basis for the informant’s knowledge and facts sufficient to establish its veracity or reliability.” Id. As Judge Becker pointed out, the Supreme Court later abandoned this two-part test in favor of a totality of the circumstances approach under which the Aguilar/Spinelli factors are “better understood as relevant considerations in the totality-of-the-circumstances analysis that traditionally has guided probable cause determinations: a deficiency in one may be compensated for, in determining the overall reliability of a tip, by a strong showing as to the other, or by some other indicia of reliability.” Id. (quoting Illinois v. Gates, 462 U.S. 213, 233, 103 S.Ct. 2317, 76 L.Ed.2d 527 (1983)). This totality of circumstances test was extended to the reasonable suspicion context in Alabama v. White, 496 U.S. 325, 110 S.Ct. 2412, 110 L.Ed.2d 301 (1990).
In White, the Court upheld a Terry stop based upon an anonymous informant’s telephone tip. 496 U.S. at 332, 110 S.Ct. 2412. The informant in that case alleged that a certain woman was in possession of illegal drugs, and predicted that this woman would leave a particular apartment, at a particular time, in a particular car, and then drive to a particular location. Id. at 327, 110 S.Ct. 2412. Most of the tip was corroborated by police before they stopped the woman. Id. Despite the plethora of descriptive and predictive information contained in the tip (as opposed to the case before us here), the Court recognized that White was “a close case,” id. at 332, and only concluded that there was reasonable suspicion because “police observation showed that the informant had accurately predicted the woman’s movements.” Florida v. J.L., 529 U.S. 266, 270, 120 S.Ct. 1375, 146 L.Ed.2d 254 (2000).
Most recently, in J.L., officers received an anonymous phone call alleging that a young African-American male wearing a plaid shirt was standing at a particular bus stop and carrying a gun. 529 U.S. at 268, 120 S.Ct. 1375. Even though the police officers who investigated the tip did indeed find a man fitting that description at the bus stop, the Court held that the call did not provide them with reasonable suspicion to stop the man, because the tipster in that case had “neither explained how he knew about the gun nor supplied any basis for believing he had inside information.” Id. at 271,120 S.Ct. 1375.
As Judge Becker observed in Roberson, while WMe stressed the importance of the officer’s ability to corroborate significant aspects of the tip and the tip’s ability to predict future events, it also reiterated that the Aguilar/Spinelli factors “remain ‘highly relevant in determining the reliability of [an informant’s] report.’ ” White, 496 U.S. at 328, 110 S.Ct. 2412 (quoting Gates, 462 U.S. at 230, 103 S.Ct. 2317), cited in Roberson, 90 F.3d at 80; cf. J.L., 529 U.S. at 271, 120 S.Ct. 1375 (holding tip insufficient because it’“left the police without means to test the informant’s knowledge or credibility”). Here, the corroboration of descriptive or predictive information generally relied upon in finding reasonable suspicion in the anonymous tip context is lacking. More disturbing, there is nothing in the tip to establish the infor*174mant’s basis of knowledge. This is critical because without something to demonstrate that the van driver was not only honest but also “well informed,” White, 496 U.S. at 332, 110 S.Ct. 2412, Sullivan is left acting purely upon the van driver’s impermissible “hunch,” Terry, 392 U.S. at 27, 88 S.Ct. 1868.
I examine the factual issues separately, for the sake of clarity. This does not accord them “independent status,” but rather facilitates application of the totality of circumstances test, recognizing, as stated above, that “a deficiency in one [factor] may be compensated for, in determining the overall reliability of a tip, by a strong showing as to the other, or by some other indicia of reliability.” Gates, 462 U.S. at 233, 103 S.Ct. 2317, quoted in Roberson, 90 F.3d at 77.
A.
The lack of anything in the tip to establish a basis of knowledge on the part of the van driver is the most glaring defect in this case. See J.L., 529 U.S. at 271, 120 S.Ct. 1375 (holding that even where police verified the description of the suspect given by the anonymous informant, the fact that the tipster “neither explained how he knew about the gun nor supplied any basis for believing he had inside information” precluded finding of reasonable suspicion); cf. White, 496 U.S. at 332, 110 S.Ct. 2412 (concluding reasonable suspicion was present where “there was reason to believe not only that the caller was honest but also that he was well informed, at least enough to justify the stop”). As discussed above, basis of knowledge continues to be a relevant consideration in the totality of the circumstance analysis. Here, the van driver had at best an “inchoate and unparticu-larized suspicion or ‘hunch,’ ” Terry, 392 U.S. at 27, 88 S.Ct. 1868, that the men he saw get on the bus were the same men Sullivan was looking for. When he imparted this hunch to Sullivan it gained no greater substantive content, and Sullivan did nothing to determine whether the driver had any basis at all for his hunch. Sullivan himself thus had no more than a hunch that there was criminal activity afoot on the bus. This is simply not enough to justify a stop. Id.
Roberson concluded that since the descriptive information provided by the tip in that case was readily observable, all the Government was left with was the fact that the defendant was standing on a “hot” high-crime corner, and “[t]his is not enough.” 90 F.3d at 80. Similarly in this case, all that Captain Sullivan could glean from the informant’s tip was that (1) some men had gotten on a bus and (2) the informant assumed these men were sought by the police. There was nothing in the tip to provide a “reason to believe” that the van driver was “well informed” and had any basis for knowing that the men getting on the bus were those sought by the police.1 White, 496 U.S. at 332, 110 S.Ct. 2412. Nor did Captain Sullivan seek *175any details to establish the basis for this assumption or hunch. To paraphrase Judge Becker in Roberson: I simply cannot accept the Government’s position that any citizen getting on a bus who, without otherwise engendering suspicion, is unlucky enough to be the subject of a non-predictive anonymous tip is subject to a Terry stop simply because a crime has been committed in the area. See 90 F.3d at 80.
The court seeks to distinguish Roberson. Supra at 169-70. It argues that this case is different because Captain Sullivan already knew a crime had been committed and the van driver merely pointed out where the men Captain Sullivan was looking for had gone. This but accentuates the lack of any basis for the van driver’s knowledge of who Captain Sullivan was looking for. To establish the basis for this knowledge, the court itself must add to the tip the words “pursued,” supra at 168-69, and “having observed the fleeing suspects,” supra at 170, to support the bald, conclusory statement, “The guys you are looking for, they just got on that bus.” Had the tip contained such additional language, we would have a far different case. But this language is not in the record — the van driver did not mention fleeing or pursued suspects. This addition of language is a vital step in the court’s decision, as it supplies for itself the support for the tip which is lacking in the record. The van driver stated only that the men Captain Sullivan was looking for had gotten on the bus. He gave no information to support his raw conclusion or assumption that the men he saw were the “suspects” that Sullivan was looking for. He said nothing to demonstrate how he reached this conclusion, and the court errs today in supplying it for him. The facts in the record before us are not sufficient to justify the court’s conclusion.2
B.
In order for an anonymous tip to supply reasonable suspicion it must bear indicia of reliability. J.L., 529 U.S. at 274, 120 S.Ct. 1375. The Supreme Court has looked for these indicia of reliability in the form of predictive or descriptive details. White, 496 U.S. at 332, 110 S.Ct. 2412. In Roberson, this court recognized the importance of the corroboration of details and the value of predictive information, pointing out that in both Gates and White predictive and descriptive information was corroborated by .police investigation prior to any seizure. 90 F.3d at 77, 79. Judge Becker concluded his analysis with an observation relevant to this case: “The tip in the case at bar contained no ‘details of future actions of third parties ordinarily not easily predicted.’ Thus, no future actions could be corroborated, and an important basis for forming reasonable suspicion was absent.” Id. at 80 (internal citations omitted) (quoting White, 496 U.S. at 332, 110 S.Ct. 2412). Here, too, the tip con*176tained no details of future actions. Rather, the prediction made — that the men the officers were looking for would be found on the bus — consisted entirely of an assumption: that the men on the bus were the ones the officers were looking for. The van driver gave no basis for this assumption, which was the key reason the officers acted.3 Indeed, whom the officers were looking for is something that only the officers could know.
The Government argues that the lack of descriptive or predictive information here is not fatal given the totality of the circumstances.4 It contends the tip merely served to fill in the last part of the puzzle: Where did the two suspects go? While there may be cases that support this argument in a general sense, the Government cites none where officers relied, even in part, upon a tip so lacking in detail as the one here. For example, in United States v. Allen, the Tenth Circuit concluded that an anonymous tip was sufficiently corroborated “[f]or the limited purpose of establishing identity.” 235 F.3d 482, 488 (10th Cir.2000), cert. denied, 532 U.S. 989, 121 S.Ct. 1643, 149 L.Ed.2d 501 (2001). However, the informant in that case identified the suspect by name, and described the suspect’s physical stature and car. Id. at 485. The tip here is utterly lacking in detail, and we should “refus[e] to stretch Alabama v. White any further.” Roberson, 90 F.3d at 80.
The Government also cites this court’s recent decision in United States v. Valentine, 232 F.3d 350 (3d Cir.2000), cert. denied, 532 U.S. 1014, 121 S.Ct. 1748, 149 L.Ed.2d 670 (2001), to argue that the tip in this case is not truly “anonymous” as that term is used in cases such as J.L. and White, but rather deserves additional credit because it was a face-to-face report. In Valentine this court concluded that “officers had reasonable suspicion after they received [a] face-to-face tip, were in a high-crime area at 1:00 a.m., and saw Valentine and his two companions walk away as soon as they noticed the police’ car.” Id. at 357. J.L. was distinguished on the basis of the tip being face-to-face and because the officers “knew that the informant was reporting what he had observed moments ago.” Id. at 354.
While it is true that “a tip given face to face is more reliable than an anonymous *177phone call,” id. at 354 (citing cases), that proposition does not stretch so far as to provide reasonable suspicion here. In White, the Supreme Court stated that as a tip becomes more reliable, less information will be required to establish reasonable suspicion — it did not say that no information would be required. 496 U.S. at 330, 110 S.Ct. 2412. The tipster in Valentine in fact provided a detailed description which the police matched before making the stop. 232 F.3d at 352-53. Furthermore, Valentine recognized that the reliability inferred from a face-to-face tip was not sufficient in and of itself: “The reliability of a tip, of course, is not all that we must consider in evaluating reasonable suspicion; the content of the tip must also be taken into account....” Id. at 355. Here, the content of the van driver’s tip contained nothing to support the speculative conclusion or assumption that the men he saw get on the bus were the ones Sullivan was chasing.
III.
While this court said in Valentine that “we are not going to second-guess the officers’ decision to pursue the suspect immediately,” 232 F.3d at 355, the officer in that case actually stopped the defendant only after he had received and corroborated descriptive information, id. at 352-53. Sullivan did not attempt to satisfy Terry with a quick question such as: “Were they running?” or “What were they wearing?” If the officers in Valentine had time to verify the tip in that case even though an armed gunman was involved, Sullivan could reasonably be expected to do the same here.5 A thorough analysis of the totality of the circumstances, as required by White and Gates, and as specifically focused by this court in Roberson, requires the conclusion that Terry was not satisfied.
I would reverse.
. While it may be argued that sufficient basis of knowledge should be inferred from the fact that the van driver was in the area and pointing to a bus the suspects could conceivably have reached, we should decline to make such an inference absent something more in the content of the tip. Cf. 4 Wayne LaFave, Search and Seizure, § 9.4(g) at 211 ("[A]n unexplained assertion by tire police (or, to the police by another) that a person looks suspicious is not entitled to weight.”) (footnote omitted) (3d ed. 1996). This is not the same thing as ‘‘categoriz[ing those] factors ... as simply out of bounds in deciding whether there was ‘reasonable suspicion’ for the stop.” United States v. Arvizu, 534 U.S. 266, 122 S.Ct. 744, 748, 151 L.Ed.2d 740 (2002). Rather, I simply conclude that the totality of circumstances here does not give rise to reasonable suspicion in light of the binding anonymous informant precedent set out in this dissent.
. The court also attempts to distinguish Roberson because "[h]ere, Captain Sullivan, in hot pursuit, did not have time to ask for details without risking the suspects’ disappearance.” Id. Suffice it to say that this argument rests on the proposition that a public bus may disappear in the time it takes to ask: "Were they running?” or "Was one of them wearing red pants?” Finally, the court suggests this case is different because the suspects here were alleged to be armed. There is no firearm exception to Terry, however. See Florida v. J.L., 529 U.S. 266, 272, 120 S.Ct. 1375, 146 L.Ed.2d 254 (2000) ("Our decisions recognize the serious threat that armed criminals pose to public safety; Terry’s rule, which permits protective police searches on the basis of reasonable suspicion rather than demanding that officers meet the higher standard of probable cause, responds to this very concern.”). Of most significance is that the relevance of Roberson is the analytical framework it sets out, not any particular set of facts it does or does not encompass.
. The court states as fact that the van driver said he saw "two guys.” Supra at 4. However, the record, which on this point consists entirely of the testimony of Sullivan, is not so clear. See App. for Appellant at 33a ("[H]e stated to me: 'Officer, those guys you’re looking for just got on that bus up there.” ’); id. at 52a ("[H]e says 'Officer, them two guys you're looking for just got on that bus’ .... ”); id. ("[He] say[s]: 'Officer, the guys you're looking for, they just got on that bus.' "); see also United States v. Robertson, 81 F.Supp.2d 579, 580 (E.D.Pa.2000) ("[A] motorist came by and informed [Sullivan] that the men he was looking for had boarded a SEPTA bus....”).
. It is important to remember that the circumstances in this case include two men running toward the scene of the crime, at midday, in metropolitan Philadelphia, cf. 4 Wayne R. LaFave, Search and Seizure, § 9.4(g) at 206 (3d ed. 1996) ("[L]ess will suffice in the early morning hours when few persons are about than would be a basis for a stopping at high noon.”), as well as an informant whose identity is unknown, cf. J.L., 529 U.S. at 270, 120 S.Ct. 1375 (distinguishing anonymous tip from one where tipster "can be held responsible if her allegations turn out to be fabricated”); Roberson, 90 F.3d at 79 n. 3 (distinguishing descriptions given by anonymous informants from those given by “identified witnesses”). See generally United States v. Jones, 998 F.2d 883, 885-86 (10th Cir.1993) (concluding reasonable suspicion was not present in vehicle stop case where suspects were "in Albuquerque, a major population center, at 4:00 p.m. on a weekday,” and the "information that the police were acting on came from an informant with whom they had no experience”).
. The fact that Sullivan was confronted with the possibility of an armed criminal on a passenger bus does not change this conclusion. It is certainly not obvious that the best reaction to such a scenario is to pull over the bus with sirens blaring and lights flashing. Cf. J.L., 529 U.S. at 273, 120 S.Ct. 1375 ("The facts of this case do not require us to speculate about the circumstances under which the danger alleged in an anonymous tip might be so great as to justify a search even without a showing of reliability.”).