Opinion by Judge O’SCANNLAIN; Partial Concurrence and Partial Dissent by Judge FISHER.
O’SCANNLAIN, Circuit Judge:
The constitutionality of the City of Portland’s policy on the use of deadly force by its police officers is squarely presented by this appeal from grant of summary judgment by the decedent’s estate.
I
On March 28, 2004, in the course of a routine traffic stop, City of Portland, Oregon Police Officer Jason Sery shot and killed James Jahar Perez, the driver of the stopped vehicle. Certain key facts surrounding the shooting are in dispute, but they are not relevant to this limited appeal. The district court, however, found a [964]*964number of facts to be undisputed, which we recite here to provide adequate context.
A
Sery and another officer, Sean Macom-ber were on a routine patrol in the St. John’s neighborhood of North Portland on Sunday afternoon, March 28, 2004, when Macomber noticed a white luxury sedan with tinted windows and chrome wheels that struck him as atypical “for cars driven in that working class neighborhood.” The officers were aware of local complaints of illegal drug activity, and drove by the car for a closer look. Upon running a registration check and learning that the car was registered to a man born in the 1950s, Macomber concluded that the age of the driver did not match. He also felt that the car’s two occupants “appeared nervous and did not want to make eye contact.”
As the officers drove by, the car remained stopped at a stop sign, leading Macomber to suspect that the driver was waiting to leave the area without being observed by the officers. After passing the car, the officers temporarily lost visual contact with it. When the officers regained sight of the car, the driver was now the sole occupant. The officers witnessed the driver signal and make a right turn into a strip mall parking lot, but it did not comply with Oregon traffic laws requiring vehicles to signal continuously for at least 100 feet prior to executing a turn. Ma-comber parked the patrol car behind the parked car, blocking it from any means of exit.
Although what transpired after the officers exited their patrol car and confronted Perez is disputed and awaits determination by a jury, it is undisputed that no more than 25 seconds elapsed from the time the officers left their patrol car until the time that Sery shot Perez. At the time of his death, Perez’s seatbelt remained fastened, and he was unarmed.
B
Gwen Price (“Price”), on behalf of Perez’s estate and his son, and Deborah Perez, sued Sery, Macomber, and the City of Portland (“City”) under 42 U.S.C. § 1983, alleging that the officers unconstitutionally used deadly force for which the City is liable under Monell v. Department of Social Services of the City of New York, 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978). In addition, the complaint alleged a state law claim for wrongful death based on negligent acts of the officers and the City, and other claims not relevant to this appeal.
Because the Portland Police Bureau’s (“PPB”) policy, training, and discipline practices with respect to the use of lethal force are relevant to the constitutional claims, we recite them as well. The use of deadly force is governed by PPB General Order (“G.O.”) § 1010.10, which reads in relevant part as follows:
The Bureau recognizes that members may be required to use deadly force when their life or the life of another is jeopardized by the actions of others. Therefore, state statute and Bureau policy provide for the use of deadly force under the following circumstances:
a. Members may use deadly force to protect themselves or others from what they reasonably believe to be an immediate threat of death or serious physical injury.
b. A member may use deadly force to effect the capture or prevent the escape of a suspect where the member has probable cause to believe that the suspect poses a significant threat of death or serious physical injury to the member or others.
[965]*965c. If feasible, some warning has been given.
Members must be mindful of the risks inherent in employing deadly force. A member’s reckless or negligent use of deadly force is not justified in this policy or State statute. Members are to be aware that this directive is more restrictive than state statutes.
G.O. § 1010.10.
C
Because we are reviewing a district court’s ruling on a motion for summary judgment, we must also consider facts alleged but not yet proven in order to decide this appeal.1 In particular, Price’s complaint makes allegations concerning the City’s history of disciplining officers for the inappropriate use of deadly force, as well as its training of officers. Price alleges specifically that, according to Portland Chief of Police Derrick Foxworth’s deposition, no Portland police officer has “ever” been “successfully disciplined” for shooting at an unarmed citizen “in the last twenty years.” But as District Judge Mosman’s opinion notes, Chief Foxworth’s second affidavit in the case, filed under seal, indicates that Chief Foxworth has demoted at least one officer for inappropriate use of deadly force, that former Chief Mark Kroeker disciplined an officer for unsatisfactory performance leading to the use of deadly force, and that former Chief Charles Moose terminated an officer for firing at a fleeing suspect without justification. The record also reflects that some decisions by the PPB to discipline officers have been overturned by arbitrators.
Price’s pleading further cited a report issued by the Police Assessment Resource Center (“PARC”) in August 2003. The findings of the PARC report, as presented by Price, did not reveal a failure to discipline officers but posited a need for improvement in the PPB’s approach to reviewing deadly force incidents. In addition to the PARC report, Price submitted the declaration of an expert in police tactics, Thomas Streed, Ph.D. Streed’s Declaration (“Streed Declaration”) asserts that, after reviewing 30 police shootings over the past 20 years, “at least fifteen” were not based on “probable cause.” Streed repeats the contention that no officer has “ever” been disciplined by the PPB for the use of lethal force, though he does acknowledge two unsuccessful attempts to discipline.
Price also made allegations concerning the City’s training of police officers in the use of deadly force. She relies on the Streed Declaration to contend that the City’s use of so-called “no-win” scenarios, together with its deadly force policy (also a part of training), creates a mind-set encouraging officers to “shoot first” and ask questions later. The Streed Declaration devotes just one paragraph to the City’s training program, and bases its conclusion that PPB training is “particularly” likely to lead to shooting of unarmed persons in significant part upon a distinction between “reasonable belief’ and “probable cause,” as used in the PPB G.O. § 1010.10 (quoted above).
D
Price moved for partial summary judgment on her Monell claims, and the City [966]*966moved for summary judgment on all claims. The district court denied Price’s motion, granted the City’s motion for summary judgment as to the Monell claims, and denied the City’s motion for summary judgment on the state law negligence claims. Pursuant to Fed.R.Civ.P. 54(b), the district court entered a final judgment as to the Monell claims, so that this appeal could be decided prior to a jury trial on the related unresolved issues.
II
Price offers three principal arguments that the City is liable under Monell for the alleged violation of Perez’s rights under the Fourth Amendment by the use of deadly force against him. We consider each in turn.
The Supreme Court has held that municipalities may be held liable as “persons” under § 1983 “when execution of a government’s policy or custom, whether made by its lawmakers or by those whose edicts or acts may fairly be said to represent official policy, inflicts the injury.” Monell, 436 U.S. at 694, 98 S.Ct. 2018. A plaintiff may also establish municipal liability by demonstrating that (1) the constitutional tort was the result of a “longstanding practice or custom which constitutes the standard operating procedure of the local government entity;” (2) the tortfeasor was an official whose acts fairly represent official policy such that the challenged action constituted official policy; or (3) an official with final policy-making authority “delegated that authority to, or ratified the decision of, a subordinate.” Ulrich v. City & County of San Francisco, 308 F.3d 968, 984-85 (9th Cir.2002).
Price first contends that, as written, the City’s official policy governing the use of lethal force by police officers violates the Fourth Amendment’s requirements, as explicated by the Supreme Court in Tennessee v. Garner, 471 U.S. 1, 105 S.Ct. 1694, 85 L.Ed.2d 1 (1985). Specifically, Price argues that the City’s policy, expressed in G.O. § 1010.10, that an officer “reasonably believe” a suspect poses an immediate threat of serious physical injury or death falls short of the “probable cause” requirement set forth in Gamer and this court’s precedents. 471 U.S. at 11-12, 105 S.Ct. 1694; Brewer v. City of Napa, 210 F.3d 1093, 1098(9th Cir.2000) (referring to Gamer’s “probable cause” deadly force standard as a “more specific and demanding standard” than Graham’s excessive force standard for the use of non-lethal force).
Price contends that the Supreme Court’s decision in Garner sets a requirement of “objective probable cause” for any use of deadly force by a police officer. Price notes that the City’s policy requires “objective probable cause” in the case of fleeing suspects, whereas for cases where an officer fears an imminent threat of death or serious physical injury, the City’s policy only requires that the officer “reasonably believe” that he or she is confronted by an immediate threat. Price argues that “reasonable belief’ is a different, and lesser, standard from “probable cause.”
Both Garner and Graham v. Connor, 490 U.S. 386, 109 S.Ct. 1865, 104 L.Ed.2d 443 (1989) are recognized as the leading Supreme Court cases explicating the requirements for the use of force by law enforcement officers under the Fourth Amendment. In Gamer, the Supreme Court for the first time considered the constitutionality of the common law rule permitting the use of lethal force to prevent the escape of a fleeing felon (but not of a misdemeanant). The Court held that [967]*967the Fourth Amendment’s reasonableness standard required “probable cause to believe that the suspect poses a threat of serious physical harm, either to the officer or to others” before using deadly force. Garner, 471 U.S. at 11, 105 S.Ct. 1694. In its lengthy analysis leading to this conclusion, the Court noted that the movement away from the common law rule in a number of states suggested that what was “reasonable” was not frozen in time with the common law. Id at 13-20, 105 S.Ct. 1694. In particular, the Court found it significant that police departments were “overwhelmingly” moving away from the common law rule. Id at 18, 105 S.Ct. 1694. Indeed, it quoted approvingly from the Commission on Accreditation for Law Enforcement Agencies’ language restricting the use of deadly force “to situations where the officer reasonably believes that the action is in defense of human life ... or in defense of any person in immediate danger of serious physical injury.” Id (internal quotation marks omitted and emphasis added).
The narrow question in Gamer was therefore whether, to justify the use of deadly force, an officer must believe only that a suspect is fleeing or also that the suspect represents a serious and immediate threat to the officer or others. In either case, the Court assumed that the belief would have to be reasonable, an inquiry that under the Fourth Amendment always depends upon objective factors and not upon sincerity of belief.
In Graham, the question was whether excessive force claims (a broader category than deadly force claims) should be analyzed as substantive due process claims or as Fourth Amendment claims. Graham, 490 U.S. at 388, 109 S.Ct. 1865. Chief Justice Rehnquist referred in a general way to the Fourth Amendment as requiring an “objective reasonableness” standard — not matching perfectly either the “reasonable belief’ or “objective probable cause” formulations. Id
The Graham Court clarified that the reasonableness inquiry turned upon the circumstances confronting the officer, rather than the officer’s subjective beliefs or intentions: “As in other Fourth Amendment contexts, however, the ‘reasonableness’ inquiry in an excessive force case is an objective one: the question is whether the officer’s actions are ‘objectively reasonable’ in fight of the facts and circumstances confronting them, without regard for their underlying intent or motivation.” Id at 397, 109 S.Ct. 1865.
The Gamer and Graham decisions are the Court’s leading cases bringing claims about the use of force — deadly or allegedly excessive — by law enforcement officers under the rubric of modern Fourth Amendment search-and-seizure analysis, rather than under the common law or substantive due process. Both cases focused on the “totality of the circumstances” and the “perspective of a reasonable officer on the scene, rather than with the 20/20 vision of hindsight.” Graham, 490 U.S. at 396, 109 S.Ct. 1865; see also Garner, 471 U.S. at 9, 105 S.Ct. 1694.
In this fight, the Supreme Court’s careful discussion of “probable cause” in a recent Fourth Amendment arrest context in Maryland v. Pringle, 540 U.S. 366, 124 S.Ct. 795, 157 L.Ed.2d 769 (2003), is instructive. There, Chief Justice Rehnquist, the author of Graham, wrote that “[t]he probable-cause standard is incapable of precise definition or quantification into percentages because it deals with the probabilities and depends on the totality of the circumstances” and “the substance of all the definitions of probable cause is a reasonable ground for belief.” Pringle, [968]*968540 U.S. at 371, 124 S.Ct. 795(internal quotation marks and citations omitted). Pringle does not address “probable cause” in the context of the use of deadly force by law enforcement officers. Nevertheless, the phrase “probable cause” itself should not mean one thing in one context and something different elsewhere. Rather, what an officer has probable cause to believe dictates the level of force he may justifiably use in a given scenario. In other words a law enforcement officer’s use of force will be justified, or not, by what that officer reasonably believed about the circumstances confronting him.
The Supreme Court very recently confirmed and clarified this analysis of the relationship between Gamer, Graham, and the Fourth Amendment’s reasonableness requirement in Scott v. Harris, — U.S. —, 127 S.Ct. 1769, 167 L.Ed.2d 686 (2007). In considering the reasonableness of a police officer’s use of likely deadly force to end a high-speed car chase, the Court noted that “Graham did not establish a magical on/off switch that triggers rigid preconditions whenever an officer’s actions constitute ‘deadly force.’ Gamer was simply an application of the Fourth Amendment’s ‘reasonableness’ test to the use of a particular force in a particular type of situation.” Id. at 1777 (citing Graham, 490 U.S. at 388, 109 S.Ct. 1865). The Court went on to state that “[wjhether or not [the police officer’s] actions constituted application of ‘deadly force,’ all that matters is whether [the officer’s] actions were reasonable.” Id. at 1778 (emphasis added). Accord Acosta v. Hill, 504 F.3d 1323, 1324 (9th Cir.2007) (holding that, under Scott, “there is no special Fourth Amendment standard for unconstitutional deadly force”); United States v. Gorman, 314 F.3d 1105, 1111 (9th Cir.2002) (“[W]e now hold that the ‘reason to believe,’ or reasonable belief standard ... embodies the same standard of reasonableness inherent in probable cause.”).
For this reason, Price’s attempt to use Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968), to establish some daylight between “reasonable belief’ and “probable cause” is unavailing. Terry permitted police officers to “stop and frisk” individuals where there is “reason to believe” that the individual is “armed and dangerous” even absent “probable cause to arrest the individual for the crime.” Id. at 27, 88 S.Ct. 1868. But Terry was not addressed to the quantity of evidence required for such belief, nor how much confidence the officer must have in it. Rather, the distinction was that a belief that an individual is armed and dangerous could justify a brief detention, whereas arrest requires a belief that the individual has committed a crime. Id. at 26-27, 88 S.Ct. 1868(“a perfectly reasonable apprehension of danger may arise long before the officer is possessed of adequate information to justify taking a person into custody for the purpose of prosecuting him for a crime”). The Terry Court made clear that the touchstone of the reasonableness inquiry was not the subjective strength of the officer’s belief, but its grounding in the objective facts: “[I]n determining whether the officer acted reasonably in such circumstances, due weight must be given, not to his inchoate and unparticularized suspicion or ‘hunch,’ but to the specific inferences which he is entitled to draw from the facts in light of his experience.” Id. at 27, 88 S.Ct. 1868.
That the phrases employed in the PPB’s General Order are not distinguishable along the lines Price suggests is also manifest in one of the lead eases upon which Price relies, Ting v. United States, 927 F.2d 1504, 1511 (9th Cir.1991). There, in order to conduct a qualified immunity analysis, we were required to describe the state of the law as a California police [969]*969officer would have known it in a pre-Garner context. We concluded that it “was generally established at the time of the shooting that an officer could use deadly force to effect an arrest if, under the circumstances, he reasonably believed such force was necessary to protect himself or others from death or serious bodily harm.” Ting, 927 F.2d at 1511 (emphasis added). There, we were relying upon a leading California case dealing with the use of lethal force by police officers, Kortum v. Alkire, 69 Cal.App.3d 325, 138 Cal.Rptr. 26 (1977). Ultimately, considering Gamer and Kortum together, we concluded that “under both the U.S. Constitution and California law, an officer may not use deadly force unless it is necessary to prevent the escape and the officer has probable cause to believe that the suspect poses a significant threat of death or serious physical injury to the officer or others.” Ting, 927 F.2d at 1513. We cannot fairly read Ting to stand for the proposition that there is a significant legal difference between the terms “reasonably believes” and “has probable cause to believe.”
Price draws our attention to several of our decisions to suggest that there is a difference between the two phrases employed in the PPB’s deadly force guidelines in G.O. § 1010.10(a) and § 1010.10(b): “reasonable belief’ and “objective probable cause.” We examine each in turn.
In Brewer v. City of Napa, 210 F.3d 1093 (9th Cir.2000), which involved the use of police dogs, we had to decide whether jury instructions in an excessive force claim were required to invoke Gamer’s “probable cause” language. In considering both Gamer and Graham, we stated that “the existence of probable cause, a more specific and demanding standard, was simply not relevant.” Id. at 1098. The opinion did not specify, however, what was “more specific and demanding” about the Gamer formulation.
We declined to hold in Brewer that a jury instruction should have required that the officer have probable cause to believe that the plaintiff was armed in order to justify the use of police dogs. Id. at 1097. Rather, we held that the less demanding standard from Graham, requiring only that the officer believe there was an “immediate threat to the safety of the officers or others” was sufficient. Id. at 1098. The “more specific and demanding” standard of Gamer is more specific and demanding in that it requires a specific belief — that a fleeing suspect poses a threat of death or serious physical harm.
Brewer was merely following this court’s precedent in requiring less of a showing of threat to justify the use of police dogs than to justify the use of deadly force. See, e.g., Vera Cruz v. Escondido, 139 F.3d 659, 661 (9th Cir.1997) (“[T]he Supreme Court in Gamer established a special rule concerning deadly force”) (overruled on other grounds, Smith v. City of Hemet, 394 F.3d 689 (9th Cir.2005) (en banc)). What is crucial is not the “objectivity” of the officers’ beliefs but the object of those beliefs. That is, when there is objective reason to fear for one’s safety (as in Brewer), but not one’s life, then force short of deadly force might be justified; to justify deadly force, an objective belief that an imminent threat of death or serious physical harm is required. In neither case would a merely subjective sense of threat justify the use of force; rather, the objectively describable totality of the circumstances would have to be such as to justify the use of force. Sincerely held but unreasonable belief does not justify the use of force under Garner, Graham, or our own precedents.
Nor do our cases turn on the quantum of evidence, as it were, necessary to justify the use of force by a police officer. All the [970]*970evidence in the world that a police dog might be necessary to locate a suspect (as in Brewer) would not justify the use of lethal force; not because the quantity of evidence would be insufficient, but because the kind of evidence would not justify the decision to use deadly force.2 It is the specification of the threat (as an immediate one involving death or serious physical harm), not the putative difference between “reasonable belief’ and “probable cause,” that controls. Thus, in Fikes v. Cleghorn, 47 F.3d 1011, 1014 n. 2 (9th Cir.1995), when we stated in a footnote that, “[wjhile the use of ‘force’ is reasonable under the Fourth Amendment if it would seem justified to a reasonable police officer in light of the surrounding circumstances, the use of ‘deadly force’ is only justified if the officer has probable cause to believe that a suspect poses a threat of serious physical harm to the officer or others,” we were specifying the kind of surrounding circumstances that uniquely justify the use of deadly force — a threat of serious physical harm or worse.3 The use of “probable cause” is, as in Pringle, synonymous with “reasonable belief,” and the use of deadly force is tied to the nature of the threat, not to the quantum of evidence required to believe in it.4
Another case invoked by Price actually decisively demonstrates our point. In Monroe v. City of Phoenix, 248 F.3d 851 (9th Cir.2001), we relied on the language from Brewer and Fikes, inter alia, to hold that a Gamer deadly force instruction was required where the use of deadly force was at issue, rather than a more general instruction on excessive force. Notably, we relied also upon our decision in Quintanilla v. City of Downey, 84 F.3d 353, 357 (9th Cir.1996), which itself considered whether the Graham formula had not simply subsumed the Gamer formula. We found the error in Monroe to be harmless because the jury had also been instructed that, under Arizona law, deadly force could only be used when “the police officer believes that deadly physical force is immediately necessary to protect himself against the other’s use or imminent use of unlawful deadly physical force.” 248 F.3d at 860. Because the jury found for the police officer on that claim, we held that “the jury must have concluded that Sgt. Sherrard had ‘probable cause to believe’ that Monroe ‘posed a threat of serious physical harm’ to him or others.” Id. (quoting Garner, 471 U.S. at 11, 105 S.Ct. 1694).
[971]*9717
We are satisfied that our case law does not support Price’s contention that “reasonable belief’ is a lesser standard than “probable cause” as a matter of law. Both standards are objective and turn upon the circumstances confronting the officer rather than on the officer’s mere subjective beliefs or intentions, however sincere. Our case law requires that a reasonable officer under the circumstances believe herself or others to face a threat of serious physical harm before using deadly force. Moreover, as the Supreme Court clarified in Scott, the touchstone of the inquiry is “reasonableness,” which does not admit of an “easy-to-apply legal test.” 127 S.Ct. at 1777-78. The City’s policy requires that an officer have a reasonable belief in an “immediate threat of death or serious physical injury” and thus comports with the requirement.
Accordingly, the district court correctly concluded that the City’s policy governing the use of deadly force was not, as written, contrary to the requirements of the Fourth Amendment.
Price next contends that the City’s policy, as interpreted by statements made in depositions by Chief Foxworth and the City’s legal arguments before the district court, suffices to sustain a Monell claim based on a “longstanding procedure,” even if the City’s policy as written is constitutional. See Adickes v. S.H. Kress & Co., 398 U.S. 144, 168, 90 S.Ct. 1598, 26 L.Ed.2d 142 (1970); see also Ulrich, 308 F.3d at 984-85.
Price emphasizes the City’s concession in legal arguments before the district court that “reasonable belief’ was a label for a standard “less specific and demanding” than “probable cause.” As the discussion above has shown, the City was tracking the language of Brewer and similar cases, which, while admittedly confusing in their references to the Gamer and Graham framework, did not actually distinguish between “reasonable belief’ and “probable cause” but between the circumstances justifying the use of deadly force and those justifying the use of lesser force.
Moreover, the City’s legal argument before the district court cannot, by itself, establish a “longstanding custom” related to the shooting at issue here. As the City rightly notes, this court is not bound by the concessions of parties concerning the meaning of the law. United States v. Ogles, 440 F.3d 1095, 1099 (9th Cir.2006) (en banc) (court “not bound by a party’s concession as to the meaning of the law, even if that party is the government and even in the context of a criminal case”).
Price specifically notes that in depositions related to this litigation, Chief Fox-worth stated that he believed there was a difference between “reasonable belief’ and “probable cause.” In the first instance, Chief Foxworth agreed with a question from an attorney that “there’s some difference in the amount of confidence that the officer needs to have before using deadly force for self-defense ... that that’s a slightly different standard than the probable cause standard where someone is escaping?” But in the same colloquy, when pressed as to whether a reasonable belief left “room for difference of opinion,” Chief Foxworth replied, “no, that’s why I think they’re pretty close.” Chief Foxworth’s descriptions of the need to look at the circumstances, rather than the subjective beliefs of the officer, are clear from the record.
That Chief Foxworth assumed there was some difference between the two different phrases applying to two different scenarios is nevertheless troubling, a point on which [972]*972we agree with our colleague in dissent. From the same deposition testimony, Price points to a question which suggests that Chief Foxworth thought the “reasonable belief’ standard to be subjective. The question and answer went as follows:
Q. All right. When we talk about subparagraph A, when we talk about reasonable belief, I’ve read a lot of the training material that the department has put out. And there’s discussion about subjective reasonable belief and objective reasonable belief, and I’m wondering what your understanding is of a situation. And let’s just talk about self-defense rather than defense of others.
If a police officer is in a situation where he or she personally, reasonably, honestly believes that they’re in a self-defense situation and they’re facing the immediate threat of death or immediate serious physical injury, is that sufficient to be within subparagraph A, or does it have to be some objective, you know, kind of imaginary objective officer who has the reasonable belief?
A. It’s the first one that you’ve described.
This exchange between the lawyer and Chief Foxworth is more ambiguous than Price suggests, since the question itself includes both subjective and objective elements. The question includes the term “reasonably” in the first scenario, and tracks the language that “they’re facing the immediate threat of death or immediate serious physical injury.” Thus, it is possible that Chief Foxworth was not embracing a subjective standard at all.
To dispose of a case on summary judgment, however, ambiguity in favor of the defendant is not sufficient. If a reasonable person could side with the plaintiffs interpretation of events, the issue must survive for trial. We conclude that there is a genuine issue of material fact as to whether the City’s interpretation of the differing phrases in G.O. § 1010.10(a) and § 1010.10(b) represents the sort of “longstanding” custom or practice that can establish Price’s Monell claim even though the formal written policy does not.
Price’s arguments about the application of the City’s deadly force policy are not limited solely to the interpretation of the policy. She also contends that the City has failed adequately to discipline officers for the inappropriate use of deadly force, and has trained them in such a fashion as to lead to the unjustified use of deadly force. Price’s case depends largely upon the Streed Declaration, which in turn relies upon the purported difference between “reasonable belief’ and “objective probability” discussed above. It is hard to know how much weight to give an expert report that seems, as the district court noted, to incorporate that assumption throughout its own analysis. In addition to the Streed Declaration, however, other evidence in the record, such as the PARC report, could support Price’s claims.
The district court was unwilling to give the Streed Declaration sufficient weight to survive summary judgment once it found error in the legal distinction relied upon therein. Further, the district court concluded that the City’s unsuccessful attempts at discipline should be credited, and could not serve as evidence that “no officer had been disciplined.” While the district court’s conclusions concerning this evidence may well be reasonable and even persuasive, those are arguments for trial, and do not comport with the obligation of the court, on summary judgment, to draw all inferences in favor of the nonmoving party. Thus, we are unable to conclude that no rational trier of fact could agree with Price’s interpretation of the City’s history of discipline and training and its [973]*973relevance to an allegedly unconstitutional “longstanding” use of deadly force in situations where the objective facts did not support such force.
Finally, Price contends that the City’s alleged failure to train its police officers appropriately as to the use of deadly force amounts to a constitutional violation by itself. The “inadequacy of police training may serve as the basis for § 1983 liability only where the failure to train amounts to deliberate indifference to the rights of persons with whom the police come into contact.” City of Canton v. Harris, 489 U.S. 378, 388, 109 S.Ct. 1197, 103 L.Ed.2d 412 (1989). Price argues that her failure-to-train claim meets the standard set by Harris and should itself suffice to make out a Monell claim against the City.
Under Harris and progeny, one must demonstrate a “conscious” or “deliberate” choice on the part of a municipality in order to prevail on a failure to train claim. See Blankenhorn v. City of Orange, 485 F.3d 463, 484 (9th Cir.2007). Harris’s standard is objective in that it does permit a fact finder to infer “constructive” notice of the risk where it was “obvious” — but this is another way of saying that there needs to be some evidence that tends to show a conscious choice. See Farmer v. Brennan, 511 U.S. 825, 841, 114 S.Ct. 1970, 128 L.Ed.2d 811 (1994) (“It would be hard to describe the Canton understanding of deliberate indifference, permitting liability to be premised on obviousness or constructive notice, as anything but objective.”).
Price points to nothing in the record or in the Streed Declaration that evinces the deliberate indifference that Harris requires for a free-standing failure-to-train claim to succeed. The City’s description of the many hours of training required for PPB officers is undisputed. Given our conclusion that the City’s policy on the use of deadly force is constitutional as written, the undisputed fact that the City trains according to that policy does not advance Price’s argument. Considering the high burden laid out in Harris, the district court was correct to find that even if the City’s training may not have been ideal, Price offers nothing that would establish the kind of “conscious” or “deliberate” choice by the City to risk a “likely” violation of constitutional rights. See Harris, 489 U.S. at 389, 109 S.Ct. 1197.5
Accordingly, construing the evidence in the light most favorable to Price, we conclude that no rational finder of fact could decide that the City violated Perez’s constitutional rights by a failure to train. The district court properly granted summary judgment to the City on the claim that its training practices amount to a violation under Harris.
Ill
For the foregoing reasons, we agree with the district court that the City’s official policy concerning the use of deadly force, as written, does not violate the requirements of the Constitution. Further, we agree with the district court that Price has not made a sufficient showing of a failure to train on the part of the City to survive summary judgment. We conclude, however, that a genuine issue of material [974]*974fact exists as to whether a “longstanding” practice or custom of the City might in fact have deprived Perez of his constitutional rights.
The decision of the district court is therefore
AFFIRMED in part, REVERSED in part, and REMANDED.