Judge JOSÉ A. CABRANES filed a separate opinion, concurring in part and dissenting in part.
Judge JANET C. HALL filed a separate opinion, concurring in part and dissenting in part.
CALABRESI, Circuit Judge,delivering the majority opinion with respect to Part II, in which CABRANES, Circuit Judge, joins, and the majority opinion with respect to Part III, in which HALL, District Judge, joins.
I.
Sometime in the early afternoon of November 8, 2003, local law enforcement officials executed a search warrant at 205 Glenwood Avenue in Rochester, New York (“the Glenwood residence”). The warrant, which had issued on the basis of reports by two confidential informants that they had separately bought dime bags of cocaine from persons at that address, authorized a no-knock entry and search of the Glen-wood residence for cocaine, for records showing the sale and trafficking of cocaine, for proceeds from such sales, and for documents indicating the occupancy, residency and/or ownership of the premises. The warrant did not, however, authorize any arrests, nor was there evidence suggesting the involvement of any specific individuals.
Lieutenant Eric Paul (“Paul”) of the Rochester Police Department was one of the first law enforcement officials to enter the Glenwood residence.1 Upon entering *54the residence, Paul observed Lionel Sum-mersett (“Summersett”) step out of the bathroom at the top of the home’s stairwell. From the bottom of the stairs, Paul ordered Summersett to lie on the floor. With gun drawn, Paul held Summersett in that position until the other officers completed an initial safety sweep of the home’s ground floor.2
Paul then went up the stairs. While other officers took Summersett into custody, Paul went into the upstairs bedroom closest to the stairwell. There he found defendant-appellee Jerrell Heath (“Heath”) sitting on the bed, talking on a cell phone. Paul placed Heath in handcuffs, ordered him to lie on a pile of clothes, and pat-frisked him. The frisk revealed nothing incriminating. Paul and other officers then searched the bedroom. Behind a dresser, concealed from plain view, Paul discovered a small bag with one-half ounce of marijuana. The dresser was approximately five feet from where Heath was found sitting on the bed talking on the phone.
Once Paul discovered the marijuana, other members of the Rochester Police Department took Heath from the bedroom and formally arrested him. A search attendant to that arrest revealed that Heath had $3,073 in cash.
While Heath was being arrested in the hallway, Paul continued to search the bedroom and discovered a loaded 9mm handgun in the top dresser drawer. Paul then left the bedroom and was informed that other officers had found several small bags of cocaine in plain view in the residence.
Ultimately, Heath was charged with the following offenses: possession of a firearm in furtherance of drug trafficking, in violation of 18 U.S.C. § 924(c)(1)(A); being a felon in possession of a firearm, in violation of 18 U.S.C. 922(g)(1); possession of marijuana, in violation of 21 U.S.C. § 844; and possession of cocaine base with intent to distribute, in violation of 21 U.S.C. § 841(a)(1). Thereafter, in proceedings before Magistrate Judge Marian W. Pay-son, Heath argued that he had been arrested without probable cause, and moved to suppress the seized currency. In due course, the magistrate judge issued a report recommending that, because the only evidence linking Heath to the marijuana was his proximity to that hidden contraband, the police lacked probable cause for his arrest. The magistrate judge based this conclusion on three Fourth Amendment cases discussing “premises” and “proximity” liability: United States v. Di Re, 332 U.S. 581, 68 S.Ct. 222, 92 L.Ed. 210 (1948), Sibron v. New York, 392 U.S. 40, 88 S.Ct. 1889, 20 L.Ed.2d 917 (1968), and Ybarra v. Illinois, 444 U.S. 85, 100 S.Ct. 338, 62 L.Ed.2d 238 (1979). The magistrate judge further held that the Supreme Court’s recent decision in Maryland v. Pringle, 540 U.S. 366, 124 S.Ct. 795, 157 L.Ed.2d 769 (2003), did not justify Heath’s arrest, as Heath was arrested in a “fixed residence,” not in an automobile. Finally, the magistrate judge determined that the currency found on Heath’s person was not admissible under the “inevitable discovery” doctrine.
The district court (Telesca, J.) agreed with the magistrate judge’s recommendation and ordered that the evidence found on Heath be suppressed. In so doing, the district court declared its unwillingness to extend Pringle to the situation in this case, because “Fourth Amendment jurisprudence recognizes a distinct difference between a passenger in an automobile and an occupant of a fixed premises.”
*55In this interlocutory appeal, the government argues that the district court erred by suppressing the evidence. We need not, at this time, address the government’s primary contention on appeal — that Heath’s arrest was constitutional under Pringle — because, assuming arguendo that Heath was arrested without probable cause, we find that the evidence at issue may well have been admissible under the inevitable discovery doctrine, notwithstanding the putative Fourth Amendment violation. In so doing, we emphasize that the inevitable discovery doctrine is available only where there is a high level of confidence that each of the contingencies required for the discovery of the disputed evidence would in fact have occurred. In circumstances such as those before us, where the government contends that the challenged evidence would inevitably have been discovered during a search incident to a valid arrest, one of the contingencies that must be resolved in the government’s favor involves a police officer’s discretionary decision to arrest and search the person on whom the evidence would presumably have been found.
On the record before us, we believe that the district court erred in reaching a firm conclusion that the currency seized on Heath’s person would not have been inevitably discovered, and we conclude that further consideration of that issue is needed. At the moment of Heath’s arrest, Paul and the arresting officers knew only that Heath was in the same room as a small, concealed quantity of narcotics. Even if this was not an adequate basis upon which to arrest him — and we assume, arguendo, that it was not — -a sufficient reason to justify the arrest and a search attendant to the arrest arguably became evident a few moments later. But whether the relevant officers would have acted on that reason, and would then have arrested Heath, thereby bringing the inevitable discovery doctrine into play, is not clear to us. Nor do we have findings by the district court on the matter. We therefore remand the case for further findings.
* * *
Under the “inevitable discovery” doctrine, evidence obtained during the course of an unreasonable search and seizure should not be excluded “if the government can prove that the evidence would have been obtained inevitably” without the constitutional violation. Nix v. Williams, 467 U.S. 431, 447, 104 S.Ct. 2501, 81 L.Ed.2d 377 (1984); see also United States v. Eng, 997 F.2d 987, 990 (2d Cir.1993) (inevitable discovery doctrine “requires the district court to determine, viewing affairs as they existed at the instant before the unlawful search, what would have happened had the unlawful search never occurred” (emphasis in original) (internal quotation marks omitted)). In essence, the inevitable discovery doctrine’s application turns on a central question: Would the disputed evidence inevitably have been found through legal means “but for” the constitutional violation? If the answer is “yes,” the evidence seized will not be excluded. In the case before us, this inquiry requires the government to prove: (a) that law enforcement officers could have validly arrested Heath on the basis of evidence that became available moments after Heath was (arguably improperly) arrested, and (b) that law enforcement officers would have acted on that evidence and, given the circumstances, ivould have arrested Heath. That is, the currency should not be excluded if there is sufficient proof that a later arrest was valid and sufficiently likely to occur as to qualify as “inevitable.”
II. Validity of a Later Arrest
The arguably valid reason for a later arrest, and hence for inevitable dis*56covery, came about in the following way. As soon as Paul emerged from the bedroom, he learned that other officers had discovered several “loose dime bags of crack cocaine” on the landing of the stairwell, on the stairwell itself, and at the bottom of the stairwell. At that point, Paul observed what appeared to be small bags of cocaine in plain view in each of these locations. Three of the small bags of cocaine were at the bottom of the stairs. The magistrate judge determined that Summersett must have dropped the bags of cocaine on the stairwell and on the landing (which were found underneath or around Summersett) when Paul initially commanded him to lie down. The magistrate judge’s factual finding — that the “logical inference” to be drawn from the circumstances was that the cocaine found underneath or immediately around Sum-mersett was dropped by him upon exiting the bathroom — -was not clearly erroneous. And so we cannot say that the cocaine assertedly dropped by Summersett created probable cause to arrest Heath, because these facts would not, in and of themselves, lead a “person of reasonable caution,” United States v. Rogers, 129 F.3d 76, 79 (2d Cir.1997), to believe that Heath was complicit in Summersett’s apparent possession of the cocaine found on the stairwell and on the landing.
But neither the magistrate judge nor the district court reached any similar conclusion as to the origin of the cocaine in plain view at the bottom of the stairwell. While the magistrate judge noted that the “logical inference” was that the cocaine on the stairwell, and on the landing, was dropped by Summersett “on his way to or from the bathroom,” no such finding was made regarding the origin of the cocaine at the bottom of the stairwell.3 Similarly, *57although the district court’s order noted that the government’s evidence did not establish that Heath had actually seen narcotics (or firearms) anywhere in the house prior to the officers’ entry into the residence, the district court did not, at any point, make specific findings on this question.
Viewing the facts in the light most favorable to the prevailing party (here, Heath), see United States v. Harrell, 268 F.3d 141, 145 (2d Cir.2001), we cannot say that Heath actually saw the cocaine on the stairwell, on the landing, or at the bottom of the stairwell before he entered the bedroom and began to speak on the cell phone. Still, undisputed evidence indicates that the bags of cocaine at the bottom of the stairs were in plain sight when the officers discovered them. And a reasonable officer could, therefore, conclude that the bags at the bottom of the stairs were likely to have been there prior to the arrival of law enforcement officials at the Glenwood residence.
Under the circumstances, whether Heath actually saw the bags of cocaine at the bottom of the stairs before the police entered the Glenwood residence is of no import. He could see them, and in fact would be expected to see them, in the regular course of walking through the small home’s public spaces. As a result, a “person of reasonable caution,” Rogers, 129 F.3d at 79, could properly have concluded that Heath had committed or was committing a crime. After all, those who are permitted to observe obvious criminal activity in a home are, absent indications to the contrary, likely to be complicit in the offense. See, e.g., United States v. Pennington, 287 F.3d 739, 747 (8th Cir. 2002) (“Having entered the residence lawfully, the officers could act upon the drug manufacturing evidence in plain view. That evidence provided probable cause to arrest [the occupant of the apartment], which justified a search of his person incident to the arrest.”) (internal citation omitted); United States v. Jones, 72 F.3d 1324, 1332-33 (7th Cir.1995) (concluding that probable cause existed based on defendant’s “evident status as a resident of a private household in which counterfeit money was being sold on a regular basis and in which the evidence of counterfeiting lay in plain view” (emphasis added)); United States v. Holder, 990 F.2d 1327, 1329 (D.C.Cir.1993) (holding that keeping narcotics “openly on display” in a private residence was indicative that the residence’s owner considered a visitor “sufficiently complicit to allow him a full view”); Hollyfield v. United States, 407 F.2d 1326, 1326 (9th Cir.1969) (holding that visible contraband in plain view within the apartment provided probable cause for an arrest of person found therein); cf. United States v. MacDonald, 916 F.2d 766, 768, 770 (2d Cir.1990) (holding that undercover agent’s observation of cocaine and marijuana in plain view throughout an apartment containing six men provided probable cause to arrest all six occupants).
In the case before us, there is no evidence that any precautions were taken to prevent people in the house from coming across the putatively visible contraband, and from doing so in the ordinary course of their expected behavior. As a result, reasonably cautious police officers could have concluded that the home’s adult occupants were complicit in the illegal activities involving the contraband. It follows that the law enforcement officials seemingly had a valid ground for arresting both of the adult occupants of the Glenwood resi-*58denee.4 And, accordingly, we conclude that Heath’s arrest, though effectuated in quick order and arguably prematurely, might well have validly occurred in an appropriate manner as soon as the plainly visible contraband came to the attention of the arresting officers.5
III. Inevitability of a Later Arrest
In cases where the government argues that disputed evidence would inevitably have been discovered during a search incident to arrest, the government must establish that the arrest would have been valid. But, evidence that an arrest would have been supported by probable cause and, therefore, could legally have taken place, only makes the inevitable discovery doctrine potentially applicable. That is, to find that reasonable police officers could properly believe that they had probable cause to arrest Heath shortly after they, seemingly improperly, arrested him does not mean that they would inevitably have done so at the later time. The circumstances of this case force us to focus on the possible gap between, on the one hand, facts that are sufficient to qualify as probable cause and thereby enough to validate an arrest, and, on the other hand, facts that are sufficient to establish that a legal arrest and subsequent search would inevitably have taken place very soon after the arguably unconstitutional arrest. Even if the present record provides enough evidence to suggest that a reasonable police officer could have made a valid arrest supported by probable cause, it does not establish with a sufficiently high degree of certainty that a reasonable police officer would have made the arrest under the circumstances. Because we cannot, on the record before us, resolve this latter contingency, we cannot at this time decide whether Paul and his fellow officers would in fact have discovered the currency whose exclusion is at issue in this case. We therefore remand the case for more findings.
We have previously indicated that the government cannot prevail under the inevitable discovery doctrine merely by establishing that it is more probable than not that the disputed evidence would have been obtained without the constitutional violation.6 See United States v. Cabassa, *5962 F.3d 470, 472-73 (2d Cir.1995). On the contrary, we held in Cabassa that proving that a judge could validly have issued a warrant supported by probable cause was not necessarily enough to establish that a judge would have issued the warrant in question. See id. at 474. In that case, agents from the Drug Enforcement Agency (“DEA”) entered the defendant’s home and seized controlled substances, weapons, and money, all without a warrant. See id. at 472. At the time of the search, other DEA agents were in the process of requesting, but had not obtained, a warrant from a magistrate judge. Id. Rejecting the district court’s conclusion that the contraband would inevitably have been found because a warrant would surely have been issued, we suppressed the illegally-acquired evidence on the grounds that the magistrate judge might not have granted a warrant to conduct the search. We did so even though we conceded that the district court might well have been correct that the “government’s draft affidavit demonstrated probable cause” to support the issuance of a warrant. See id. at 473. But, because there was “some room for disagreement,” we found “a residual possibility that a magistrate judge would have required a stronger showing of probable cause” before authorizing the search and seizure. Id. at 473-74. Since this eventuality could not be resolved in the government’s favor, we declared the evidence inadmissible. Id.
Our decision in Cabassa established that evidence minimally sufficient to support probable cause would not always be enough to demonstrate that a governmental actor vested with discretion — e.g., a magistrate judge asked to issue a warrant — would act on that evidence.7 On the basis of Cabassa, district courts in our circuit have held that the inevitable discovery exception to the exclusionary rule is available only where a court has a “high level of confidence” that “each of the contingencies” needed to obtain the evidence legally would be resolved in the government’s favor;
The teaching of Cabassa, which is supported by principles of probability, thus is straightforward. It suggests that a *60trial court’s task in [the] context [of inevitable discovery] is to deny the motion to suppress on the ground of inevitable discovery only if it has a high level of confidence that the warrant in fact would have issued and that the specific evidence in question would have been obtained by lawful means. Inevitable discovery analysis therefore requires a court to examine each of the contingencies that would have had to have been resolved favorably to the government in order for the evidence to have been discovered legally and to assess the probability of that having occurred.
United States v. Lavan, 10 F.Supp.2d 377, 389 (S.D.N.Y.1998) (Kaplan, J.) (emphasis added); see also United States v. Arms, 2002 WL 32781, *5 (E.D.N.Y.2002) (Block, J.) (finding that discovery of the evidence was inevitable when there is “no question that a warrant would in fact have issued”).
The Tenth Circuit has similarly required substantial certainty with respect to each of the contingencies involved in the causal chain of inevitability posited by the government. See United States v. Cunningham, 413 F.3d 1199, 1203 (10th Cir.2005) (“[W]e held it permissible for a court to apply the inevitable discovery doctrine when it has a high level of confidence that the warrant in fact would have been issued and that the specific evidence in question would have been obtained by lawful means. Inevitable discovery analysis thus requires the court to examine each of the contingencies involved that would have had to have been resolved favorably to the government in order for the evidence to have been discovered legally and to assess the probability of the contingencies having occurred.” (internal quotation marks and citation omitted)). See also United States v. Romero, 692 F.2d 699, 704 (10th Cir.1982) (“Under the inevitable discovery exception, unlawfully seized evidence is admissible if there is no doubt that the police would have lawfully discovered the evidence later.” (emphasis added)).
In contrast, other circuits have required only that there be a “reasonable probability that the contested evidence would have been discovered by lawful means in the absence of the police misconduct.” United States v. Chambers, 132 Fed.Appx. 25, 33 (5th Cir.2005); see also Jefferson v. Fountain, 382 F.3d 1286, 1296 (11th Cir.2004); United States v. James, 353 F.3d 606, 617 (8th Cir.2003); United States v. Kennedy, 61 F.3d 494, 499 (6th Cir.1995); United States v. Thomas, 955 F.2d 207, 210 (4th Cir.1992). And before our decision in Ca-bassa, two district courts of this circuit also used the “reasonable probability” threshold. See United States v. Yanes, 671 F.Supp. 927, 933 (D.Conn.1987) (Dorsey, J.); United States v. Levasseur, 620 F.Supp. 624, 631 (S.D.N.Y.1985) (Glasser, /.).
To the extent that any confusion lingers in our circuit after Cabassa, we now expressly eschew the “reasonable probability” framework that some of our sister circuits have used to analyze “inevitable discovery” cases. In its place — and consonant with the “teachings of Cabassa,” with the Tenth Circuit’s methodology in Cunningham, and with the approach most recently taken by our district courts — we conclude that illegally-obtained evidence will be admissible under the inevitable discovery exception to the exclusionary rule only where a court can find, with a high level of confidence, that each of the contingencies necessary to the legal discovery of the contested evidence would be resolved in the government’s favor.
Applying this analysis to the instant case, the issue becomes: can we conclude, with sufficient confidence, that law enforcement officers — assuming that they were legally entitled to arrest Heath — would have done so under the cir*61cumstances? For the purposes of this inquiry, we apply Cabassa’s reasoning— which analyzed a judge’s discretionary decision to grant a search warrant — to a police officer’s discretionary decision to make an arrest.8 With respect to such a decision by a police officer, we hold that relatively weak evidence of the right to arrest is not sufficient, without further findings, to establish that an officer would, in fact, have made an arrest. To hold otherwise would undermine the discretion we give to police officers to decide — on the basis of police protocol, of the circumstances confronting them, and of their trained judgment — whether and in what cases they should use their power to arrest.9
For some circuits, evaluating a police officer’s discretionary decision to arrest as part of the inevitable discovery analysis is nothing new. Although, in the decisions we have found, this evaluation resulted in the admission of the evidence, in various key eases, the evidence was admitted only after an express determination that an arrest would, in fact, have occurred. And, the validity of the arrest seemed to be a necessary, but not sufficient, condition for the application of the inevitable discovery rule. Thus, in United States v. Cherry, 759 F.2d 1196 (5th Cir.1985), the Fifth Circuit admitted contested fingerprint evidence after explicitly agreeing with the district court’s conclusion “that [the defendant] eventually would have been lawfully arrested and fingerprinted.” Id. at 1208 (emphasis added).10 Similarly, in United States v. White, 326 F.3d 1135 (10th Cir. *622003), the Tenth Circuit permitted evidence that purportedly would have been found during a lawful arrest only after deciding: (a) that a background cheek would inevitably have been conducted, and (b) that “ample evidence support[ed] the conclusion that, once the [background] check revealed [the defendant’s] outstanding warrant, [the defendant] would, have been lawfully arrested ...” Id. at 1139 (emphasis added). Hence, in both Cherry and White, the circuit court accepted the government’s argument that the disputed evidence would inevitably have been found during an arrest, but did so after making specific findings, based on a careful examination of the record, not just that the contingent arrest was proper, but also, and importantly, that the arrest was inevitable.11
In contrast, in the circumstances of this case, it is possible that an arrest, even if potentially valid, would not have been made. And if no such arrest would have been made, then there would have been no inevitable discovery, and, hence, the currency at issue would be properly excluded. It is, of course, also possible that the arrest and discovery of the cash would have occurred. But, on the record before us and without further findings by the court below, we cannot say whether the likelihood of such an arrest was great enough to justify the application of the doctrine of inevitable discovery.12
Because we cannot at this time and on this record know whether inevitable discovery would have occmred, we cannot decide whether the district court’s exclusion of the currency at issue was correct. Accordingly, we Vacate the order of the district court and Remand for further proceedings consistent with this opinion.
. Paul was the only witness to testify at the suppression hearing. The account of the search and of Heath’s arrest is based entirely on Paul’s undisputed testimony.
. The initial downstairs sweep revealed two children.
. The magistrate judge did state, in a general comment about the bags of cocaine found throughout the Glenwood residence, that "the record ... does not indicate whether Sum-mersett dropped the bags on his way to or from the bathroom or when he was ordered to lie down on the stairwell, or whether they were recovered during the officers' search of him." There is nothing in the record currently before us, however, that supports the position that these three possibilities were the only ways in which the cocaine could have found its way to the bottom of the stairs. We do not know whether the magistrate judge's list of possible origins of the found drugs was meant to be exclusive. But if it was, that finding would be clearly erroneous. See, e.g., Krizek v. Cigna Group Ins., 345 F.3d 91, 100 (2d Cir.2003) (recognizing that factual findings were clearly erroneous where the "record before the court [was] simply devoid of any basis for the [district] court’s conclusion”).
It is possible to read the magistrate judge’s order, and Paul's testimony, as asserting that cocaine was found in only two locations — the top of the stairwell (which, on this reading, includes the landing), and at the bottom of the stairwell. While we do not read the record in this fashion, we note that even if the cocaine at issue were found in just two locations, the magistrate judge’s findings still would not appear to address the origin of the cocaine at the bottom of the stairwell.
The court's statement that "the logical inference, it seems to me, is that [the bags of cocaine] fell from Summersett on his way to or from the bathroom,” is best read as a statement that the bags of cocaine at the top of the stairwell were dropped during Sum-mersett’s exit from the upstairs bathroom. There is nothing in the record as we have it at this time to support the “inference” that Sum-mersett, while held at gunpoint and observed by Paul, or once taken into custody by the other agents, somehow managed, undetected, to throw several bags of cocaine onto the floor of the apartment at the bottom of the stairwell. Accordingly, any finding that that is how the bags of cocaine got there would, on the existing record, be clear error. On the basis of the facts presently in the record, we, and the magistrate judge, can conclude no more than that the bags of cocaine were at the bottom of the stairwell, and were there in plain view, when Paul observed them. Our factual perspective at this juncture — in which we review ambiguities in the light most favorable to Heath, the prevailing party below, see *57United States v. Harrell, 268 F.3d 141, 145 (2d Cir.2001) — does nothing to change this. For, as we have said, there is currently no evidence in the record that creates an ambiguity.
.Heath cannot be said to have actually seen the cocaine at the bottom of the stairs on his way to the upstairs bedroom. But, as we have already said, evidence demonstrating Heath’s actual observation of the contraband was not necessary to create probable cause for the arrest. What mattered was that contraband was left openly, in a public space, so that a reasonably cautious person could easily conclude that guests and residents of the home were highly likely to see the contraband in the ordinary course.
The presence of any quantity of narcotics (or related contraband) in plain view in any location in a home does not, automatically, create probable cause to arrest every person present in that home. We need not, and do not, today decide exactly what limitations the Fourth Amendment would place on such arrests. For present purposes, we need only recognize that Heath — one of only two adults inside a small home containing, in a highly public and common space, a relatively large quantity of drugs — could reasonably be seized by the arresting officers.
. We find nothing to the contrary in Ybarra, Sibron, or Di Re, which all dealt with conduct that was, on its face, innocent. See Ybarra, 444 U.S. at 88, 100 S.Ct. 338; Sibron, 392 U.S. at 45, 88 S.Ct. 1889; Di Re, 332 U.S. at 593, 68 S.Ct. 222.
. This requirement of certitude should not be confused with the government's burden of proof, which, it is well settled, requires that inevitable discovery be established by a preponderance of the evidence. See Nix, 467 U.S. at 444, 104 S.Ct. 2501 ("If the prosecution can establish by a preponderance of the evidence that the information ultimately or inevitably would have been discovered by lawful means ... then the deterrence ratio*59nale has so little basis that the evidence should be received.” (emphasis added)); see also United States v. Cunningham, 413 F.3d 1199, 1203 (10th Cir.2005) ("The government possesses the burden of proving by a preponderance of the evidence that the evidence at issue would have been discovered without the Fourth Amendment violation.”).
In adopting this standard, we have acknowledged, but have failed to resolve, the paradox of applying the preponderance of the evidence standard in the context of inevitable discovery; "There are, of course, semantic problems in using the preponderance of the evidence standard to prove inevitability.... Given the present facts, we need not probe further into the semantic puzzle other than to note the difference between proving by a preponderance that something would have happened and proving by a preponderance that something would inevitably have happened.” United States v. Cabassa, 62 F.3d 470, 474 (2d Cir.1995) (emphasis added). Because further clarification is not needed to resolve the case before us, we leave to another day the task of ironing out this wrinkle in the doctrine.
. In Cabassa, we distinguished an earlier inevitable discovery case, United States v. Whitehorn, 829 F.2d 1225 (2d Cir.1987), on the grounds that, in Whitehom, there had been " 'overwhelming' probable cause to support the warrant application,” which strengthened the government’s contention in that case that the issuance of a valid warrant was, under the circumstances, truly inevitable. See Cabassa, 62 F.3d at 473 (emphasis added). Accordingly, at least with respect to a judge’s decision to issue a warrant, Cabassa could be read to require "overwhelming probable cause” to demonstrate inevitable discovery. Cabassa's reasoning, however, only establishes that (absent indications to the contrary) overwhelming probable cause is sufficient to find inevitable discovery, not that it is necessary to reach such a conclusion.
. Lending support to the proposition that the discretionary decisions of police officers qualify as "one of the contingencies" necessary to establish inevitable discovery, the First Circuit has intimated that courts should examine a police officer’s decision to seek a warrant in the first place, not just the judge’s decision to grant one, in evaluating a claim of inevitable discovery. See United States v. Silvestri, 787 F.2d 736, 745 (1st Cir.1986) ("[T]he requirement of active pursuit could be viewed as ensuring the independent inevitability of the police decision to seek the search warrant, i.e., to ensure that the evidence turned up in the illegal search did not influence this decision.”).
. In Wayne v. United States, 318 F.2d 205 (D.C.Cir.1963), then-Judge Warren Burger justified the admission of testimony, which was originally taken from a coroner after an illegal search uncovered a dead body, on the grounds that the coroner’s testimony would inevitably have materialized: "It was inevitable that, even had the police not entered appellant’s apartment at the time and in the manner they did, the coroner would sooner or later have been advised by the police of the information reported by the [decedent’s sister], would have obtained the body, and would have conducted the post mortem examination prescribed by law.” Wayne, 318 F.2d at 209. Importantly, to support this conclusion, Burger cited police regulations requiring officers to contact a coroner after finding a dead body, as well as a local law mandating that the coroner conduct an autopsy. See id.
We too have looked to internal police practices to determine whether a police officer's decision was, in fact, inevitable. See, e.g., United States v. Mendez, 315 F.3d 132, 138 (2d Cir.2002) (applying the inevitable discovery exception in part because "[t]he record amply supports the district court’s conclusion that that [the police department] had a consistent, if unwritten, inventory search policy,” which would have led to the evidence notwithstanding the earlier invalid search).
Accordingly, evidence that proper police protocol would have required Heath’s arrest would certainly bear on whether the arrest was in fact inevitable. Conversely, the absence of such guidelines might perhaps suggest that the officers would not necessarily have arrested Heath, and hence that the currency would not inevitably have been found. These are precisely the type of considerations that are suitable for review on remand.
. To justify this conclusion, the Cherry court listed the considerable evidence in support of the government’s position that an arrest both would have been supported by probable cause and would have inevitably happened:
By the time [the defendant] was brought before a magistrate, the FBI agents had *62uncovered the following facts independently of prior misconduct: (1) [the defendant’s] military identification and driver’s license was in the victim's abandoned taxicab; (2) the taxicab’s last dispatch was to [the defendant's] barracks at Fort Bliss; (3) the victim’s wallet had been left in a trash receptacle in the latrine area in [the defendant's] barracks; and (4) members of [the defendant’s] unit had seen him recently in possession of a .32 caliber pistol. We think that the laminated total of this information amounts to probable cause and that the district court’s finding that the agents eventually would have acted upon the probable cause to airest [the defendant] is not clearly erroneous.
Cherry, 759 F.2d at 1208 n. 16 (emphasis added).
. We do not mean to suggest that such an examination is needed in every case. There obviously will be some in which the inevitability of the arrest is sufficiently obvious as not to require discussion. But, as Cherry and White demonstrate, there clearly are others in which a careful review of the record is essential. And, in such cases, where it is possible, this type of parsing of record evidence is best done, in the first instance, by the district court.
. A remand is not rendered futile either by the fact that Heath was actually arrested, or by the possibility that the police officers might testify, on remand, that they would have arrested Heath under circumstances of this case. An officer's isolated statement that an arrest would have occurred (which we do not have on the record before us) — even if seemingly buttressed, as it would be in the instant case, by the fact that the defendant was, in reality, illegally arrested — would not, by itself, be enough to satisfy the government’s burden to establish that the arrest was inevitable. We noted in Eng, after all, that an officer's testimony “as to what he would have done or how he would have behaved, while deserving of careful consideration as part of the analysis, is not necessarily conclusive on the question of whether a particular piece of evidence inevitably would have been found through lawful means.” Eng, 971 F.2d at 862. The question is whether an officer would have "inevitably” acted in a certain way, and that inquiry remains an objective one, not necessarily turning on an officer's own testimony as to what he believes he would have done.