concurring in part, dissenting in part.
This case considers the circles of blame for the criminal act of another person. When an injured plaintiff asks a court to hold a defendant financially liable for the criminal act of a third person, Oregon law has a test to decide whether the defendant shares blame. It is a test of general foreseeability, but, when a third person inflicted the injury, it is a test with a sharper focus. It is a test that considers the risk of a particular, violent person or of a location made unsafe by the violence of others. Did the defendant, who is to be faulted for the act of another person, actually know or have reason to know of the specific danger to the plaintiff posed by this third person or by an unknown person at this location?
I write separately, concurring with most of the majority’s analysis. I do not disagree with the majority’s disposition of the claim against the national fraternity. Nor do I disagree with the majority’s remand for further proceedings on the claim against the Beta Chapter for negligence per se based on the prospect of violation of former Oregon Administrative Rules. The defendant’s argument on appeal that plaintiff “could not show that her injury resulted from any [OAR] violation” was not made to the trial court in terms of an inability to show that she was in the class of persons meant to be protected by a rule or that a sexual assault was not the sort of injury that alcohol regulations were intended to prevent. Compare McAlpine v. Multnomah County, 131 Or App 136, 144, 883 P2d 869 (1994), rev den, 320 Or 507 (1995) (cited by the Beta Chapter to the trial court and this court; listing four requisites for negligence per se), with Gattman *429v. Favro, 306 Or 11, 24, 757 P2d 402 (1988) (assault victim was not within the class of persons intended to be protected by liquor liability statute). Although the trial court did not explain the reason for its dismissal of the negligence per se claim, this court does not address new reasons on appeal, which were not developed and which, if squarely addressed, could have led to a different record. See Outdoor Media Dimensions Inc. v. State of Oregon, 331 Or 634, 659-60, 20 P3d 180 (2001) (examining the “right for the wrong reason” review of a summary judgment). Unaddressed reasons remain undecided.
As to the claim of ordinary negligence, I dissent to suggest that our case law requires more than plaintiff offers to hold the Beta Chapter liable. To have “facilitated” a crime, by having carelessly provided a setting or having made the crime more likely, does not mean, without more, that the defendant had reason to foresee the crime and thereby become liable. Beyond that setting, the national statistic, to which plaintiffs claim reduces, does not mean what it might appear to mean, and it is insufficient evidence as a matter of law. To add that misunderstood and inadequate statistic to this setting does not mean that the local fraternity had reason to foresee the crime and become liable for it.
I. LIABILITY FOR ANOTHER’S INTENTIONAL ACT
A. Prototype Precedents
Two principal cases not only typify the lines of cases on an unsafe person or location, they are the modern source of Oregon’s law of negligence. In Fazzolari v. Portland School Dist. No. 1J, 303 Or 1, 734 P2d 1326 (1987), a high school student, while waiting before school, was beaten and raped by an unknown assailant. Only 15 days beforehand, another woman had been sexually assaulted on the same school grounds. Id. at 21. The plaintiff attempted to introduce evidence of other incidents. The plaintiff alleged, among other things, that the school district had failed to warn and to provide security personnel when the district knew of previous similar attacks. The Supreme Court explained that the special duty of a school toward its students “does mean that negligence toward a student is tested by an obligation *430of reasonable precautions against foreseeable risks beyond those that might apply to other persons.”1 Id. at 20 (emphases added). Given the prior attack or attacks at that location, the court held that the risk of sexual assault was foreseeable and that dismissal on a motion for a directed verdict was error. Id. at 22. Although the case involved a special and perhaps higher duty, the fact pattern typified that in which later negligence cases could find potential liability for a location made unsafe by the risk of third-party violence.
In Buchler v. Oregon Corrections Div., 316 Or 499, 853 P2d 798 (1993), a prisoner escaped from custody, found keys in a state van, stole a gun from his mother’s home, and shot two people two days later 50 miles away. The plaintiffs alleged, among other things, the state’s negligence in failing to give warning of the escape to the public and especially to the plaintiffs’ neighborhood. The Supreme Court evaluated the allegation of fault involving the van keys and the failure to warn in terms of ordinary negligence or general foreseeability, not a special duty. Id. at 504-05.
The court began by rejecting the possibility that liability would follow from simply “facilitating” the harmful act, a distinction that is critical to the case at hand. Id. at 510. Previously, in Kimbler v. Stillwell, 303 Or 23, 734 P2d 1344 (1987), the court had reinstated the plaintiffs claim of negligence for a store’s failure to lock a gun display case, which led to a thiefs use of the gun to inflict injury later. Buckler overruled Kimbler, explaining that a defendant should not face liability simply because a defendant is involved in the chain of events — even those without which the death or injury would not have occurred. The court elaborated:
“While it is generally foreseeable that criminals may commit crimes and that prisoners may escape and engage in criminal activity while at large, that level of foreseeability does not make the criminal’s acts the legal responsibility of everyone who may have contributed in some way to the criminal opportunity. In other words, in our society it is *431foreseeable that crimes may occur and that the criminals perpetrating them may cause harm. Thus, in a general sense, it is foreseeable that anyone whose conduct may in any way facilitate the criminal in committing the crime has played some part in the resulting harm. But mere ‘facilitation’ of an unintended adverse result, where intervening intentional criminality of another person is the harm-producing force, does not cause the harm so as to support liability for it.”
316 Or at 511-12 (emphasis added). The Buckler court considered the state’s carelessness in leaving the keys in the van within the chain of causation, but, ultimately, deemed it not to be actionable, like the store’s failure to lock the gun case in Kimbler. The court concluded that the “generalized foreseeability principle” did not reach so far.
To find the defendant liable for the intentional act of a third person, the court required more than “facilitation” of the crime. The court required “evidence that defendant here either knew [or] had reason to know of the specific danger presented by the prisoner to plaintiffs.” Id. at 516 (emphasis added). Liability did not turn on hindsight, nor the perspective of an all-knowing “reasonable person.” Liability turned on what this particular defendant knew or had reason to know. The prisoner’s history gave the state no indication of violent crimes. The plaintiffs had not shown why the state should have known the prisoner would be near the plaintiffs, should have known where the prisoner’s mother lived, and should have known he would have stolen a gun. The court concluded that the state had “no duty to warn in the absence of [the] knowledge” of the specific danger presented to the plaintiffs by the prisoner. Id. The court affirmed a dismissal on summary judgment. Although Buckler dismissed the claims, its fact pattern typified which negligence cases could, with something more, find a defendant liable for the risk of a third person’s violence.
B. Liability for a Dangerous Person
Two lines of cases, involving a defendant’s liability either for a dangerous person or an unsafe location, resemble these two prototype cases, and both lines of cases contribute to the answer to the case at hand. I draw first from the cases *432on a dangerous person, then draw on cases on an unsafe location, before turning to the facts at hand, and, as to the local fraternity, disagreeing with the majority’s conclusion.
The fact patterns with dangerous third parties often involve alcohol, robbery, assault, and sexual violence. The decisions caution courts against making easy assumptions. In several cases involving alcohol, it did not suffice to allege that serving alcohol to an intoxicated person gave reason for a defendant to foresee that a third person would intentionally assault someone. In Moore v. Willis, 307 Or 254, 260, 767 P2d 62 (1988), the Supreme Court declared:
“The fact that someone is visibly intoxicated or underage, standing alone, does not make it foreseeable that serving alcohol to the person creates an unreasonable risk that the person will become violent.”
The court affirmed a judgment on the pleadings in favor of two tavern owners after a patron drew a gun and killed a taxi driver.
In Sparks v. Warren, 122 Or App 136, 856 P2d 337 (1993), it did not suffice to allege that a fraternity failed to regulate the use of alcohol by its members and knew or should have known that underage drinkers frequently become abusive and violent. The plaintiff was assaulted by a fraternity member after drinking there and at a tavern. We responded:
“The risk flowing from the negligence alleged here is not that a minor will drink but that someone predictably will be exposed to danger of an assault if defendants were negligent as alleged.”
Id. at 140. Because the plaintiff had failed to present evidence that the violence was foreseeable, we affirmed summary judgment for the defendants.
Recently, in Chapman v. Mayfield, 263 Or App 528, 329 P3d 12, rev allowed, 356 Or 400 (2014), a fraternal lodge was alleged to have served a visibly intoxicated patron. The patron shot and injured two men, who sued the lodge and others. We recalled that,
“to establish foreseeability, a plaintiff must first plead and then prove specific facts — beyond the fact of visible *433intoxication — from which an objectively reasonable fact-finder could find or reasonably infer that the tavern owner who served the visibly intoxicated person knew or had reason to know that serving that person created the unreasonable risk that that person would become violent”
Id. at 531 (emphases added). Like Sako in the case at hand, the lodge did not know any “specific facts” about the patron, which would have made violence foreseeable. Like the Beta Chapter here, the lodge had suffered no prior incidents in which patrons, to whom the lodge had served alcohol, had become violent. And, like plaintiff here, the plaintiff in Chapman offered only generalized information. A medical doctor testified that “[i]ntoxicated drinkers frequently become violent,” and a bartender from a different bar testified that, when a patron becomes violent, “[t]hat’s the alcohol talking.” Id. at 532, 546.
We held “that evidence is insufficient to permit a rational factfinder to make the finding,” which the law requires, that the defendant, just by being in the alcohol business, had reason to foresee that serving an intoxicated patron would pose an unreasonable risk of violence. Id. at 533. The generalizations, whether from a doctor or a bartender, did not suffice to prevent summary judgment for the defendant, because “too many intermediate inferences and assumptions” or “guesswork” would be required to find that the defendant had knowledge or reason to know the risk from this patron.2 Id. at 535. See also Allstate Ins. Co. v. Tenant Screening Services, Inc., 140 Or App 41, 914 P2d 16 (1996) (nothing in the mixed criminal record of a tenant/ security guard “foreshadowed his particular crimes,” i.e., violent crimes of sexual assault and sodomy, so as to state a claim for liability against a tenant-screening service that had reported a clean record).
In contrast to the alcohol cases from Moore to Chapman, the outcome favors the plaintiff when the plaintiff offers some evidence that the defendant knew or had *434reason to know the “specific danger” that a third person presented to plaintiff. See Buchler, 316 Or at 515 (no knowledge of specific danger to plaintiffs). For example, in Brown v. Washington County, 163 Or App 362, 987 P2d 1254 (1999), rev den, 331 Or 191 (2000), an inmate escaped from a corrections center and killed his brother. The defendant county knew that the inmate had a violent history, was agitated just before his escape, used his wife’s address as his address, suspected her of an adulterous affair, and likely would go to her address. Yet, the county failed to warn anyone at the address of his escape. Such knowledge of the specific threat posed by a third person permitted the possibility of liability. See also Washa v. DOC, 159 Or App 207, 979 P2d 273 (1999), aff'd by an equally divided court, 335 Or 403, 69 P3d 1232 (2003) (the defendant, who inadequately supervised a parolee, was well aware of the parolee’s violent assaults on women and the risk that he would remain “a potential mutilator and killer of women,” such that the defendant had reason to have foreseen the risk of two rapes and a murder).
In McAlpine v. Multnomah County, 166 Or App 472, 999 P2d 522 (2000), rev den, 336 Or 60 (2008), a parole officer knew or had reason to know that the parolee had a history of violent offenses, including possession of a gas bomb, armed robbery, and assault and knew that he had committed new offenses, but failed to issue an updated parole report, allowing the parolee to be released. The plaintiff was subsequently injured in a traffic-related altercation with the parolee. The allegation of such knowledge permitted the county’s potential liability for the parolee’s conduct. See also Panpat v. Owens-Brockway Glass Container, 188 Or App 384, 394-95, 71 P3d 553 (2003) (the employer knew it had an employee with an explosive disorder, that he was not authorized to return to work without a mental health evaluation, and that the employee had twice engaged in obscene verbal confrontations with the victim).3
*435C. Liability for an Unsafe Location
These cases involving a dangerous third person parallel the cases involving a location made unsafe by unknown persons. There is a similar requisite in common: In location cases, liability can arise when the defendant knew or had reason to know from specific evidence that the location presented a risk of criminal harm or violence from third persons, even if the identity of the bad actor could not have been foreseen. Two early cases, with contrasting outcomes, illustrate that local experience is the specific evidence needed. Two recent cases, with contrasting outcomes, will sharpen our focus.
In Uihlein v. Albertson’s, Inc., 282 Or 631, 580 P2d 1014 (1978), a supermarket shopper was assaulted and her pocketbook stolen by an unknown assailant. She alleged a lack of security, argued the market was in a high-crime area in Portland, and relied on the comment of the store’s former manager that it was the “roughest store” at which he had ever worked. The Supreme Court, however, found no evidence in the record about the other stores where the former manager had worked, which would give his statement context and probative value. Id. at 636. The only evidence of a “high crime rate” was shoplifting at the store. In the prior couple years, the store had suffered no prior incident of robbery or assault on its premises. The court concluded that the defendant did not know or have reason to know of the likelihood of the harmful acts of the kind the plaintiff suffered. Id. at 640-41. Summary judgment for the defendant was affirmed.
More specific evidence concerning the location produced a different outcome in Brown v. J. C. Penney Co., 297 Or 695, 688 P2d 811 (1984) (J. C. Penney). A customer was attacked and robbed of her purse in the parking lot of the defendants’ shopping center in Eugene. She alleged inadequate security and offered evidence of 268 incidents of criminal activity in the immediate vicinity in the prior six-month period. The shopping center knew of the incidents and acted on the reports in allocating security forces. Because the defendant had actual knowledge or reason to know of the risk — a risk specific to the location — the court affirmed the judgment for the plaintiff. Id. at 710.
*436Subsequent cases illustrate the same point. See McPherson v. Oregon Dept. of Corrections, 210 Or App 602, 617-18, 152 P3d 918 (2007) (defendant had reason to know of risk of assault of son and sexual assault of mother in apartment laundry building when defendant knew neighborhood was unsafe, neighborhood had 86 emergency calls to police in nine years, and apartment managers had called police about problems, including vandalism and trespass in the laundry building); Sande v. City of Portland, 185 Or App 262, 271-72, 59 P3d 595 (2002) (given repeated assaults and robberies of lone victims by a “mountain-bike” assailant, the city had reason to foresee harm to plaintiff from its instruction to neighbor not to warn plaintiff).4
A more recent pair of contrasting cases provide a sharper focus for this case. In Stewart v. Kids Incorporated of Dallas, OR, 245 Or App 267, 261 P3d 1272 (2011), rev dismissed as improvidently allowed, 353 Or 104 (2012), a 13-year-old girl, volunteering at a car wash at a Dairy Queen restaurant, was forced into the men’s restroom and sexually assaulted. Her guardian ad litem alleged that the carwash sponsor and restaurant were negligent and had reason to foresee that sexual predators would be attracted to a carwash with teenage girls participating. We found the plaintiffs theory no different in essence than the “discredited theory in Kimbler” that, because criminals are likely to steal guns and harm people, the failure to secure a gun display created a foreseeable risk. Id. at 284. We rejected the plaintiffs argument that the “‘place or character’ of the Dairy Queen” (where youth could be found) permitted a reasonable inference of risk. Id. at 285. Finding nothing more than a “theoretical possibility” of sexual assault at that location, we affirmed dismissal for failure to state a claim. Id. at 286.
*437In Piazza v. Kellim, 271 Or App 490, 354 P3d 698 (2015), a 17-year-old student was fatally shot while waiting in line outside an underage nightclub in the “Old Town/Chinatown neighborhood” in downtown Portland. The shooter was mentally ill and had gone to the “nightclub looking to shoot ‘preppies’ or ‘pop tweens.’” Id. at 492. Seven years before, a shooter had fired into a crowd at the same location, and there had been a history of fights and assaults in line outside the nightclub. The surrounding neighborhood, the plaintiff alleged, was “plagued” by recurrent violence and gang activity. Id. at 494. Managers of local bars, including the defendant, attended a police “summit” to reduce violence. Id. The plaintiff alleged that the nightclub defendants had failed to take reasonable measures to protect customers from the criminal acts of third parties.5 The trial court dismissed the case, under ORCP 21 A(8), for failure to state a claim. On appeal, we stressed the requisite for a defendant’s liability for an unsafe location, explaining that
“a plaintiff must allege facts demonstrating that the harm by third-party criminal conduct was foreseeable to the defendant in a concrete way and may not rely on the abstract proposition that ‘crimes may occur and that the criminals perpetrating them may cause harm.’”
Piazza, 271 Or App at 504 (quoting Buchler, 316 Or at 511-12) (emphasis added). We distinguished Stewart because that plaintiff had failed to “allege specific factual support— as opposed to relying on generalized abstractions about the existence of criminal activity.” Id. at 506 (emphasis in original). In Piazza, the plaintiff had alleged that the defendants knew the alleged criminal activity in the vicinity and at the specific location. We concluded that, for the purpose of stating a claim, a jury could find “the prior physical assaults at or around” the nightclub to be “similar enough” so as give the nightclub defendants reason to have foreseen the shooting. Id. at 511.6
*438II. LIABILITY OF THE BETA CHAPTER
If the principles of these cases were applied to the facts here, the majority should have concluded that plaintiff has failed to offer evidence that the local fraternity knew or had reason to know of an unreasonable risk to plaintiff from a dangerous person or an unsafe location. Each alternative requires separate discussion.
A. No Knowledge or Reason to Know of a Dangerous Person
The majority opinion recites, and plaintiff concedes, that there was no evidence that the Beta Chapter had reason to know that Sako had a propensity for violence. Without contradiction, Gerritz, who was the chapter president at the time of the party, declared, “Mr. Sako had no known prior sexual assault record or any history of becoming violent with or without the influence of alcohol.” Absent any reason to know that Sako might assault plaintiff or another person, the chapter could not become liable, under the first line of cases, for the intentional or criminal act of a person who the chapter knew or should have known was dangerous. See Moore, 307 Or at 261 (claim dismissed); Sparks, 122 Or App at 140 (claim dismissed). This is doubly true because, unlike several of the cases reviewed above, the Beta Chapter did not serve him alcohol when he was visibly intoxicated; it did not serve him any alcohol at all. See Moore, 307 Or at 261 (where liability for another’s violence does not follow even from serving someone who is visibly intoxicated). Under the first line of precedents, the local fraternity cannot be held liable for a specific “dangerous person,” even for allowing an underage member to possess his own alcohol or, by logical extension, to bring a guest to his room, when Sako had no history of assault or threatening misbehavior. See Moore, 307 Or at 260-61 (intoxicated or underaged drinker); Sparks, 122 Or App at 140 (fraternity drinker). The Beta Chapter lacked-the requisite reason to know any “specific danger” that this particular person presented to plaintiff. Buchler, 316 Or at 516 (requiring defendant to know specific facts of threat of violence); Chapman, 263 Or App at 532 (need for specific facts showing risk of violence from this intoxicated person); see also Brown, 163 Or at 362 (liability *439where threat of person known); McAlpine, 166 Or App at 472 (same).
B. No Knowledge or Reason to Know of an Unsafe Location
The majority is quite right to the extent it observes that “a risk of third-party criminal conduct” is not limited to “a defendant’s knowledge of a specific perpetrator’s propensity for violence.” 273 Or App at 404 (referring, however, to the propensity for violence of “Sako or other chapter members” (emphasis added)). But the precedent that the majority cites for broader liability is an early case, predating Buckler and involving a hazardous location, a bank’s night depository. Torres v. United States Nat. Bank, 65 Or App 207, 670 P2d 230, rev den, 296 Or 237 (1983). It was a claim initially dismissed at the pleading stage but reversed on appeal to allow the plaintiff to offer evidence of other robberies of night depositories on the street or in the area. Id. at 214.
Even at that, Torres drew a dissent, contending what later cases taught. The dissent declared that “ [p]lain-tiff should have pleaded with specificity that defendant knew, or should have known, that in the past [the] patrons at the 72nd and Fremont Branch had been robbed or assaulted while using the night depository.” Id. at 216 (Van Hoomissen, J., dissenting); see, e.g., Stewart, 245 Or App at 284 (criminal threat to teen carwash). Regardless of the difference in majority or dissenting views, both views in Torres contemplated that, ultimately, a plaintiff must show specific evidence from the experience of the locale to demonstrate that a particular location was made unsafe by unknown third parties. Such localized specificity is what the complaint in Piazza illustrated. As it was, Torres was only an early-stage, pleading case. After Buchler and the parallel line of location cases, Torres cannot be read to permit a defendant’s liability for providing a setting at which the general risk of crime might be realized, when no local history foretells the risk.
Unlike Torres, the case at hand is not a pleading case. Defendant’s evidence challenged plaintiff to offer evidence to create a dispute of fact to show that the local fraternity was a location known or knowable by the local fraternity *440to be hazardous from past experience. See J. C. Penney, 297 Or 695; Uihlein, 282 Or at 631; Stewart, 245 Or App at 267; McPherson, 210 Or App at 605-06 (risk of criminal activity at location). To borrow from Piazza, a more recent unsafe-location case, this plaintiff must have responded with facts demonstrating that the risk of a hazardous location “was foreseeable to the defendant in a concrete way” and not by reference to the general risk of such crime. Piazza, 271 Or App at 504 (citing Buchler, 316 Or at 511-12).7
Prior to this incident, there is no evidence whatsoever of sexual assault at the Beta Chapter. Gerritz recounted, “To my knowledge, the sexual assault committed by Mr. Sako during the event was the first and only known sexual assault to have occurred at the Phi Psi House and/or to have involved Mr. Sako or any other member of the Local Chapter.” Collinsworth, the executive director of the national fraternity organization, confirmed the same fact. Plaintiff offered no evidence of any violence of any kind at the local fraternity and, for that matter, no evidence of any sexual assault at fraternities, sororities, or dormitories at Oregon State University. Absent any evidence to dispute the Gerritz or Collinsworth reports, plaintiff failed to offer “specific facts” to show that the local fraternity knew or had reason to know from its experience that its location threatened plaintiff with an unreasonable risk of criminal assault. See Buchler, 316 Or at 515-16 (dangerous person risk); Moore, 307 Or at 261 (alcohol risk of violence); Uihlein, 282 Or at 640 (dangerous location risk); Sparks, 122 Or App at 140 (alcohol risk of violence). Under the line of cases involving a defendant’s liability for an unsafe location, the Beta Chapter should not be found liable for contributing to an unsafe location, when no “specific facts,” such as incidents at or around the fraternity, gave the Beta Chapter reason to foresee a sexual assault by an unexpected or unknown assailant.
*441III. DIFFERING VIEWS
A. Evidence Distinguished
The majority reaches a different conclusion, but our difference of opinion is not due to defense arguments that the local fraternity employed so many measures to manage the party. I concur with the majority’s rejection of those arguments on this issue of law. It is certainly true that the Beta Chapter followed the advice of the national fraternity by not furnishing alcohol at the party. Participants were required to bring their own beer, check it at the bar, and retrieve one beer at a time after showing proof of age. Four fraternity members served as “sober monitors,” and two security guards helped supervise the party. When circumstances outside the house became questionable, the fraternity shut down the party early.
Although those precautions would reduce the risk of misbehavior, they are simply evidence, albeit significant evidence, for a jury’s ultimate evaluation of the facts when deciding whether, all in all, the Beta Chapter breached its duty of care. Those precautionary measures may reduce some of the reasons that the Beta Chapter could foresee sexual assault in common areas, but those measures do not resolve foreseeability as a matter of law when the local fraternity allowed members to possess alcohol and invite guests into their rooms. I concur with the majority in distinguishing such evidence and leaving its significance for a jury.8
B. Evidence Misconstrued
Where I respectfully disagree with the majority is in evaluating two matters: (1) the role of the setting and *442(2) the meaning and sufficiency of a national statistic. I do not disagree with the majority in a desire to see an innocent victim compensated for injury, to see anyone legally responsible held to task, and, by assuring a tort remedy, to encourage safeguards to avoid repetition of such an offense in the future. Yet, as to this local fraternity on this occasion, Oregon law should compel a different conclusion for these two reasons: (1) the party setting is “facilitation,” not a basis for liability, and (2) the national statistic, upon which critical arguments are based, is misunderstood and is insufficient evidence.
1. Facilitation Is Not Foreseeabilty
The majority is not entirely misguided in recognizing the role of the setting. Kerr, the coordinator of “Greek Life” at Oregon State University, opined that it was his advice to fraternity presidents to close access to private rooms during social events with alcohol. In a private room, the sponsor’s “control disappears.” He said, “[T]he opportunity for there to be sexual misconduct rises because they’re no longer in an area that is managed.” Kerr’s evidence was simply that it was “good practice to close [the] private rooms” to preclude the “opportunity” of sexual misconduct.
When asked if he had relayed his advice to the Beta chapter, Kerr said he did not recall. Over the last 10 years he “probably” had the conversation with a representative of each fraternity. If he talked to the Beta Chapter, it may not have been during the tenure of its current, student officers. Gerritz, the Beta president, testified he had not heard of a fraternity at OSU closing its private rooms during a party, although he had heard of sororities doing that. When Kerr was asked again if he had given his advice to the Beta Chapter, he concluded, “[T]hat I can’t confirm.” Ultimately, it does not matter whether Kerr’s knowledge should be imputed to the Beta chapter, either as a matter of reasonable inference or as a matter of impermissible speculation. See Chapman, 263 Or App at 536 (describing the “intermediate inferences” as “guesswork”). Kerr’s testimony was about “good practices” or preventative measures. His testimony was about the setting, not about the reasonable foreseeability of alcohol-related, third-party violence. Kerr *443did not testify to relative likelihood of violence, the propensity for alcohol to lead to sexual assault, nor, so far as the record reflects, did he testify that a sexual assault had ever happened at an OSU party. See id. (rejecting inferences about defendant’s knowledge to be made from the testimony of a doctor and a nearby bartender). For the critical fact about the foreseeability of assault, plaintiffs claim relies on the national statistic from the fraternity’s guide — to be discussed next.
Construing the evidence most favorably to plaintiff, the fraternity may have “facilitated” the sexual assault by tacitly permitting a fraternity member to keep alcohol in his room and to invite a party guest there. However, facilitating a crime, even carelessly, does not suffice to impose financial liability for a third person’s intentional act. Buchler, 316 Or at 510-11. We know from the series of cases like Moore, Sparks, and Chapman that actually serving a visibly intoxicated person does not give the alcohol provider reason to know “that the person would become violent.” See, e.g., Chapman, 263 Or App at 533. “Specific facts” about the person, which the server had reason to know, must be offered in order to create the possibility of liability.
We know from cases like J. C. Penney and Stewart that to provide a setting like a parking lot or a restroom, where the crime is easier to accomplish, is not enough to make a defendant liable for an unsafe location. And we know from Buckler, where leaving keys in the van hastened the escape, that carelessly facilitating a crime does not make that crime foreseeable by the defendant. Even when a defendant’s act is within the chain of causation, that does not make the defendant responsible for another’s intentional act. Buckler’s observation bears repetition:
“While it is generally foreseeable that criminals may commit crimes and that prisoners may escape and engage in criminal activity while at large, that level of foreseeability does not make the criminal’s acts the legal responsibility of everyone who may have contributed in some way to the criminal opportunity.”
Buchler, 316 Or at 511 (emphasis added). Under Oregon law, the fraternity, at worst, did no more than “facilitate” this *444crime by allowing Sako to have alcohol and a guest in his room. Without more, that is not enough to make the Beta Chapter financially responsible for sexual assault.
2. A Statistic, Misunderstood and Insufficient
The majority concludes that the Beta Chapter “knew that alcohol-related sexual assaults were a foreseeable risk of hosting social events where alcohol is available, and that permissive alcohol use by underage members and access to private rooms during such social events increases the risk of sexual misconduct.” 273 Or App at 407. In part, the conclusion rests simply on the spectre of a Halloween party at a fraternity — a setting that, at best, amounts to “facilitation,” not reasonable foreseeability. In part, the majority’s conclusion is based on “computer-based programs,” which include a topic on alcohol at social events and a topic on sexual assault, and on the Risk Management and Insurance Guide (guide), provided by the national fraternity. The content of the online training, however, is not in this record, and to assume its content is a speculative inference. If we could reasonably assume that the online training contains information akin to the guide, then that online training would tell a fraternity no more than what the guide does, and that is not enough with which to make a reasonable inference. The guide is in the record and warrants scrutiny. It is the document that provides the national statistic on alcohol-related sexual assault that is the heart of plaintiff’s case and the majority’s conclusion.
The national statistic on alcohol-related sexual assault, offered'in this case, is not the same as plaintiffs argument about it. When properly understood, this incomplete and generalized statistic is not a substitute for the requisite, specific evidence of experience at this location. In its summary of facts, plaintiffs brief takes the liberty to argue that the national fraternity’s guide purportedly shows “both defendants’ knowledge of the epidemic of sexual assault and alcohol abuse at fraternities” in this nation. (Emphasis added.) The majority opinion succumbs to this tempting spin on a misconstrued statistic. In its analysis, the majority finds that plaintiff offered evidence “that the chapter knew that alcohol-related sexual assaults were a *445risk in certain circumstances at fraternities on college campuses nationwide.” 273 Or App at 411; see also id. at 408 (“potential problem” and “increased the risk of sexual misconduct”). In truth, the evidence in this record is much less than plaintiffs interpretive spin. It is critically incomplete.
The guide underscores the importance of its recommended precautions, like the measures used at this party, with an introduction on “social programming and alcohol.” It recounts:
“In the last decade, over 700 Greek-related accidents and injuries were reported. In many, insurance claims and/ or legal action resulted. Of the most serious incidents, alcohol was a factor in 96% of falls from roofs, 97% of sexual abuse, 96% of fights and 87% of automobile accidents.”
In a later section on sexual assault, the guide advises:
“As with so many problems on college campuses, alcohol plays a prevalent role in sexual assaults. Many campus rapes involve alcohol or, in some cases, drugs. One report estimates that 50% of the women who fall victim to rape attempts and 75% of their attackers have been drinking prior to the incidents. Of sexual abuse cases which have been brought against fraternities, 97% involved alcohol.”
These statistics speak of alcohol within a subset of injuries. These statistics do not speak of the larger set of individual and social activities with alcohol. Context is missing. These statistics tell how often alcohol is involved when some circumstances eventuated in falls, fights, automobile accidents, and sexual assaults, but these statistics do not foretell how often alcohol consumption, by how many tens of thousands of fraternity members and guests at how many thousands of organized social events and unplanned individual interactions over 10 years, resulted in 700 accidents and injuries. To say that alcohol was consumed in 97 percent of incidents of sexual abuse does not mean that 97 percent of fraternity members who consume any alcohol will commit rape. The guide does not report how many sexual assaults occurred, let alone how often they occurred in comparison to how often fraternity members consumed alcohol alone or with an acquaintance, friend, guest, or date. Without knowing the proportion of sexual abuse in the larger field of all *446activities, ranging from individual encounters to planned social events, the guide’s incomplete statistic cannot rationally be used to foretell or to foresee that alcohol consumption portends rape.
The guide introduces the topic of sexual assaults, saying “alcohol plays a prevalent role in sexual assaults.” It follows with the 97 percent statistic. In context, the term “prevalent role” refers to the statistic on the frequency of alcohol within the subset of injuries. The term does not mean that any alcohol consumption leads foreseeably to rape. Again, without knowing how often alcohol is consumed, it begs the question to say that alcohol use is “prevalent” in a high number of assaults.
In its section on “social programming and alcohol,” the guide does address the effects of alcohol, indicating:
“Alcohol is a depressant. It slows down your bodily functions and the ability to respond. It does not increase the sex drive, but will decrease your inhibitions.”
Although decreased inhibition suggests misjudgments, the statement still does not foretell the frequency of sexual assault in the larger context of all alcohol consumption. It is a statement, much too generalized, and insufficient to make a fraternity reader, who left alcohol in the possession of another fraternity member, legally liable for another member’s rape of a victim. See Chapman, 263 Or App at 532-33 (doctor’s testimony that “[i]ntoxicated drinkers frequently become violent” did not suffice to show server knew the same and had specific reason to foresee violence).
To be sure, no evidence in this record says that there is an “epidemic” of sexual abuse. The raw numbers and the proportions are not here. Nor is there any evidence in this record that alcohol-related sexual assaults were prevalent in fraternities on college campuses nationwide. To say so is to misconstrue the guide’s statistic. To say so is to speculate. To say so is to make an inference that is not rational or reasonable. See Chapman, 263 Or App at 536 (too many “intermediate inferences represent guesswork”).
*447The misuse of the guide’s statistic is easier to recognize if the same rationale is applied to the companion statistic that 87 percent of automobile accidents were alcohol related. Here again, we do not know how many thousands of fraternity gatherings or individual encounters included the consumption of alcohol. We should know that we are not permitted to speculate how many instances of alcohol consumption eventuate in how many vehicle accidents. We should know that the 87 percent figure cannot mean that a local fraternity should foresee and become liable, without more, for any alcohol-related collision involving a fraternity member. But, plaintiffs statistical rationale would mean that, if a fraternity member was allowed to possess alcohol in his room and, if a guest had suffered injuries on a ride home with a fraternity member in an alcohol-related accident, then the fraternity would be held liable. Plaintiffs statistical rationale, which the majority necessarily assumes, would hold the fraternity liable although the fraternity had not provided him alcohol while visibly intoxicated, nor even provided him any alcohol at all. It would suffice that he was allowed to possess his own alcohol.
Liability for a third person’s intentional violence should not be easier to prove than liability for a third person’s drunk driving. Oregon cases have established that the risk of violence is not reasonably foreseeable simply from having served someone who is visibly intoxicated. See, e.g., Moore, 307 Or at 260-61; Chapman, 263 Or App at 531; Sparks, 122 Or App at 139-40. Logically, therefore, to hold a defendant liable for a third person’s deliberate violence, purportedly because defendant allowed -the third person to possess alcohol, should not be even easier than to hold a defendant liable for another’s violence when defendant served a visibly intoxicated, third person. Because the fraternity could not be held liable for deliberate violence if it had served Sako while visibly intoxicated, it is difficult to understand how the fraternity could be liable for his assault when having allowed him to consume his own alcohol.
Even if plaintiff’s statistic were what it is not — a national statistic on the frequency of sexual assault among all occasions of alcohol consumption — it would only be the *448same sort of evidence that Buchler rejected as a reason to foresee injury by a third party. In Buchler, it might have been plausible to fear that an escaped prisoner would use violence in order to remain free. Yet, the Supreme Court declared that the generalized idea that crime may occur when a criminal is at large does not suffice to give a defendant, even someone who may have contributed to the escape, reason to foresee violence from subsequent crime by a previously nonviolent offender. 316 Or at 511.
The Beta Chapter correctly contended that plaintiffs reliance on a national statistic was too generalized to give it reason to have foreseen a crime which had no precedent in Sako’s past or in the group’s experience. We should remember from the cases on unsafe locations that “specific facts” showing a risk of crime at or near this location was necessary to show that a defendant had reason to foresee harm to plaintiff from an unexpected, third person. Uihlein, 282 Or 631 (summary judgment for lack of local evidence); Piazza, 271 Or App 490 (specific facts of location’s experience permitted claim for failure to protect patron); Stewart, 245 Or App 267 (dismissal for failure to state a claim based on general criminal risk to teen carwash). The misconstrued and incomplete statistic, which is the heart of plaintiffs claim to foreseeability, is not “the specific factual support,” based on Sako’s behavior or the location’s experience, that Oregon law requires to have given the Beta Chapter reason to foresee and become liable for the intentional offense of a third person.
III. CONCLUSION
I concur that the claim based on negligence per se cannot be dismissed for the reasons defendant had urged, but I cannot concur that the law permits a claim of ordinary negligence against the local fraternity. The offense is reprehensible, but on this record and on the negligence claim, the circle of blame for another’s crime should not go further than the offender. Therefore, I reluctantly dissent.
Speaking of cases without a special duty, the court commented that “the issue of liability for harm actually resulting from defendant’s conduct properly depends on whether that conduct unreasonably created a foreseeable risk to a protected interest of the kind of harm that befell the plaintiff.” Fazzolari, 303 Or at 17.
In Chapman, we recognized that ordinarily foreseeability is a jury question but noted that that principle “is not a rule of law, it is a rule of thumb.” 263 Or App at 537. It is guidance that “affords the least guidance where, as here, the Supreme Court has articulated with some precision what facts a plaintiff must prove in order to establish foreseeability.” Id. at 537.
This line of cases involving specific knowledge of a dangerous person has extended to the negligence of an off-duty employee. The employer’s knowledge of an overworked employee’s schedule permitted the possibility of liability, when the weary, off-duty employee was involved in an auto accident. Faverty v. McDonald’s Restaurants of Oregon, Inc., 133 Or App 514, 525-26, 892 P2d 703 (1995), rev dismissed, 326 Or 530 (1998) (the employer “knew or had reason to know of the number of hours [the employee] had been working”).
In a variation on the unsafe location theme, the defendant’s summary judgment was reversed in Cunningham v. Happy Palace, Inc., 157 Or App 334, 970 P2d 669 (1998), rev den, 328 Or 365 (1999). Late at night, the defendant’s bouncer prevented the intoxicated plaintiff from calling her daughter for a ride home and instead left her outside, impaired, and vulnerable. Forced to hitchhike, she was picked up and raped by three men. We concluded that a jury could find that the defendant had reason to foresee that forcing the plaintiff to leave the safety of the restaurant before she could make her call and abandoning her outside, at night, and impaired, placed her at a risk of criminal assault.
Certain Rotary defendants, sponsors of the international exchange program in which the student participated, were also named defendants.
Dissenting, a member of this court was unpersuaded that the incidents at or around the nightclub gave the nightclub defendants reason to foresee the incident. Id. 271 Or App at 520 (Edmonds, S. J., dissenting).
The majority borrows instead from Chapman, a dangerous person case, to set up the majority’s proposition that this rape should have been foreseeable to the local fraternity by reason of its training about the general risk of alcohol-related misconduct. The majority holds that, notwithstanding Buchler or Moore, defendant’s imagined knowledge of the general risk of alcohol-related assault satisfies foreseeability to render a defendant liable for facilitating this crime. 273 Or App at 405-06.
Because it reverses, the majority opinion does not address the declaration of plaintiff’s counsel that plaintiff retained an expert to support the claims. Plaintiff indicated that the expert would testify that the fraternity breached a duty of care. Because the critical issue is what the Beta Chapter knew or had reason to know, and, because witnesses were deposed and provided concrete information on'what the local fraternity knew or had reason to know, an outside expert’s opinion would seem to amount to no more than the doctor or the bartender down the street in Chapman, who could offer little that was helpful to the requisite foreseeability involving what the local defendant knew or, by reason of local circumstances, should know. See Belgarde v. Linn, 205 Or App 433, 441-42, 134 P3d 1082, rev den, 341 Or 197 (2006) (affidavit under ORCP 47 E ineffective where inappropriate).