TORRUELLA, Circuit Judge.
This is the second appeal concerning a Terry stop which resulted in the entry of a conditional guilty plea by appellant Gregory Wright for being a felon in possession of a firearm, 18 U.S.C. § 922(g)(1). In the Terry stop at issue, Boston police officers recovered a gun from Wright’s sweatshirt pocket after Wright was observed leaning forward from the backseat of a car to identify the officers, quickly exiting and running from the car, clutching at the right side of his sweatshirt while running, and ignoring the officers’ order to stop. Below, Wright moved to suppress the gun, but the district court denied the motion, concluding that the officers had reasonable suspicion to stop him. In Wright’s first appeal, we ruled that the district court’s denial of the suppression motion was tainted with legal error, and remanded for further proceedings. United States v. Wright, 485 F.3d 45, 54 (1st Cir.2007). On remand, the district court again denied the motion, and Wright renewed his appeal. This time, after a careful review of the record, we affirm.
I. Background
A. The Prior Proceedings
The district court first denied Wright’s motion to suppress after an evidentiary hearing at which three officers testified about the circumstances leading to his arrest. On appeal, we held that the district court’s conclusion that the officers had reasonable suspicion to stop Wright was based on a legal error. The court interpreted Wright’s running as “flight,” and accepted the police officers’ testimony that they saw Wright clutch his sweatshirt, by linking those findings to the subsequent discovery of the gun. Id. at 48.
We further held that the court’s self-described “backwards” reasoning fatally tainted its factual findings. See id. at 48, 52 (quoting the district court as stating “Can I reason backwards from the fact that what happened next was that the police officers discovered the weapon on Mr. Wright? I think it is undisputed he was carrying a weapon and I do so reason”). We observed that it was “impossible to discern whether the court would have concluded that Wright knowingly fled from the police if it had not considered the eventual recovery of the gun.” Id. at 52. Similarly, we could not evaluate the court’s finding on Wright’s hand movement because the court had used its “commonsense assumption that the gun was heavy ... [to] ma[k]e a factual finding that Wright grabbed his sweatshirt because he was carrying a heavy gun.” Id. at 53. We concluded that the flaw in the underlying factual findings invalidated the court’s legal conclusion that the officers had reasonable suspicion to stop Wright, requiring us to vacate the denial of the suppression motion. Id.
In remanding the case for reconsideration, we also addressed the district court’s discussion of the area where Wright was stopped. The court had stated that it did not conclude that the area was a “high crime area,” a characterization that would have been relevant to the inquiry into whether “the circumstances [were] sufficiently suspicious to warrant further investigation,” Illinois v. Wardlow, 528 U.S. 119, 124, 120 S.Ct. 673, 145 L.Ed.2d 570 (2000); see also id. (“[W]e have previously noted the fact that the stop occurred in a ‘high crime area’ among the relevant contextual considerations in a [reasonable suspicion] analysis.”). Because the district [203]*203court expressed uncertainty about whether the high crime area finding was a legal question or a mixed question of fact and law, we clarified that the character of a stop’s location is a factual issue. Wright, 485 F.3d at 53. We observed that the court might choose to revisit the question upon remand. To assist its possible reevaluation of the issue, we identified a number of relevant factors to be considered. See id. at 53-54.
B. The Remand Proceedings
On remand, the district court solicited supplemental briefing and heard oral argument from the parties, but took no additional evidence. In its ruling from the bench, the court expressly adopted the description of the stop set out in our prior decision, with some modifications:
On the evening of November 8, 2004, a caravan of four unmarked police cars was patrolling in Dorchester, Massachusetts. The cars were Crown Victorias, a model widely associated with police departments. The plainclothes officers in the caravan were members of the Boston Police Department Youth Violence Task Force.
At about 7:45 p.m., the caravan was driving north on Blue Hill Avenue and slowed down as the lead car passed a vehicle that had just pulled over in front of a mini-mart at 1216 Blue Hill Avenue. The parked car was partially blocking one of two driveway entrances to the mini-mart parking lot. Officer Brown, who was seated in the lead car’s front passenger seat, looked to his right as they passed the parked vehicle and observed three people, one of whom he recognized as Omar Edwards, a neighborhood resident. He did not recognize the driver or the passenger seated in the back seat of the parked car.
Immediately after passing this parked vehicle, Officer Brown’s car pulled over to the right parking lane, in front of the parked car. The rest of the caravan came to a stop in the right travel lane to the rear of the parked car. The front passenger of the second police car, Officer Bordley, then observed the back seat passenger of the parked car, later identified as appellant, lean forward as though he was looking at the Crown Victoria that had just pulled over in front of his car. Wright then exited his car, on the passenger side, and began to run southward down Blue Hill Avenue. As he ran, Wright “grabbed toward the front of his sweatshirt in the vicinity of his waist.” 1
Officer Brown quickly exited his car, as did a number of the other officers in the caravan. The police ordered Wright to stop running, but he did not obey this directive. Within a matter of seconds, the officers caught up to Wright, who resisted the officers’ attempts to frisk him.2 The police patted Wright down and recovered a silver pistol from his sweatshirt pocket. He was charged with being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g)(1).
See id.
Based on these facts, Wright argued that the officers lacked reasonable suspi[204]*204cion to stop him. He noted that the officers’ testimony suggested no suspicious motive either for Wright’s leaning forward in the car or for grabbing at his sweatshirt. Instead, Wright argued that the officers’ testimony indicated that they had stopped Wright only because he had run from the car. Wright asserted that running in those circumstances, without more, was insufficient to justify the stop.
The government countered that the totality of the circumstances demonstrated that the officers had reasonable suspicion to stop Wright. The government argued that Wright’s running from the car, which it considered “flight,” along with his “clutching at either the waist or the side of the sweatshirt” were sufficient to justify the stop. The government further contended that a finding that the area in which the stop occurred was a “high crime area” under Wardlow was not necessary. However, the government argued that the district court could take into account “the officers’ testimony regarding what ... their particular knowledge of that area was at that time.”
After hearing argument from both sides, and after setting forth its findings of fact, the district court ruled as follows:
So, there must be an adequate constitutional basis to chase after [Wright] and seize him under the Fourth Amendment, and it is to that issue the Court now turns.
These are mixed issues of law and fact. I conclude that the officers had no right to seize Mr. Wright at the point when their, the vehicles, both the one in front, the police cars in front, and the two police cars behind, came to park in front and back. They had no right to seize him under the Fourth Amendment simply because Officer Bordley saw him lean forward. I do think that it is a reasonable inference from Officer Bordley’s testimony that Officer Bordley thought, though he did not expressly so testify, that Mr. Wright had made [i.e., recognized] the unmarked police car that had parked in front.
I infer from the testimony that, though things happened in split second intervals, Mr. Wright had started to run, I do not at this point say flee, he had started to run before the officers had started to run after him. The Court infers, and again the testimony is what it is, it’s not explicit, I infer that the officers ascribed some significance as they are competent and experienced police officers in that area of Boston to the fact that Mr. Wright grabbed his side, clutched at something in his sweatshirt. The Court infers that the officers did in that split second draw the inference that he might well possess a weapon.
Having drawn that inference, that is sufficient under all the circumstances to order Mr. Wright to stop, and when he did not stop, t[o] chase after him and seize him, the actual seizure is when the officers came in contact with Mr. Wright.
The court also revisited whether the Terry stop occurred in a high crime area:
It’s necessary I think to go a little further in honor, out of respect to the reasoning of the Court of Appeals, because I think this is a very close case. I do not find that this was a high crime area as the language is used in Ward-low. The first time round I expressed some unease with that definition. And my unease continues. I don’t make that finding. I honor what I infer is the police officers’ experience and knowledge of the communities which they patrol .... [T]he problem I have with that Wardlow type of analysis, candidly, is that the rights of the citizens in the Dorchester area of Boston have to be identical to the rights of citizens in Milton and Wellesley. They have to be scrupulously identical.
[205]*205So, while I treat all the circumstances taken as a whole and I draw inculpatory inferences ... from what I infer was the observations of the officers and their response, their immediate order to Mr. Wright to stop, I rule that that would have been enough if the officers had acted in the same fashion in Wellesley or Milton, not because of this particular area along Blue Hill Avenue is a high crime area. I think I’ve said all that needs to be said.
Wright renewed his appeal.
II. Discussion
On appeal, Wright again contends that the district court erred in concluding, based on the totality of the circumstances, that the officers in this case had reasonable suspicion to stop him.
A. The Legal Framework
Under Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968), and its progeny, a police officer may briefly detain an individual for questioning if the officer has “reasonable suspicion to believe that criminal activity ‘may be afoot.’ ” United States v. Arvizu, 534 U.S. 266, 273, 122 S.Ct. 744, 151 L.Ed.2d 740 (2002) (quoting United States v. Sokolow, 490 U.S. 1, 7, 109 S.Ct. 1581, 104 L.Ed.2d 1 (1989)); see also Terry, 392 U.S. at 30, 88 S.Ct. 1868 (permitting such a stop when “a police officer observes unusual conduct which leads him reasonably to conclude in light of his experience that criminal activity may be afoot”). “While no perfectly precise definition of reasonable suspicion exists, it is well established that, in terms of the continuum of knowledge, reasonable suspicion requires more than a mere hunch but less than probable cause.” United States v. Ruidíaz, 529 F.3d 25, 29 (1st Cir.2008).
Reasonable suspicion requires “ ‘a particularized and objective basis’ for suspecting the person stopped of criminal activity.” Ornelas v. United States, 517 U.S. 690, 696, 116 S.Ct. 1657, 134 L.Ed.2d 911 (1996) (quoting United States v. Cortez, 449 U.S. 411, 417-18, 101 S.Ct. 690, 66 L.Ed.2d 621 (1981)). “Th[e] particularity requirement means, in effect, that such a finding must be ‘grounded in specific and articulable facts.’ ” United States v. Espinoza, 490 F.3d 41, 47 (1st Cir.2007) (quoting United States v. Hensley, 469 U.S. 221, 229, 105 S.Ct. 675, 83 L.Ed.2d 604 (1985)). The “objective” component requires courts to “focus not on what the officer himself believed but, rather, on what a reasonable officer in his position would have thought.” Id.
We review the district court’s findings of historical fact, as well as inferences draw from those facts, for clear error, which exists when we are left with a “ ‘definite and firm conviction that a mistake has been committed.’ ” See id. at 46 & n. 2 (quoting United States v. U.S. Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 92 L.Ed. 746 (1948)). We review the district court’s ultimate reasonable suspicion determination de novo, Ornelas, 517 U.S. at 699, 116 S.Ct. 1657, although the factual component of this “mixed question of law and fact” remains subject to clear error review. Id. at 696, 116 S.Ct. 1657. In reviewing this determination, we are instructed by the Supreme Court to view the “totality of the circumstances,” and not engage in a “divide-and-conquer analysis” whereby we determine whether each of the facts supporting reasonable suspicion are “susceptible to an innocent explanation.” Arvizu, 534 U.S. at 274, 122 S.Ct. 744. Thus, “[a] determination that reasonable suspicion exists ... need not rule out the possibility of innocent conduct.” Id. at 277, 122 S.Ct. 744.
Finally, when a district court has denied a motion to suppress, “ ‘we will [206]*206uphold a denial of a motion to suppress if any reasonable view of the evidence supports it.’ ” United States v. Coccia, 446 F.3d 233, 237 (1st Cir.2006) (quoting United States v. Garner, 338 F.3d 78, 80 (1st Cir.2003)).
B. Application of the Framework
1. The Historical Facts
The district court adopted the recitation of the facts contained in our prior opinion, and thus found the following sequence of events that resulted in the instant Terry stop, immediately after the car in which Wright was sitting “had just pulled over in front of a mini-mart at 1216 Blue Hill Avenue .... partially blocking one of two driveway entrances to the mini-mart parking lot:”
(a) Wright “lean[ed] forward” from the backseat of his car and then “exited his car on the passenger side”;
(b) Wright then “began to run southward down Blue Hill Avenue”;
(c) “As he ran, Wright put one hand on the right side of his sweatshirt grabbing or holding onto the sweatshirt pocket”; and
(d) “The police ordered Wright to stop running but he did not obey this directive.”
The parties largely do not dispute these facts on appeal.3 As to the last finding above, Wright does not dispute that the stop did not occur until Wright was physically restrained after the officers’ order to stop. See California v. Hodari D., 499 U.S. 621, 626, 111 S.Ct. 1547, 113 L.Ed.2d 690 (1991) (holding that a directive to stop is not itself a seizure). Accordingly, we consider Wright’s refusal to stop as a historical fact in determining whether the officers had reasonable suspicion to stop Wright.
Of significance to this appeal, the district court did not find that the area where the events took place was a high-crime area as defined under Wardlow. Although the officers testified at length that the area was a high-crime area,4 the government does not press this position on appeal, nor does it claim that the district court committed clear error with respect to this finding.
On appeal, Wright challenges two inferences made by the district court based upon these historical facts. First, the district court inferred that Officer Bordley, who observed Wright leaning forward, “thought, though he did not expressly so testify, that Mr. Wright had made [i.e., recognized] the unmarked police car that [207]*207had parked in front.” Second, the district court inferred that the officers inferred from Wright’s clutching of the side of his sweatshirt that the officers “ascribed some significance” to that clutching and that “in that split second dr[e]w the inference that he might well possess a weapon.”
Wright argues that the district court’s inferences concerning the officer’s subjective inferences are irrelevant. Specifically, he argues that the proper focus is “not on what the officer himself believes but, rather, on what a reasonable officer in his position would have thought.” Espinoza, 490 F.3d at 47; cf. Whren v. United States, 517 U.S. 806, 814, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996) (“[T]he Fourth Amendment’s concern with ‘reasonableness’ allows certain actions to be taken in certain circumstances, whatever the subjective intent.”) (emphasis in original).
We agree that the proper focus is an objective one, but we disagree that the inferences made by police officers are irrelevant in all instances. The Supreme Court has instructed courts to afford “due weight” to the inferences made by police officers based on their “experience and expertise.” See Ornelas, 517 U.S. at 699, 116 S.Ct. 1657. In Ornelas, for example, an officer testified that “over the past nine years [he] had searched approximately 2,000 cars for narcotics,” and, upon searching the car in that ease, noticed a loose panel that he inferred “might have been removed and contraband hidden inside.” Id. at 693, 116 S.Ct. 1657. The Court stated:
To a layman the sort of loose panel below the back seat armrest in the automobile involved in this case may suggest only wear and tear, but to [the officer], who had searched roughly 2,000 ears for narcotics, it suggested that drugs may be secreted inside the panel. An appeals court should give due weight to a trial court’s finding that the officer was credible and the inference was reasonable.
Id. at 700, 116 S.Ct. 1657.
Here, however, and as the district court admitted, the officers did not testify to the inferences they purportedly made. In the district court’s words, the officers “did not expressly so testify” as to their inferences and “the testimony is what it is.” Thus, unlike in Ornelas, the district court did not make credibility determinations as to the officers’ stated inferences. Nor did the officers testify as to how their “experience and expertise” provided support for the inferences they made. Instead, the district court speculated as to what those inferences were, and, with respect to the second challenged inference, even further speculated that the inferences were reasonable because the officers were “competent and experienced police officers in that area of Boston.” We cannot condone such speculation and agree with Wright that, to the extent that the district court sought to supplement the historical facts with such inferences, this was improper. Cf. United States v. Spinney, 65 F.3d 231, 234 (1st Cir.1995) (noting, in the context of a sufficiency of an evidence challenge, that “[t]he appellate function, properly understood, requires the reviewing court to take a hard look at the record and to reject those evidentiary interpretations and illations that are unreasonable, insupportable, or overly speculative”).
However, we agree with the government that the district court’s inferences with respect to the officers’ inferences are better understood as implicit conclusions as to what a “reasonable officer ... would have thought.” Espinoza, 490 F.3d at 47. The record strongly suggests that this was the district court’s intention. After recounting its findings concerning the events leading up the stop, the district court switched gears and stated “there must be an adequate constitutional basis to chase after [208]*208[Wright] and seize him under the Fourth Amendment, and it is to that issue that the Court now turns.” (Emphasis added). Thus, the district court indicated that it was switching from its factual findings to its ultimate reasonable suspicion analysis. This is confirmed by the district court’s following statement that “[t]hese are mixed issues of fact and law,” which is not true of the district court’s findings of historical fact, but is true of its ultimate determination whether reasonable suspicion exists in this case. See Ornelas, 517 U.S. at 696, 116 S.Ct. 1657 (noting that “[t]he first part of the analysis involves only a determination of historical facts, but the second is a mixed question of law and fact”).
Construing the district court’s inferences as inferences that a reasonable officer would have made does not end the matter. To the extent that the district court’s inferences are factual inferences drawn from the historical facts, they are subject to clear error review, “mindful throughout that when two or more legitimate interpretations of the evidence exist, the factfinder’s choice between them cannot be deemed clearly erroneous.” Espinoza, 490 F.3d at 46. To the extent that the district court’s inferences are ultimate conclusions as to what a reasonable officer would have concluded based on the historical facts, and thus conclusions as to whether reasonable suspicion exists, those conclusions are subject to our de novo review.
Having reviewed the historical facts, we turn to whether those facts support reasonable suspicion in this case.
2. The Reasonable Suspicion Analysis
In its analysis, the district court reviewed the historical facts in sequence, building upon each prior event in determining whether a reasonable officer would have reasonable suspicion to stop Wright. For example, the district court noted that, upon observing Wright leaning forward, that, at that point in time, the officers “had no right to seize [Wright] under the Fourth Amendment.” Only later, after subsequent events, did the district court rule that it “was constitutionally appropriate” to stop Wright.
We take the same approach. We review the sequence of events, and any reasonable inferences drawn from those events, to determine whether the officers had reasonable suspicion to stop Wright. After examining this sequence and the totality of the circumstances, we conclude that they had such reasonable suspicion.
a. Leaning Forward
We begin with the officers’ testimony that they observed Wright “lean forward” from the backseat.5 Although [209]*209Wright does not dispute that he leaned forward when the lead ear passed and parked in front of the car in which he was sitting, the district court made a further finding as to what a reasonable officer would have inferred from that action:
I do think that it is a reasonable inference from Officer Bordley’s testimony that Officer Bordley thought ... that Mr. Wright had made the unmarked police car that had parked in front.
Again, we note that although the district court attributed this inference to Officer Bordley, this inference is better understood as one a reasonable officer would have made.
Implicit in the district court’s conclusion is that a reasonable officer would have inferred from Wright’s leaning forward that Wright attempted to identify the lead car. Officer Bordley, in fact, expressly testified that Wright leaned forward “to observe the unmarked motor vehicle that had pulled over.” Since this inference is “plausible ... based on the raw facts as supportably found,” Espinoza, 490 F.3d at 48, we consider it in our analysis.
But the district court went even further and concluded that a reasonable officer would have inferred that Wright “made,” that is recognized, “the unmarked police ear that had parked in front.” We only credit such a factual inference to the extent that it has some basis in the record or upon the “background facts” known by the district court. See Ornelas, 517 U.S. at 699, 116 S.Ct. 1657.
We conclude that the evidence supports a reasonable inference that Wright “made” the lead car as a police car. In United States v. Aitoro, for example, we concluded that it was reasonable for the officers to stop two suspects who, upon seeing them in plainclothes, exclaimed “Oh, shit” prior to fleeing. 446 F.3d 246, 252-53 (1st Cir.2006). Granted, Wright did not state “Oh, shit” or any other exclamation that was indicative that he recognized the officers. However, Wright was observed leaning forward in an attempt to identify the unmarked cars, and, shortly thereafter, quickly exiting from the car. While less colorful than the expression “Oh, shit,” the timing of Wright’s conduct in quickly exiting the car after leaning forward would permit a reasonable officer to infer that Wright “made” the lead car as a police car.
On appeal, Wright contends that an inference that Wright “made” the lead car as a police car is unreasonable because the police cars in this case were unmarked. We disagree. In Aitoro, the officers were in plain clothes, and we did not require evidence of the recognizability of the officers by the suspect to support an inference that the suspect recognized the officers in fleeing.6 In fact, in response to the sus[210]*210pect’s argument that it was “dubious” to conclude that he recognized the officers given that they “were not in uniform and were standing perhaps 80 feet from the corner when Aitoro and Williams came around the bend,” we stated:
The reasonableness of a [stop] entails an objective inquiry into the [stop] from the perspective of the ... officers, however, so what is relevant is not whether Aitoro actually perceived the officers as police officers, but whether the officers reacted reasonably on seeing him flee.
Id. at 253 (emphasis added). Thus, like in Aitoro, where the suspect’s exclamation of “Oh, shit” permitted the officers to infer that they were recognized, here Wright’s actions in leaning forward and shortly thereafter exiting the car are also sufficient to permit a reasonable officer to “react[ ] reasonably” and infer that Wright “made” the unmarked car as a police car.
b. Running Southward
The parties do not dispute that Wright, shortly after leaning forward, exited the car in which he was sitting and started running southward, away from the lead car. The district court, in describing the running, stated that:
I infer from the testimony that, though things happened in split second intervals, Mr. Wright had started to run, I do not at this point say flee, he had started to run before the officers had started to run after him.
On appeal, Wright seizes on the district court’s use of the term “flee,” and argues that the district court made an express finding that Wright’s running southward does not constitute “flight.” We disagree, because the district court found all of the factual components necessary to support a finding of flight. We thus conclude that the district court’s statements cannot be reasonably read as a finding that Wright’s running did not constitute flight.7
Although not strictly defined, the Supreme Court in Wardlow described “flight” as the “unprovoked” running “upon noticing the police.” 528 U.S. at 124, 120 S.Ct. 673 (noting that “unprovoked flight upon noticing the police” contributed to the officers’ reasonable suspicion in that case, and that “[h]eadlong flight — wherever it occurs — is the consummate act of evasion”). As we have previously concluded, the district court reasonably inferred that Wright “made” the lead car. Moreover, and as clumsily noted by the district court above, Wright “had started to run before the officers had started to run after him,” which is supported by the officers’ testimony. Accordingly, Wright’s running was “unprovoked ... upon noting the police.”
Wright argues on appeal that his running could not be interpreted as flight because, in exiting the car, he ran in the direction of the other unmarked police cars, which were behind the car in which Wright was sitting. While true, this fact does not defeat a finding of flight. As an initial matter, and as pointed out by the government, the unmarked police cars were not directly behind the car in which Wright was sitting. Although the testimo[211]*211ny showed that Wright’s car was in the parking lane, the other police cars were, as testified to by Officer Bordley, “stopped in the flow of traffic,” just to the right of the parking lane. Moreover, Wright’s recognition of the police cars after leaning forward was only as to the lead car. Thus, in identifying the lead car as a police car, Wright, like the suspects in Aitoro, “did an abrupt about-face and sprinted in the reverse direction.” 446 F.3d at 249 (internal quotation marks omitted). As events occurred “in split second intervals,” Wright did not have a great deal of time to identify the unmarked cars that were roughly ahead of him and change course accordingly-
c. Clutching
The parties do not dispute that, while running, Wright made a clutching or grabbing motion to his sweater. However, the district court included a further inference as to what a reasonable officer would have inferred from that clutching:
I infer that the officers ascribed some significance as they are competent and experienced police officers in that area of Boston to the fact that Mr. Wright grabbed his side, clutched at something in his sweatshirt. The Court infers that the officers did in that split second draw the inference that he might well possess a weapon.
As with the district court’s earlier inference concerning whether Wright “made” the car, we construe the Court’s “infer[enee]” as one that a reasonable officer would have made.8
The evidence certainly supports a reasonable inference that Wright clutched at contraband. Wright made his clutching motion while in flight, which supports a reasonable inference that, whatever he was trying to secure, he wanted to keep it from the officers. In fact, all of the officers testified to the closeness of the clutching motion to his running away from the lead car. Officer Bordley testified that Wright “grabbed his sweater, took off running” and that, shortly thereafter, “[h]e grabbed his sweater on the right side where the right pocket was.” Officer Celester also testified that, shortly after Wright “jumped out and ran,” he “appeared to pull something out of his waist area” and that “[h]e was tugging with his right hand” around his “waist area.” Finally, Officer Brown testified that he observed “the rear passenger exit the vehicle, stand up, and ... he immediately turned to his right and he grabbed onto like his sweatshirt pocket,” and further demonstrated where he grabbed the pocket, stating to the district court that he “grabbed onto his hood sweatshirt pocket right around here and began to run up Blue Hill Avenue.”
On appeal, the government “does not believe the inference that Wright might possess a gun is essential to reasonable suspicion,” but nonetheless argues that the evidence supports such an inference. We agree that the inference that Wright “might well possess a gun” is not essential, as the evidence is sufficient to support an inference that Wright was clutching at contraband, but disagree that the evidence supports such an inference. The government first argues that the district court found on remand that the object in Wright’s pocket was “heavy,” but the weight of the contents of Wright’s pocket [212]*212does not support an inference that the pocket contains a weapon, as opposed to some other object. Moreover, the government argues that “the court could also have considered that guns are with some frequency carried in the pockets or waist area and police know this.” See, e.g., Aitoro, 446 F.3d at 249 (noting that the officer there “saw Aitoro grab at the waist of his pants, where [the officer] saw a bulge that he thought was a gun”); United States v. Woodrum, 202 F.3d 1, 5 (1st Cir.2000) (“When he freed his right hand and exited the taxi, a gun fell out of his jacket.”); United States v. Alston, 112 F.3d 32, 33 (1st Cir.1997) (“Realizing that there was a gun in the pocket, the officer removed it and arrested Alston.”). But, again, the fact that guns are frequently found in pockets (where else would they be?) does not provide support for an inference that Wright was carrying a weapon as opposed to something else that is frequently kept in pockets, such as keys. Finally, the government points to the presence of Omar Edwards to support an inference that Wright was armed; however, we have already concluded that his presence does not support a suspicion that Wright was engaged in similar criminal activity. Thus, the fact that Wright clutched at his pocket, even while in flight, cannot support an inference that the object he clutched was specifically a weapon, and it was clear error for the district court to so infer. See United States v. McKoy, 428 F.3d 38, 41 (1st Cir.2005) (“It is simply not reasonable to infer that a driver is armed and dangerous because the officers believe that he appears nervous and reaches toward the car’s console when approached by police, even in a high-crime neighborhood”).
d. Refusal to Stop
Unlike the other facts so far discussed, Wright’s refusal to obey the officers’ order to stop, without a doubt, contributes to a reasonable suspicion. Thus, as we have previously held, the failure to heed an officer’s order, while not conclusive of reasonable suspicion itself, is supportive of it. See, e.g., United States v. Soares, 521 F.3d 117, 121 (1st Cir.2008) (holding that a defendant’s refusal to follow police orders to keep still and keep hands visible contributed to reasonable suspicion that suspect was armed and dangerous). Other courts have so held. See United States v. Lenoir, 318 F.3d 725, 729 (7th Cir.2003) (“A suspect’s failure to halt upon police command to do so,” along with other factors, “support a finding of reasonable suspicion.”); United States v. Johnson, 212 F.3d 1313, 1316-17 (D.C.Cir.2000) (fact that suspect did not comply with order to put hands up but continued to make motions consistent with hiding or retrieving something contributed to finding of reasonable suspicion); United States v. Valentine, 232 F.3d 350, 358-59 (3d Cir.2000) (citing cases finding failure to comply with police orders supported reasonable suspicion).
e. The Totality of the Circumstances
We have noted that, in conducting a reasonable suspicion analysis, “a fact that is innocuous in itself may in combination with other innocuous facts take on added significance.” Ruidíaz, 529 F.3d at 30. Accordingly, in our prior decisions we have upheld Terry stops where the combination of “innocuous” facts culminates in reasonable suspicion. See, e.g., id. at 30-32, 88 S.Ct. 1868 (combination of reliable tip from a 911 caller plus belligerence of suspect resulted in reasonable suspicion); Soares, 521 F.3d at 120-21 (combination of time of night, high-crime area, and unusual behavior resulted in reasonable suspicion); United States v. Romain, 393 F.3d 63, 72 (1st Cir.2004) (combination of 911 call, suspect’s visible agitation, and suspect’s belligerence resulted in reasonable suspicion).
[213]*213Here, reasonable suspicion arises not just from the combination of facts, but from their progression. We have previously held that reasonable suspicion can be based on “unfolding events,” with suspicion accumulating as more innocent interpretations of the historical facts fall by the wayside. United States v. Sowers, 136 F.3d 24, 27 (1st Cir.1998) (“Based on unfolding events, the trooper’s attention (and, thus, his reasonable suspicions) shifted away from the equipment violations that prompted the initial stop toward a belief that the detainees were engaged in more serious skulduggery. Such a shift in focus is neither unusual nor impermissible.”); see also Soares, 521 F.3d at 120 (“Several additional facts became known as the stop progressed, which, taken together, created reasonable suspicion that Soares might be armed and dangerous.”).
So it is here. In this case, the lead police car was first drawn to the car in which Wright was sitting when it parked and partially blocked an entrance to a mini-mart. In short order, Wright leaned forward, quickly exited from the car as if recognizing the lead car as a police car, ran in the opposite direction of the lead car, clutched at his side, and, crucially, refused to stop when ordered to do so. As is clear from the discussion above, each event recast each previous event in a different light, such that, by the time that Wright refused to stop, there was sufficient suspicion to question why he had leaned forward, exited the car, clutched, and started running.
We conclude by emphasizing that the issue here is not whether Wright’s conduct could lead a reasonable officer to conclude that it was certain that he was engaged in criminal activity, or even that it was more probable than not that he was engaged in such activity. Rather, the Supreme Court has stressed that a Terry stop is permitted even if “the conduct justifying the stop was ambiguous and susceptible of an innocent explanation.” Wardlow, 528 U.S. at 125, 120 S.Ct. 673. In fact, “the very purpose of [Terry ] stops is to clarify ambiguous situations.” 2 La-Fave et al., Criminal Procedure § 3.8(d), at 327 (3d ed.2007). We conclude, based upon our review of the record, that Wright’s actions were sufficiently ambiguous as to whether there was criminal activity afoot to justify the stop in this case.
Accordingly, we conclude that the totality of the circumstances supports the district court’s conclusion that the officers had reasonable suspicion to stop Wright.
III. Conclusion
For the foregoing reasons, the denial of Wright’s motion to suppress is affirmed.
Affirmed.