dissenting.
A reader of the majority opinion would be forgiven for thinking that the death of plaintiffs decedent was the tragic result of the kind of gang-, drug-, or alcohol-related violence that plaintiff alleges to be common in the neighborhood around the Zone nightclub. But that is not what happened here, and plaintiff does not allege otherwise. Instead, decedent was shot during an attempted mass murder by a violent schizophrenic assailant, Ayala, who opened fire simply because he was intent on killing teenagers. Plaintiff does not allege that Ayala was a gang member, a drug dealer, or a club-goer or had any other foreseeable connection to the Zone or the Zone’s neighborhood.
The majority opinion, nevertheless, concludes that plaintiffs allegations of gang-, drug-, or alcohol-related violence could establish that the shooting incident in this case was foreseeable because both the alleged neighborhood violence and the shooting at issue here are types of “violent *96assaults.” 360 Or at 83. But the majority fails to explain why that similarity matters. Contrary to the majority’s analysis, it does not follow as a matter of law or as a matter of logic that, simply because some types of violent assault are foreseeable, therefore all types of violent assault are foreseeable. I would limit the types of foreseeable risks in this case to those that plaintiff has alleged to be present at the time and place of the shooting — namely, gang-, drug-, and alcohol-related violence, which might include some shooting incidents but would exclude the shooting incident in this case. Defining the foreseeable risks any more broadly than that, as the majority does, goes beyond the risks that the complaint alleges. For those reasons, I dissent.
The principal issue on review is whether plaintiff has alleged facts that would allow a reasonable jury to conclude that the shooting incident underlying this negligence case was a type of foreseeable risk created by defendants’ conduct. According to plaintiff, the Zone nightclub had a history of violence, because a shooting happened there in 2002 — seven years before the shooting incident here — and because underage nightclubs are inherently dangerous, due to the high proportion of young male patrons, high noise levels, crowding, and “competitive situations.” Further, plaintiff alleges that “assaults” occur regularly in the area around the Zone because of gangs, drug dealing, and over-consumption of alcohol. Those facts, plaintiff alleges, were known, or should have been known, to defendants. As a result, plaintiff contends that a risk of violent assault was foreseeable. Plaintiff concludes, therefore, that the Rotary defendants were negligent for leaving the group of teenagers, including decedent, at the Zone, and that the Zone defendants were negligent for asking patrons to form a line outside the club and failing to provide more and better security.
Plaintiffs allegations are unquestionably thin and bear little relationship to the unprovoked shooting in this case. That is particularly true with regard to the Rotary defendants. Plaintiff alleges only that the Rotary defendants “knew, or in the exercise of reasonable care, should have known, about the dangers of leaving plaintiffs decedent in the Zone nightclub on the date and time in question.” Although “the fact that the defendant was aware of a *97particular risk” is usually a factual allegation rather than a legal conclusion, Moore v. Willis, 307 Or 254, 259, 767 P2d 62 (1988), it is not clear to me that plaintiffs allegation of “danger” is an allegation of a particular risk. And although plaintiff alleges that “instances of violence that occurred nationally at underage nightclubs and locally in the area around the Zone nightclub were all publicized in local and/or national media,” plaintiff does not allege that the Rotary defendants were aware of those “local and/or national” news reports, nor any reason why the Rotary defendants — or their alleged agents, decedent’s host family in White Salmon, Washington — should have been aware of those reports. Reports of violence are commonplace. Local and national media are so saturated with crime stories that an ordinary person cannot be expected to take note of all violent acts wherever they might travel and account for them in assessing the risks of engaging in daily life. Foresight of that kind would demand “a paranoid view of the universe,” and that is not what the legal standard of foreseeability requires. Fazzolari v. Portland School Dist. No. 1J, 303 Or 1, 21, 734 P2d 1326 (1987).
Nevertheless, I do not believe that that issue needs to be resolved because, even assuming that the Rotary defendants had the same knowledge as the Zone defendants, those allegations are insufficient to establish that the shooting incident underlying this case was foreseeable. We generally make that determination by comparing the harm that a plaintiff suffered to the general types of risks foreseeably created by a defendant’s conduct. Id. at 17 (describing the standards of foreseeability). The trial court ruled that plaintiffs allegations, even if true, could not establish that the shooting incident was foreseeable and therefore dismissed plaintiffs second amended complaint under ORCP 21 A(8) for failing to state ultimate facts supporting her negligence claim against both the Zone defendants and the Rotary defendants.
When reviewing a trial court’s ruling dismissing a plaintiffs complaint under ORCP 21 A(8), we assume, like the trial court, that the facts alleged in the complaint are true, and we draw all reasonable inferences in the plaintiffs favor. Bailey v. Lewis Farm, Inc., 343 Or 276, 278, 171 P3d *98336 (2007). The majority agrees that, when assessing the sufficiency of a plaintiffs allegations to state a claim, it is the court’s role to “‘determine whether upon the facts alleged *** no reasonable factfinder could decide one or more elements of liability.’” Chapman v. Mayfield, 358 Or 196, 205, 361 P3d 566 (2015) (quoting Fazzolari, 303 Or at 17).
Foreseeability is central to establishing the elements of a negligence claim. A defendant may be subject to liability for negligence only if the defendant’s conduct “unreasonably created a foreseeable risk * * * of the kind of harm that befell the plaintiff.” Fazzolari, 303 Or at 17. Therefore, a plaintiff fails to state a negligence claim if the plaintiff fails to allege facts that would allow a reasonable jury to find that the plaintiff was harmed by a foreseeable risk created by the defendant’s unreasonable conduct. Solberg v. Johnson, 306 Or 484, 490-91, 760 P2d 867 (1988) (stating that the elements of a negligence claim include “that defendant’s conduct caused a foreseeable risk of harm, * * * that defendant’s conduct was unreasonable in light of the risk, [and] * * * that plaintiff was within the class of persons and plaintiffs injury was within the general type of potential incidents and injuries that made defendant’s conduct negligent”).
As a result, defining the scope of foreseeable risks has important implications for tort law generally and particularly when liability will be imposed for negligent conduct. A risk is foreseeable if a reasonable person would have “reasonably expected the injury to occur.” Stewart v. Jefferson Plywood Co., 255 Or 603, 609, 469 P2d 783 (1970). If the scope of foreseeable risks is described more broadly, then that necessarily means that an actor should expect his or her conduct to lead to more or greater harms. A more broadly defined risk makes it easier to establish both that a defendant’s conduct is unreasonable (because the probability and severity of foreseeable harms are factors in determining reasonableness) and that a plaintiffs harm falls within the risks created by the defendant’s unreasonable conduct (because the risk encompasses more harms). The opposite is true when the foreseeable risks are defined narrowly.
Although foreseeability is a component of both the reasonableness determination and the harm-within-the-risk *99determination, defendants do not challenge the sufficiency of plaintiffs allegations to establish that defendants’ conduct was unreasonable. Thus, for the sake of reviewing the trial court’s dismissal order, we assume that the Rotary defendants should not have taken the teenagers to the Zone. And we assume that the Zone defendants should not have required patrons to form a line and should have provided more and better security.
Instead of challenging the reasonable-conduct element of plaintiffs negligence claims, defendants challenge plaintiffs pleading of the element requiring that decedent’s harm was within the foreseeable risks created or increased by defendants’ conduct. See, e.g., Chapman, 358 Or at 206 (considering “whether plaintiffs’ injuries were within the type of potential harms that made defendant’s conduct unreasonable”; affirming summary judgment for defendant); Bailey, 343 Or at 281 (reversing dismissal where “[t]he type of harm that plaintiff suffered fell squarely within the scope of the risk that defendant’s negligence created”).1 That element identifies the legal limit of liability for negligently caused harms — that is, the point at which we will not hold a defendant liable for harms, even if the defendant acted unreasonably and even if the defendant’s conduct was a but-for (or substantial-factor) cause of the injury. See, e.g., Stewart, 255 Or at 607 (noting defendant’s negligence and causation but attempting to “draw[] the line at which the defendant’s liability ends”).
As it relates to that element, overly broad and overly narrow categories of risk distort the limits of liability:
“It has been observed that ‘if we use a very generalized description of the type of harm that was foreseeable and of the type of harm that occurred, an answer that the result was within the risk is inevitable.’ And on the other hand, ‘(i)f we use a detailed, mechanism-of-harm description of the result and the risks, the answer will be negative.’”
*100Id. at 610 (quoting Robert E. Keeton, Legal Cause in the Law of Torts, 51 (1963)) (footnotes omitted); see also Chapman, 358 Or at 208 (same) (citing Stewart, 255 Or at 610). Risks should not be defined so narrowly as to require the defendant to foresee the precise sequence of events, as they unfolded, that lead to the particular plaintiffs particular injury. See Fazzolari, 303 Or at 21 (so stating). Instead, we focus on generalized risks of harm. Id. at 13 (citing Stewart, 255 Or at 610-11). My disagreement with the majority, therefore, is not whether courts should use “generalized” descriptions of risk; our cases correctly require us to do so. My disagreement, rather, is over how broadly to define the generalized risks that we use in this case. As Stewart itself recognized, a generalized risk of harm may be defined in terms that are too general. Stewart, 255 Or at 610. Because liability is intended to track community standards for blameworthiness, risks must be characterized with sufficient particularity to distinguish between conduct that is blameworthy and conduct that is not. Id. at 608-09 (so stating). Risks that are defined too broadly — where a finding of foreseeability is “inevitable,” id. at 610 (internal quotation marks and citation omitted)— cannot serve as a meaningful limit on the scope of liability. See Fazzolari, 303 Or at 21 (so stating).
There is no single standard for determining the appropriate level of generality for defining risks. Last year, in Chapman, this court emphasized the importance of appropriately identifying how generally to define risks. 358 Or at 208. To do so, “this court’s practice in cases that address foreseeability as a limit on liability” has been to “[d]raw[] on the plaintiffs theory of liability” and to “view[] the defendant’s conduct through the lens of the particular factual circumstances of the case — with emphasis on what the defendant knew or should have known about the risk of harm.” Id.
For example, in Stewart, the court addressed the appropriate level of generality for defining the foreseeable risks created by negligent welding operations. In that case, there was no question that the defendant’s unreasonable conduct started a fire that spread to a neighboring warehouse. Stewart, 255 Or at 605. While attempting to prevent further fire damage on the roof of that warehouse, the plaintiff fell *101through a covered skylight, hit the floor, and suffered various injuries. Id. Thus, there was also no question that defendant’s conduct was a factual cause of the plaintiffs injury. But the court noted that establishing that the defendant’s unreasonable conduct caused the plaintiffs injury is not sufficient to establish liability, because a defendant is not liable for all injuries caused by his or her unreasonable conduct. Id. at 606. Instead, a defendant is subject to liability only for injuries that can be “anticipated.” Id. at 609.
The court concluded that the plaintiffs injury could be anticipated because a reasonable person would know that firefighters “might be injured as a result of a variety of possible circumstances — by being burned, by falling off a ladder, by falling off a roof, by falling through a burned portion of the roof, or by other similar risks normally associated with fighting fires.” Id. at 610. Although falling through a covered skylight might not immediately come to mind as a potential harm, it was foreseeable because it was within the types of risks that a reasonable person would reasonably expect a firefighter to encounter. Id.
The plaintiffs theory and supporting evidence were related to firefighting, so the risk of harm was defined in reference to firefighting. Based on that theory and that evidence, it would have been inappropriate, for example, to define the risk as being “falling through covered skylights.” That risk would have been too narrow because it would have required the plaintiff to establish that defendant had specific knowledge that the skylight was covered. That risk would have also been too broad because it would encompass risks outside the plaintiffs theory of liability and the evidence offered. Hypothetically, bystanders might have gone to the roof of a warehouse upwind from the fire simply to watch the activity. If a bystander on that warehouse fell through a covered skylight, the defendant’s negligent conduct still might have been a factual cause of the resulting injuries. But the foreseeability of that injury could not have been established through evidence of the risks encountered by firefighters.
In this case, plaintiff attempts to establish that the shooting incident was foreseeable based on defendants’ *102knowledge of prior criminal acts at or around the Zone. As noted above, we assume that defendants’ conduct was unreasonable, and, therefore, we assume that they had knowledge of prior criminal acts at or around the Zone that would cause them to be aware of some continuing risk of further criminal acts. The question is whether the shooting incident in this case was among the criminal acts that a reasonable person would foresee based on those prior criminal acts allegedly known to defendants.
Answering that question requires comparing those prior criminal acts to the shooting incident in this case. The majority identifies a number of factors relevant to that comparison — namely, the degree of similarity, frequency, recency, and geographic proximity between the plaintiffs harm and the prior criminal acts alleged by the plaintiff. 360 Or at 81. And, as the majority notes, similarity may be based on the nature of the criminal act itself, or it may be based on other considerations, such as the circumstances of the offense or the specific types of victims.
Thus, a generalized risk does not need to be defined so broadly as to capture all of the prior criminal acts that a plaintiff has alleged. Instead, the risk is based only on those prior criminal acts that indicate a continuing risk of injury at the time and place of a defendant’s allegedly negligent conduct and a plaintiffs resulting harm. In other words, the foreseeability of a plaintiffs injury cannot be based on prior criminal acts that are too dissimilar, are too remote in time or place, or otherwise fail to indicate a continuing risk at the relevant time and place.2 Prior criminal acts that are dissimilar to the criminal act that caused the plaintiffs injury might have signaled a continuing risk of some other type of harm, but not the risk of harm that the plaintiff suffered. And prior criminal acts that are too remote in time *103and place might have signaled a continuing risk of harm at some other time or some other place, but not at the time and place of the plaintiffs injury. What counts as dissimilar or too remote will depend on the plaintiffs theory of liability and factual allegations.
For example, in Fazzolari, a female student was sexually assaulted on school grounds. 303 Or at 3. In assessing the foreseeability of the sexual assault, the court noted that there was evidence of a prior sexual assault on another female at the same school only 15 days earlier. Id. at 21. Although the court in Fazzolari did not expressly apply the factors that the majority now identifies, that single prior criminal act established recency, geographic proximity, and similarity of the criminal act sufficient to allow the issue to go to the jury. In considering evidence of “various other kinds of attacks,” the court noted that “evidence of foreseeability will differ depending on whether the risk of injury is claimed to be specific to a school, or schools generally, or a neighborhood, or a class of potential victims such as women or particular ethnic groups.” Id.
Similarly, we start by assessing decedent’s injury in this case and then compare it to the prior criminal acts allegedly known by defendants. According to plaintiff, the shooting incident in this case occurred when Ayala approached a group of teenagers waiting in line at night outside the Zone and opened fire with a semiautomatic handgun with the intent to kill them. As pleaded, Ayala was a paranoid schizophrenic who was motivated to commit the shooting by an irrational and hostile fixation with teenagers.3
The prior criminal acts alleged by plaintiff to be known by defendants relate to the Zone itself and to the neighborhood around the Zone. As it relates to the Zone, plaintiff alleges that underage nightclubs are inherently dangerous because many patrons are young males in “competitive *104situations” and clubs are often noisy and crowded. But prior incidents of teenage fights are plainly insufficient to create a reasonable expectation of an unprovoked shooting. Further, plaintiff relies on a shooting in 2002 that injured three people outside the location later occupied by the Zone. Although that shooting is a similar criminal act in some ways, it happened seven years before the shooting in this case — nothing like the 15 days that separated the crimes in Fazzolari. Because of that lack of recency, we cannot simply assume that the causes of the 2002 shooting persisted and continued to pose a risk of another shooting there in 2009. And plaintiff alleges no additional facts that could establish that the 2002 shooting created a reasonable expectation of a shooting seven years later.
As it relates to the area around the Zone, plaintiff alleges that violent assaults occur regularly because of gangs, drug dealing, and over-consumption of alcohol. To substantiate that, plaintiff provides crime statistics from an expansively drawn area, covering the entire Downtown and Old Town/Chinatown neighborhoods. Plaintiff, however, makes no allegations addressing the extent to which those crimes actually happened near the Zone, affected Zone patrons, or otherwise resembled the unprovoked shooting in this case. In other words, in an attempt to establish recency and frequency, plaintiff has stretched the bounds of similarity and geographic proximity.
Nevertheless, the majority insists that a reasonable jury would be allowed to conclude that defendants should have foreseen a risk of “violent assault.” 360 Or at 83. The majority does not further limit its characterization of that foreseeable risk, which, therefore, covers all types of violent assault. But the expansiveness of that risk proves that it is untenably broad. A risk of “violent assault” would cover nearly all types of violence, including the shooting incident in this case, but also including sexual assaults, armed robberies, bombings, and even acts of terrorism — each of which is a type of violence that would not have been foreseeable based on plaintiffs allegations. If plaintiff had pleaded that the shooting incident occurred in a war zone, then we might expect a reasonable person to anticipate the wide range of harms that fall within the category of “violent assaults.” *105But plaintiffs allegations are far more limited than that. Consequently, the foreseeable risk that may be established by plaintiffs allegations requires some limitation.
That limitation separates the types of violence that could have been reasonably expected from the types of violence that could not have been reasonably expected. As explained above, many of plaintiffs allegations are too remote to indicate a continuing risk at the time and place of the shooting.4 Focusing only on plaintiffs allegations that indicated a continuing risk of injury at the time and place of the shooting, plaintiffs allegations could support, at most, a conclusion that defendants should have foreseen some types of violent assaults — namely, those types of violent assaults that plaintiff alleged to be common in the neighborhood at the time of the shooting: gang-, drug-, and alcohol-related violence. Plaintiffs allegations of prior gang-, drug-, and alcohol-related crimes indicated a foreseeable risk of harm at the time of the shooting because, as alleged by plaintiff, gang-, drug-, and alcohol-related activity continued to persist in the area around the Zone at that time.
The foreseeable risk in this case is defined by the type of criminal — i.e., whether the criminals whose activities posed the allegedly foreseeable risks of harm are connected to gang-, drug-, or alcohol-related activities — rather than the type of crime, because plaintiffs allegations provide no other grounds for defining the risk. There will likely be many cases where the type of criminal will carry less weight in the foreseeability analysis, such as cases where the prior criminal conduct is more similar, recent, frequent, and geographically proximate than those alleged by plaintiff. In those cases, it may be reasonable to infer that the cause of the prior crimes persisted at the time and place of the later crime. But, given the lack of similarity, recency, frequency, and geographic proximity, that inference is not reasonable based on plaintiffs factual allegations. See Delgado v. Souders, 334 Or 122, 135, 46 P3d 729 (2002) (whether an inference is reasonable is a question of law). Further, plaintiff makes no allegation that the Zone specifically attracted *106people wanting to commit unprovoked shootings or people directing violence at teenage victims. Although the majority notes that Ayala targeted the Zone because he sought out teenage victims, 360 Or at 82, 89-90, plaintiff does not allege that Ayala’s intention to shoot teenage victims was known to defendants prior to the shooting incident and, therefore, cannot provide grounds for establishing foreseeability.
Once the risk posed by the alleged prior criminal acts is properly defined as gang-, drug-, and alcohol-related violence, it is apparent that decedent’s harm was not within the foreseeable risk: Plaintiff makes no allegation that Ayala was involved in gang-, drug-, or alcohol-related activities or had any connection to gang members, drug dealers, or club-goers in the neighborhood. Further, eliminating the risk of gang-, drug-, and alcohol-related activity in the area around the Zone would have had no effect on the likelihood of the shooting incident in this case, plainly establishing that the risks are distinct. To be sure, gang-, drug-, and alcohol-related violence are themselves broadly defined and generalized risks of harm. But to define the foreseeable risks more broadly than that would improperly divorce the risk of harm from plaintiffs theory of liability and from the factual circumstances alleged by plaintiff to be known to defendants.
In reaching its conclusion that all violent assaults in the geographic area that plaintiff identified were foreseeable, the majority fails to consider the similarity, recency, frequency, and geographic proximity of the prior criminal acts. Instead, the majority’s analysis appears to first go off track by reading more into our prior cases than those cases can support. According to the majority, “This court’s previous decisions generally prescribe a relatively broad level of generality in assessing the reasonable foreseeability of criminal conduct as a limit on liability.” 360 Or at 83 (citing Fazzolari, 303 Or at 21).
The problem is that, as noted above, when this court most recently addressed the process for defining the foreseeable risks, that process did not “prescribe a relatively broad level of generality,” as the majority now asserts. 360 Or at 83. Instead, after reviewing the same case law *107that the majority reviews in this case, we concluded that the process for appropriately identifying how generally to define the risks requires critically assessing the connection between plaintiffs theory of liability and the facts alleged to be within defendants’ knowledge. Chapman, 358 Or at 208. As a result, our prior cases in fact do not exhibit a preference for either a broad or narrow level of generality. Rather, the appropriate level of generality turns on plaintiffs theory of liability and the facts alleged to be within defendants’ knowledge. Whether a different theory of liability and different facts alleged in a different case supported a broad level of generality is immaterial.
In any event, I do not share the majority’s reading of Fazzolari as prescribing that all types of assaults be treated as one category of foreseeable risks. That decision never defines an appropriate level of generality. Although Fazzolari states that “a school’s responsibility for students’ safety against assault is not limited to the risk of rape,” 303 Or at 21, that statement goes to assessing the reasonableness of the school’s precautions. The more foreseeable assaults that might be prevented through a particular safety precaution, the more likely that it was unreasonable for the school not to take that precaution. As the court noted earlier in its opinion in Fazzolari, “[I]t is important that all foreseeable risks of harm to other classes of persons be considered in evaluating the reasonableness or unreasonableness of defendant’s conduct.” Id. at 14. But simply because all foreseeable assaults should be considered when determining whether conduct is reasonable does not mean that all assaults are foreseeable.
The majority attempts to justify the broad level of generality by explaining that location-based risks are uniquely suited to broadly defined foreseeable risks. The majority insists that it is appropriate to use a broad level of generality
“where, as here, the risk of harm at a particular location is at issue. See, e.g., Uihlein, 282 Or at 641 (‘[T]here must be something to alert the storekeeper to the likelihood of harm of some kind from a criminal agency.’). That only makes sense. When a business owner is on notice that — in *108the absence of reasonable precautions — assaultive conduct can be anticipated where people are waiting in line to enter its premises, it rings hollow to suggest that a specific crime pattern triggers the owner’s duty to protect patrons only with regard to those specific crimes (such as gang-related homicide) that have occurred.”
360 Or at 83. But there is no question that a business owner must take reasonable precautions to protect invitees from assaultive conduct reasonably expected to occur. The question is whether the harm that decedent suffered was among those types of assaultive conduct that defendants should have reasonably expected to occur.
In answering that question, the majority refuses to distinguish between criminal conduct based on the characteristics of the shooter:
“To the extent that [defendants’] challenge focuses on the precise sequence of events that occurred in this case, defendants invite a taxing level of comparison that this court consistently has rejected. As noted, contrary to defendant’s focus, foreseeability has reference to ‘the generalized risk of the types of incidents and injuries that occurred rather than the predictability of the actual sequence of events.’ See Fazzolari, 303 Or at 13 (discussing Stewart).”
360 Or at 81-82. In effect, the majority concludes that an assailant’s methods and motivations are irrelevant to the foreseeability analysis. But, as described above, plaintiffs allegations make them relevant here because only gang-, drug-, and alcohol-related crimes could have given rise to a reasonable expectation of injury at the time and place of the shooting.
From the majority’s perspective, that does not matter because gang-, drug-, and alcohol-related violence entailed a risk of a violent assault and a violent assault happened. But, although we do not require plaintiffs to allege or adduce facts establishing that the precise sequence of events was foreseeable, we also do not focus only on the end result — e.g., a violent assault — while ignoring the manner in which the harm occurred. A foreseeable risk cannot be defined without connecting the end result to the reasons that the risk was foreseeable in the first place. Therefore, “identical injuries, *109if they occur in different ways or at different times, may be treated differently.” Restatement (Third) of Torts: Phys. & Emot. Harm § 29 comment d (2010).5
That principle is well-established in our case law. For example, in Hefty v. Comprehensive Care Corporation, 307 Or 247, 766 P2d 1026 (1988), the plaintiffs were a minor who had been a patient at the defendant’s alcohol treatment center as well as her parents. Id. at 250. The defendant prematurely discharged the minor, who was still suffering from alcoholism, without notifying her parents or doctor or otherwise providing alternative arrangements for her care. Id. The day after the discharge, the minor was involved, as a passenger, in a traffic accident that did not involve drugs or alcohol. Id. at 251.
Plaintiffs sued the defendant for negligently discharging the minor. The court resolved the case on foreseeability, asking whether the plaintiffs “injuries fell within the scope of the ‘generalized risk of the types of incidents and injuries’ created when defendant discharged [the plaintiff] .” Id. The court appropriately identified two generalized types of foreseeable risks created by the defendant’s conduct: (1) that the minor would “resume abusing alcohol and drugs”; and (2) that the minor “would engage in activities consistent with her impaired judgment and inability to control her behavior.” Id. at 252.
Notably, the court did not define the risk by the end result — a traffic accident — even though it was at least plausible that either of those risks could lead to a traffic accident. If, for example, the minor had been driving and was under the influence of alcohol, then the accident might have *110fallen within the scope of the risk that she would resume abusing alcohol. But the fact that defendant’s conduct created a risk that could lead to a traffic accident and the fact that a traffic accident occurred was insufficient, because the traffic accident that was risked was a different type of traffic accident than the one that occurred. The issue was not whether the defendant’s unreasonable conduct contributed to the traffic accident — that would merely establish factual causation. The issue was whether the reasons that made the defendant’s conduct unreasonable were related to the traffic accident that occurred. The court denied liability as a matter of law because the plaintiffs adduced no evidence establishing that the minor’s substance abuse or impaired judgment were related to the traffic accident. The court held, “‘In an extreme case a court can decide that no reasonable factfinder could find the risk foreseeable.’ This is an ‘extreme case!” Id. at 253 (quoting Donaca v. Curry Co., 303 Or 30, 38, 734 P2d 1339 (1987)).
Similarly, in Oregon Steel Mills, Inc. v. Coopers & Lybrand, LLP, 336 Or 329, 83 P3d 322 (2004), we considered the liability of an accountant who negligently misstated a steel company’s earnings, delaying the company’s stock offering, through which the company intended to raise capital. Id. at 333. By the time the company’s stock was offered, the market for steel company stocks had declined. As a result, the company’s stock sold for less than it would have if it had not been delayed by the accountant’s negligence, and the company was unable to raise as much capital as it would have otherwise. Id.
As in Hefty, the issue in Oregon Steel Mills was not factual causation; instead, the issue was foreseeability. Id. at 344 (“There is sufficient evidence of factual causation * * *. Rather, the critical issue is whether plaintiffs market losses were a reasonably foreseeable result of defendant’s wrongful conduct.”).6 Again, the court did not define the risk as the end result — stock-price reduction and resulting *111decrease in the capital raised — even though an accountant’s negligence entailed such a risk. For example, negligently misstating earnings might understate a company’s prior earnings, making the company’s stock less desirable and impairing the ability to raise capital. Then, there would be a connection between the accountant’s negligence and the stock-price reduction that impaired the company’s ability to raise capital.
But the fact that the accountant created a risk of a lower stock price and the fact that a lower stock price occurred was not sufficient, because the stock-price reduction that was risked was a different type of stock-price reduction than the one that occurred. Instead, liability required establishing that the lower stock price occurred in the manner that was reasonably anticipated. This court denied liability as a matter of law because the company’s stock price was not lowered by the accountant’s negligence; it was lowered by market fluctuations that affected the entire steel industry and had nothing to do with the accountant’s negligence. Id. (“ [P]laintiff seeks damages based solely on a decline in the price of plaintiffs stock during the delay that defendant caused in getting the offering to market, yet plaintiff admits that the price decline affected all steel stocks and was unrelated to defendant’s misconduct.”).7
Other courts have likewise distinguished between the manner of the harm and the end result under facts similar to this case. For example, Shadday v. Omni Hotels Mgmt. Corp., 477 F3d 511 (7th Cir 2007), involved a negligence claim against a hotel for failing to prevent the sexual assault of the plaintiff, who was a hotel guest. Id. at 511. The sexual assault was committed by another hotel guest. Id. In *112denying liability as a matter of law, Judge Posner explained that, based on the evidence adduced, the risk of crime was from intruders and not from hotel staff or guests. Id. at 516. Although “the precautions against guest-on-guest crime are not dramatically different from the precautions against intruder crime” such that the hotel’s failure to take those precautions might have been a but-for cause of the guest’s injury, that fact did not establish liability. Id. at 517. Instead, regardless of the risk of sexual assault presented by intruders, “[t]he hotel becomes liable for guest-on-guest crime only when it has some reason to think such crime likely,” and the plaintiff presented no such evidence in that case. Id.
It may be worth asking why the manner of harm matters when a defendant’s conduct is unreasonable and the end result is the same, regardless of the manner in which it occurs. Judge Posner addressed that question in Shadday: “[T]he puzzle of the line of cases that [require the harm to be within the foreseeable risk] is why the defendant, having been negligent, should get off scot-free just because the harm that would have been averted had he been careful was not foreseeable.” Id. at 518. He notes that, without requiring the harm to be within the risk, “negligence liability is potentially too encompassing.” Id.
Our case law reflects those concerns as well. Liability is intended to turn on fault, and fault is predicated on blameworthiness. Stewart, 255 Or at 608. “[T]he community deems a person to be at fault only when the injury caused by him is one which could have been anticipated because there was a reasonable likelihood that it could happen.” Id. at 609. Defining the risk of harm in this case by the end result — all violent assaults — would allow defendants to be subject to liability for those violent assaults that could be reasonably anticipated (gang-, drug-, and alcohol-related violence) but also allows defendants to be subject to liability for those violent assaults that could not have been reasonably anticipated (such as the shooting incident in this case).
Focusing on the end result, while ignoring the manner in which that end result was reached, is like crediting a broken clock because it happens to be right twice a day. The correctness of a broken clock turns on the coincidence that it *113was viewed at the same time of day that it previously stopped working. But in that case, whether the clock is correct has nothing to do with whether the clock is functioning properly. Likewise, whether the end result was foreseeable has nothing to do with a defendant’s blameworthiness unless the end result occurred in a foreseeable manner. Focusing only on the end result allows liability to turn on coincidence rather than blameworthiness.
Finally, the majority opinion emphasizes the notion that questions of foreseeability should be decided by the jury. 360 Or at 86, 94. Although it is true that foreseeability is a fact issue, simply identifying an issue as factual does not entitle a party to a jury trial or even to proceed to summary judgment. Rather, it depends on the relevant facts alleged in the complaint or established by the summary judgment record. Only when those facts are reasonably disputed does the issue go to the jury. The jury’s role, of course, properly can be an expansive one, particularly in deciding fact-intensive issues like foreseeability. But even as to foreseeability, the court retains its role as gatekeeper, allowing cases to proceed to the jury only if reasonable minds could reach different conclusions on whether the facts alleged or adduced could prove a plaintiffs case. Stewart, 255 Or at 607.8
The majority defines the foreseeable risk in this case broadly as “violent assaults” and leaves it to the jury to decide whether the foreseeable risk should be defined more narrowly. But the majority fails to explain how a reasonable jury could define the risk that broadly, based on the narrow facts alleged. How broadly or narrowly a reasonable jury may define the generalized risks foreseeably created by defendants’ conduct is a question of law appropriately and necessarily resolved by the courts. See Fazzolari, 303 Or at 17 (defining the role of courts).
Negligence liability is intended to reflect community standards of fairness and proportionality. Our system *114operates on the “assumption that judges as well as juries know something about the kind of conduct that is deemed acceptable or not acceptable in the community and that, at least at the higher and lower ends of the continuum of that standard, the court can say that the conduct does or does not meet the standard.” Stewart, 255 Or at 607-08. When liability is clearly inappropriate, as it is in this case, “we are charged with the duty of withdrawing the issue from the jury.” Id. at 609. Although allowing too few cases to proceed to trial intrudes on the role of the jury, allowing too many cases to proceed to trial abdicates the role of the courts. Consequently, I would affirm the trial court’s dismissal and, therefore, dissent.
Landau, J., joins in this dissenting opinion.
See Restatement (Third) of Torts: Phys. & Emot. Harm § 29 (2010) (“An actor is not liable for harm different from the harms whose risks made the actor’s conduct tortious.”); id. § 30 (“An actor is not liable for harm when the tortious aspect of the actor’s conduct was of a type that does not generally increase the risk of that harm.”).
In certain circumstances, prior criminal acts may no longer indicate a continuing risk, even if the prior criminal acts are close in time and place to the plaintiffs injury. In Commonwealth v. Peterson, 286 Va 349, 749 SE2d 307 (2013), for example, the Virginia Supreme Court held that knowledge that two university students were shot and killed on campus did not establish the foreseeability of a mass shooting by the same assailant hours later, which lead to the deaths of 32 people. At the time the mass shooting began, defendants reasonably believed that the earlier shooting was a domestic incident and that “the shooter had fled the area and posed no danger to others.” Id. at 359.
Defendants ask this court to take judicial notice of the fact that Ayala shot a total of nine people, killing two — decedent and one other victim — and that Ayala then fatally shot himself at the scene. It would have been more appropriate for defendants to raise that issue in the first instance with the trial court. Regardless, that issue is beyond the scope of this case because, for the reasons described in the text, this court should affirm the trial court’s dismissal order even if we do not take judicial notice of the facts that defendants offered.
The majority appropriately declines to give weight to all of plaintiffs allegations. 360 Or at 82; id. at 82 n 13.
In Legal Cause in the Law of Torts, which this court relied on in Stewart, 255 Or at 610, Keeton expressly warned against ignoring the manner in which the harm occurred:
“[E]xtreme freakishness of the sequence of events that produces a final state of affairs that might have been expected to come about in some other way commends itself as a relevant consideration and stands against a practice of description of the harm in a way that wholly disregards the mechanism of injury. The choice of description involves a degree of orientation toward type of harm on the one hand or toward mechanism of harm on the other hand, with neither point of view wholly rejected.”
Legal Cause in the Law of Torts at 52.
See, e.g., Solberg, 306 Or at 490 (distinguishing between whether “defendant’s conduct caused a foreseeable risk of harm,” which is an issue of foreseeability, and whether “the conduct was a cause of plaintiffs harm,” which is an issue of factual causation).
The majority attempts to distinguish Hefty and Oregon Steel Mills by asserting that, “in contrast to the circumstances in those cases, as alleged in plaintiff’s complaint, the assailant’s access to Delgado and the other teens outside the Zone was directly connected to the Zone defendants’ asserted negligence!.]” 360 Or at 93. But foreseeability does not turn on whether the connection between the defendant’s conduct and the plaintiff’s injury is direct rather than indirect — both of which may establish liability. Instead, foreseeability turns on whether the connection should have been reasonably expected. Stewart, 255 Or at 609. In any event, it is unclear why the connection is more direct in this case than in Hefty and Oregon Steel Mills. In all three cases, the harm-inducing force is extrinsic to the defendant’s conduct.
In Buckler v. Oregon Corrections Div., 316 Or 499, 853 P2d 798 (1993), this court rejected a common misreading oí Fazzolari as standing for the proposition that “all negligence claims based on general foreseeability of a plaintiffs harm would reach the jury.” Id. at 511 n 8.