dissenting.
Recognizing this as a close case, I respectfully dissent. I would affirm the district court’s suppression order on the ground that Smith was seized at least as of the point when, boxed in by the officers, he was asked for his identification. See United States v. Smith, 332 F.Supp.2d 277, 283-84 (D.Mass.2004). My difference with the majority is over application of the appellate standard of review, articulated in Ornelas v. United States, 517 U.S. 690, 116 S.Ct. 1657, 134 L.Ed.2d 911 (1996).
The government does not challenge the district court’s findings of historic fact. Nor, on appeal, does it challenge the court’s ruling that the officers had no reasonable suspicion even for a Terry stop at any time before the seizure. See Terry v. Ohio, 392 U.S. 1, 30-31, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968). Nor is there a claim that the district court used an incorrect legal standard. The correct standard, and the one used, is that if there is no reasonable suspicion of a crime, then the police may not, consistent with the Fourth Amendment, make a reasonable person feel that he or she is not free to leave or to terminate the encounter. See Florida v. Bostick, 501 U.S. 429, 439, 111 S.Ct. 2382, 115 L.Ed.2d 389 (1991); Michigan v. Chesternut, 486 U.S. 567, 573-74, 108 S.Ct. 1975, 100 L.Ed.2d 565 (1988); United States v. Mendenhall, 446 U.S. 544, 554-55, 100 S.Ct. 1870, 64 L.Ed.2d 497 (1980) (plurality opinion). When the police do so, they have “seized” a person under the Fourth Amendment. See, e.g., Bostick, 501 U.S. at 439, 111 S.Ct. 2382.
The prosecution in this case, accordingly, chooses to fight on the ground that there was no stop at all and thus no seizure: that the appellate court should, contrary to the district court, conclude that a reasonable person in Smith’s position would have felt free to leave or to terminate the encounter with the police.
I.
The facts, according to the district court’s findings, are these. Quinton Smith, a black man, was waiting for a bus at 1:30 in the afternoon on a drizzly September day in Dorchester, a largely minority neighborhood in Boston. Because there were 8 or 10 people crowding the bus stop, which had no seats anyway, Smith walked across the street to a place where he could sit, and settled down on a three-foot-high wall. His choice of a place to wait was a good one: from there he could see the bus coming and have enough time to get back and board it, and an *33adjacent tree provided some shelter from the rain. See Smith, 332 F.Supp.2d at 280-81.
The bus did not come, but a marked police cruiser did. The two officers in the cruiser patrolled this-high crime area. The officers, who knew a number of the regulars in the area, cruised by Smith and did not recognize him. They went around the block and saw him again on their return. The cruiser pulled up to the curb close to Smith. An officer leaned out of the passenger-side window, nearest Smith, and asked Smith if he lived in the house in front of which he was sitting. Smith answered, accurately, that he did not. The officer asked Smith what he was doing sitting on the wall. Smith answered that he was waiting for the bus. The officer sarcastically asked why Smith was not actually at the bus stop. Smith replied politely that he could catch the bus from where he was sitting. See id. at 279-81 & n. 2.
The officers testified that Smith’s responses did not make sense to them because someone waiting for the bus would not be sitting where Smith was. See id. at 280-81. As a result, they believed that they had a basis to be reasonably suspicious of him. That the officers believed they had reasonable suspicion is also supported by the fact that their later Incident Report stated: “[T]he officers stopped their marked [motor vehicle] to talk to the suspect and to fill out an FIO.” Id. at 281 n. 10. Officers are directed by Boston Police Department regulations to complete an FIO, or Field Intelligence and Observation Report, upon observation of “known criminals” or persons “suspected of having an unlawful design.” Id. It is clear, then, that the officers considered Smith to be a “suspect,” and that he was either a “known criminal! ]” or was “suspected of having an unlawful design.” Their perception of events is the important point.
The district court, having taken a view of the scene, sat on the wall, and looked for the bus, found that it was entirely sensible for someone to wait at that spot for the bus and thus that the officers did not in fact have reasonable suspicion of Smith. The stop was 100 feet away and could be reached in 13 to 16 seconds. See id. at 280-81.
Although the government argues that there was no stop at all, and so no need for even reasonable suspicion, the next steps taken by the officers are indeed the steps that would be taken by officers who believed that they were authorized to detain and question someone. The officers did not remain in the car to question Smith, as they easily could have done. They got out of the car, prepared, if needed, to apprehend him. See id. at 281 & n. 6, n. 7. They then came physically close to Smith, apparently positioning themselves to take him into custody. They did not stand at normal conversational distance. The district court found that the two officers placed themselves so that each was within arm’s reach of Smith; Smith testified that they were each three feet away from him. See id. at 281. The officers also spread themselves so one was on either side of Smith, again consistent with officers preparing to take a suspect into custody. See id. He was seated; they were standing. The district court found that the officers’ tone became “increasingly sarcastic and aggressive” throughout the encounter.6 Id. They soon asked Smith for his identification. See id. at 282.
*34The officers effectively boxed Smith in, given the particular geographic features of the location where Smith was sitting. Since Smith had uneven ground, a fence, and other obstacles behind him, a telephone pole in front of him, and officers on either side of him, he literally had no way out. See id. at 281. He was not free to leave, nor was he free to terminate the encounter, as the increasingly hostile tone of the officers’ questions made clear. As the district court put it:
Although there were only two officers, their presence was threatening because effectively they surrounded Smith and blocked his path to the bus stop. It was additionally threatening because, as discussed below, they approached after he had responded truthfully to the officers’ initial questions. Though holstered, both officers[ ] were armed and in uniform. Griffin’s tone was increasingly aggressive and sarcastic throughout.
Smith gave polite answers. The officers responded by closing in on him and questioning him in an increasingly aggressive tone. He told them the truth, and they became more forceful rather than less. They moved closer to him rather than further away. The timing of their approach was a signal that the questions were going to continue, whether Smith wanted to leave or not.
The government has compared the officers’ positions around Smith to that of two friends who had approached defendant to have a conversation with him. Their guns, uniforms, and the fact that they did not know each other aside, the circumstances under which they ap-proaehed Smith would have indicated to him that this was neither a friendly nor an optional encounter.
Any doubt about this was removed when Griffin asked Smith for identification, a request that Smith reasonably believed he had no choice but to answer. Putting aside the fact that the officers were blocking his ability to watch for the bus, much less catch the bus, Smith would not have had time to terminate the conversation with the officers, get his identification back, and still catch the bus. Griffin’s “request” for identification was a clear statement: forget about the bus, you are not going anywhere.
Id. at 283.
Smith had no non-suspicious alternative to turning down the officers’ inquiry.7 The district court’s conclusion was not based on the officers’ mere asking of questions, nor on their merely asking for identification.8
II.
Ornelas held that Fourth Amendment conclusions of law as to a district court’s determination of “reasonable suspicion” and “probable cause” are subject to a two-part standard of review:
The principal components of a determination of reasonable suspicion or probable cause will be the events which occurred leading up to the stop or search, and then the decision whether these historical facts, viewed from the standpoint of an objectively reasonable police officer, amount to reasonable suspicion or to probable cause. The first part of the *35analysis involves only a determination of historical facts, but the second is a mixed question of law and fact: [T]he historical facts are admitted or established, the rule of law is undisputed, and the issue is whether the facts satisfy the [relevant] statutory [or constitutional] standard, or to put it another way, whether the rule of law as applied to the established facts is or is not violated.
We therefore hold that as a general matter determinations of reasonable suspicion and probable cause should be reviewed de novo on appeal.... [A] reviewing court should take care ... to review findings of historical fact only for clear error....
Ornelas, 517 U.S. at 690, 696-97, 699, 116 S.Ct. 1657 (citation and internal quotation marks omitted; first three alterations in original). Ornelas therefore indicated that the district court’s ultimate conclusion as to whether probable cause or reasonable suspicion existed should be reviewed de novo, although it also emphasized that the district court’s historical findings of fact should be reviewed only for clear error.
The Ornelas Court gave several reasons for subjecting the ultimate conclusion on these two determinations — reasonable suspicion and probable cause — to de novo review. The first was the need for uniformity of rules throughout the nation. See id. at 697, 116 S.Ct. 1657. Another was that “the legal rules for probable cause and reasonable suspicion acquire content only through application.” Id. Finally, the Court noted that de novo review tends to “come closer to providing law enforcement officers with a defined set of rules which, in most instances, makes it possible to reach a correct determination beforehand as to whether an invasion of privacy is justified in the interest of law enforcement.” Id. at 697-98, 116 S.Ct. 1657 (citations and internal quotation marks omitted). It acknowledged, though, that since the legal “mosaic” in this area is so “multifaceted, one determination will seldom be a useful ‘precedent’ for another.” Id. at 698, 116 S.Ct. 1657 (internal quotation marks omitted).
The Ornelas Court next clarified the type of de novo review it envisioned. De novo review in this particular context is not unmindful of the district court’s reasoning (nor of the reasoning of the officers); rather, the appellate court must give:
due weight to inferences drawn from [the historical] facts found by resident judges and local law enforcement officers.
A trial judge views the facts of a particular case in light of the distinctive features and events of the community; likewise, a police officer views the facts through the lens of his police experience and expertise. The background facts provide a context for the historical facts, and when seen together yield inferences that deserve deference.
Id. at 699, 116 S.Ct. 1657; see also United States v. Arvizu, 534 U.S. 266, 276-77, 122 S.Ct. 744, 151 L.Ed.2d 740 (2002); Arvizu, 534 U.S. at 278, 122 S.Ct. 744 (Scalia, J., concurring) (“[W]e have here a peculiar sort of de novo review.”); United States v. Santos, 403 F.3d 1120, 1125 (10th Cir.2005) (noting that de novo review under Ornelas “[i]n practice ... looks more like deference — -indeed, double deference”); United States v. Townsend, 305 F.3d 537, 542 (6th Cir.2002).
The issue before us is neither reasonable suspicion nor probable cause, the subjects of the Ornelas standard, but the subsidiary issue of whether a “seizure” occurred within the meaning of the Fourth Amendment. On this issue, the circuits, before Ornelas, *36were split.9 Although one circuit has expressly decided to adhere to settled circuit precedent and to continue reviewing the determination of whether a seizure occurred using the “clearly erroneous” standard, see United States v. Mask, 330 F.3d 330, 335 (5th Cir.2003), several circuits have cited Ornelas — although generally without discussion — as support for the proposition that the ultimate determination of whether a seizure occurred is reviewed de novo, see, e.g., United States v. Williams, 413 F.3d 347, 351 (3d Cir.2005); United States v. Avery, 137 F.3d 343, 348 (6th Cir.1997); United States v. Hernandez, 93 F.3d 1493, 1498 (10th Cir.1996). This circuit has never — either before or after Ornelas — firmly addressed the question in a clear way.10
I will assume, for these purposes, that the district court’s seizure conclusion is subject to de novo review under Ornelas. The Supreme Court has described the “probable cause” and “reasonable suspicion” standards as mixed questions of law and fact. See Ornelas, 517 U. We have said in other settings that the more the issue is one of law, such as the setting of standards, the less deference is generally given to a district judge’s conclusion. See, e.g., In re Extradition of Howard, 996 F.2d 1320, 1328 (1st Cir.1993). The seizure determination is, at least here, quite heavily at the fact end of the spectrum.
The Supreme Court has described the ultimate seizure question' — whether a reasonable person would feel free to leave or terminate an encounter — as highly dependent on particular facts and circumstances. See, e.g., Bostick, 501 U.S. at 439, 111 S.Ct. 2382. That means there are fewer policy decisions involved. The inquiry here is in fact so “multifaceted” and fact-specific that, even as compared with reasonable suspicion and probable cause determinations, it is perhaps less likely to be valuable as a source of guidance for law enforcement.11 See Ornelas, 517 U.S. at 699, 116 S.Ct. 1657.
In Bostick, the Court explained:
*37[I]n order to determine whether a particular encounter constitutes a seizure, a court must consider all the circumstances surrounding the encounter to determine whether the police conduct would have communicated to a reasonable person that the person was not free to decline the officers’ requests or otherwise terminate the encounter.
501 U.S. at 439, 111 S.Ct. 2382. “The test is necessarily imprecise, because it is designed to assess the coercive effect of police conduct, taken as a whole, rather than to focus on particular details of that conduct in isolation. Moreover, what constitutes [a seizure] will vary, not only with the particular police conduct at issue, but also with the setting in which the conduct occurs.” Chesternut, 486 U.S. at 573, 108 S.Ct. 1975; see also United States v. Cardoza, 129 F.3d 6, 15 (1st Cir.1997) (“The test employed in this area is highly fact specific.”). In making the seizure determination, the pertinent facts are myriad.
This court has never said that there can be no seizure if certain facts are not present. Indeed, Supreme Court and circuit precedent has rejected any such method of analysis. See, e.g., Bostick, 501 U.S. at 439-40, 111 S.Ct. 2382 (holding that Florida Supreme Court erred in adopting per se rule that seizure occurs whenever police questioning occurs on bus, since this fails to consider totality of circumstances); Florida v. Royer, 460 U.S. 491, 506, 103 S.Ct. 1319, 75 L.Ed.2d 229 (1983) (“We do not suggest that there is a litmus-paper test for distinguishing a consensual encounter from a seizure.... Even in the discrete category of airport encounters, there will be endless variations in the facts and circumstances....”); Cardoza, 129 F.3d at 14-16; Young, 105 F.3d at 6. Utter disregard by an appellate court of the district court’s views on whether there was a seizure could place at risk the required multi-factored, totality approach to the question.
Applying the de novo standard of review under Ornelas to seizure issues, we must still, of course, give considerable deference to the district court’s historical findings of fact — for example, where the officers stood in relation to the sidewalk and telephone pole, how much space they took up and how much room this left Smith to get around them, and what tone of voice the officers used during the incident. These findings are reviewed only for clear error, see Ornelas, 517 U.S. at 699, 116 S.Ct. 1657, and any potential challenge by the government to the district court’s factual findings has not been raised on appeal and has therefore been waived, see, e.g., Diva’s Inc. v. City of Bangor, 411 F.3d 30, 39 (1st Cir.2005). We therefore must assume that the district court’s historical findings of facts are all true.
A reversal here would require departure from this rule as well as the rule that appellate courts are entitled to draw conclusions of law, but not to make findings of fact. The district court distinctly did not find that the officers stood “in the only place they could — on either side of the telephone pole directly in front of Smith.” Maj. op. at 30. Indeed, the district court concluded to the contrary and held that as the encounter escalated, the officers moved in closer to Smith, obstructing his freedom of movement. Smith, 332 F.Supp.2d at 283. In fact, the officers could have remained in the cruiser or situated themselves differently on the seven-foot-wide sidewalk. In support of its contrary fact-finding, the majority relies on its understanding of photographs of the scene. This disregards the fact that the district court was actually present at the scene, as *38the majority was not. Further, the photographs show that the officers had other choices as to where to position themselves.
As to another key conclusion, it is difficult to see how an appellate court could reach a different conclusion regarding the officers’ tone of voice and threatening presence than did the district court, which heard the testimony of all three participants to the encounter.
The government relies on Bostick, 501 U.S. 429, 111 S.Ct. 2382, 115 L.Ed.2d 389, and United States v. Brown, 169 F.3d 89 (1st Cir.1999). Bostick, which employs the totality of the circumstances test, supports the district court’s conclusion that Smith was free neither to terminate the encounter nor to walk away from the police. In Bostick, the defendant was a passenger on a bus scheduled to depart for a destination he wanted to reach, and thus would not leave even before the police encounter. Bostick, 501 U.S. at 436-37, 111 S.Ct. 2382. Here, the defendant, before the officers blocked him in, was perfectly free to leave. The government admits “[t]he physical setting of Smith’s encounter with the officers did not approach the restrictiveness of the physical setting in Bostick.” The government argued only that “limitations which were not created by the police ... are insufficient to make an encounter with police a seizure.” The government is quite correct, but the argument is misplaced, and the district court did not say otherwise. Here, physical limitations did not create the seizure; the police did. Brown likewise would not cause a reversal here. The defendant there did not claim to be seized until he assaulted the officer. Brown, 169 F.3d at 92.
For the reasons explained above, we must also follow particularly carefully' — in this context — the Ornelas Court’s dictate to give “due weight” to inferences drawn by the trial judge from the historical facts and to the perceptions of the officers. See Ornelas, 517 U.S. at 699, 116 S.Ct. 1657. In Arvizu, for instance, a “reasonable suspicion” case, the Supreme Court applied this principle to give deference to the district court’s inferences that certain individuals’ method of waving was “ ‘methodical,’ ‘mechanical,’ ‘abnormal,’ and ‘certainly ... a fact that is odd and would lead a reasonable officer to wonder why they are doing this.’ ” 534 U.S. at 276-77, 122 S.Ct. 744. Similarly, the various intermediate factual inferences made in this case by the district court, which heard the witnesses and saw the scene, are entitled to deference. These include, for example, inferences about the threatening nature of the encounter, that Smith had “no meaningful egress” from his spot on the wall once the officers had approached because his possible escape routes would have seemed suspicious, and that the officers’ request for Smith’s identification, at the time they asked for it, was a “clear statement: forget about the bus, you are not going anywhere.” See Smith, 332 F.Supp.2d at 282-83.
The government argues that Smith’s confinement was self-imposed: he remained where he was because he wanted to catch the bus to go home, and the bus shelter was already crowded. It is odd to think that a person must be put to a choice under the Fourth Amendment between catching a bus to go home and terminating an encounter with the police. Once the officers requested Smith’s identification, Smith knew he could not catch a bus while the police had his identification. Thus, the district court’s inference about the meaning of the officers’ request for Smith’s identification makes sense. See id. at 283.
These inferences of the district court must be coupled with the indications that the officers themselves saw the situation as one where they had reasonable suspi*39cion to at least engage in a Terry stop. The officers’ actions were entirely consistent with a seizure based on reasonable suspicion, which explains their failure to approach Smith in a way which communicated that he was free to leave.
According the respect due under Orne-las to the district court’s findings of historical fact and inferences from these facts, and to the actions of the officers based on their beliefs that they had reasonable suspicion, and applying de novo review, I would affirm the district court’s suppression order. .
. It is also true that the officers did not draw their weapons. See Smith, 332 F.Supp.2d at 281. They did not say anything one way or another about whether Smith was free to leave.
. The government argued that Smith was free to leave because he could have climbed over the wall or climbed through a small opening in a fence, and so forth. As the district court noted, though, "[t]here was no meaningful egress.” Id. at 281. Any such action would have been suspicious on his part.
. I note that none of the officers' actions would have been improper if there was objectively reasonable suspicion under Teny, and thus proper grounds for a Terry stop.
. See 6 W. LaFave, Search and Seizure: A Treatise on the Fourth Amendment § 11.7(c), at 448 (4th ed.2004). The de novo standard of review on the "seizure” issue was adopted in United States v. Buchanon, 72 F.3d 1217, 1222-23 (6th Cir.1995); United States v. McKines, 933 F.2d 1412, 1424-25 (8th Cir.1991); United States v. Montilla, 928 F.2d 583, 588 (2d Cir.1991); and United States v. Maragh, 894 F.2d 415, 417-18 (D.C.Cir.1990). For decisions adopting the clearly erroneous standard of review, see, for example, United States v. Gray, 883 F.2d 320, 322 (4th Cir.1989); United States v. Teslim, 869 F.2d 316, 321 (7th Cir.1989); and United States v. Archer, 840 F.2d 567, 571 (8th Cir.1988).
. A de novo standard is perhaps suggested by the language of United States v. Cardoza, 129 F.3d 6, 13-14 (1st Cir.1997) and United States v. Young, 105 F.3d 1, 5 (1st Cir.1997).
In Thompson v. Keohane, 516 U.S. 99, 115-16, 116 S.Ct. 457, 133 L.Ed.2d 383 (1995), the Supreme Court held that the question of whether a defendant was "in custody” for Fifth Amendment purposes, and therefore must be given Miranda warnings, was a question of law and not fact for habeas purposes. Thompson held that no presumption of correctness should be given to such determinations of state courts on habeas review. See id. at 116, 116 S.Ct. 457. We have cited Thompson to the effect that, "arguably,” review of a Fifth Amendment "in custody” determination is de novo in a non-habeas context. United States v. Teemer, 394 F.3d 59, 65 (1st Cir.2005); see United States v. Fornia-Castillo, 408 F.3d 52, 63 (1st Cir.2005) (applying de novo standard); United States v. Trueber, 238 F.3d 79, 93 (1st Cir.2001) (same).
. As well, since the police officer is the prime actor in the seizure, it would not be uncommon for an objective person to conclude that an officer's conduct amounted to a seizure, at the same time the officer believed his conduct to be entirely benign. In other words, an objective observer may well believe that the police had communicated to a reasonable person that the person was not free to decline the officer’s request or otherwise ter-*37mínate the encounter, while the officer believed otherwise.