¶ 11. dissenting. The tragic fatality underlying this regulatory enforcement proceeding provides tempting grounds to overlook the serious evidentiary and legal flaws in the State’s case. That is not a luxury afforded to courts of law, however, and therefore I am bound to dissent. Contrary to the majority’s holding, the record evidence here entirely fails to support a finding that licensee violated GR17a.
¶ 12. The issue at the heart of this case is whether the evidence supports the Board’s finding that licensee violated GR17a, which provides: “Licensees or [their] employees shall not serve alcoholic beverages to a person whom it would be reasonable to expect would be under the influence as a result of the amount of alcohol served to that person.” Department of Liquor Control General Regulation 17a, 4 Code of Vt. Rules 26 020 016-1 (2005).
*565¶ 13. The record evidence shows that the patron in question arrived at licensee’s establishment around 8:45 p.m. and left about three hours later, between 11:45 p.m. and midnight. According to the bartender, there were about forty people in the bar that night, close to capacity. The bartender testified that he served the patron an initial drink of Southern Comfort and orange juice in a sixteen-ounce glass, and a second identical drink about an hour later. The bartender also testified that he later discovered the second drink had been only partly consumed. The bartender observed the patron during the evening moving around the establishment, talking with other patrons and dancing. The bartender testified that the patron had no difficulty ordering or moving and showed no signs of intoxication. When the patron left, the bartender noticed that he sat in his car for a couple of minutes before driving off. A part-owner of the bar who was in the establishment that night also recalled seeing the patron dancing, and observed no signs of intoxication.
¶ 14. A chemist employed by the State testified that the patron’s blood alcohol concentration (BAC) at 3:30 a.m., two hours after the accident, was .155, or almost three times the legal limit. Using a relation-back formula, the chemist concluded that the patron’s BAC at the time he left the bar was .211, and at the time of the accident was .202. The chemist further testified that, based solely on the alcohol content of the two drinks served by the bartender, the patron’s BAC at the time of the accident would have been only .055.
¶ 15. The question for the Board thus became: what accounted for the patron’s extremely high BAC? Three possibilities — alone or in combination — presented themselves. First, the bartender could have served the patron much more than he admitted. In this regard, the chemist testified that, to achieve the patron’s measured BAC, he would have had to consume a total of six drinks like the ones described. Second, other persons could have purchased drinks for the patron that evening, although the bartender testified that he did not observe this occurring, and there was no other evidence to support the scenario. Third, the patron could have had an elevated BAC when he arrived at the bar. In this regard, a police videotape recorded later that evening, following the accident, reveals the patron telling the investigating officer that he had consumed eight or nine drinks in total that day, starting around 1:00 p.m. Assuming this to be the case, however, the State chemist testified that most of the alcohol consumed over the course of the day would have been eliminated before the patron arrived at the bar. Thus, the third possibility would depend on the patron having consumed much more alcohol earlier in the day than he admitted.
¶ 16. Which of these theories did the evidence support? The Board appears to have basically accepted the first, finding that “the drink count provided by the [licensee] is scientifically incompatible with the BAC values provided ... by the relation back process, and is therefore not credible,” and further finding that “[w]ith a BAC of .202 at the time of the crash, it is impossible to consider the testimony of [the bartender] about the number of drinks he claimed to have served to [the patron] that evening as credible.” The Board subsequently acknowledged the possibility of the second theory, as well, stating:
The Board concludes that materially all of the high BAC that [the patron] obtained was acquired while he was at the [licensee’s] premises. This was achieved either by drinks having been served directly [by the bartender] or by drinks having indirectly been served to [the pa*566tron] by allowing other patrons to buy two drinks without determining exactly where they are going or if they knew they were going to [the patron].
(Emphasis added.) Finally, the Board also appears to have recognized the possibility of the third option, noting that although the patron drove to the bar and did not appear to be intoxicated, both the chemist and licensee’s expert testified that “experienced, hard drinkers may be substantially intoxicated and yet be able to walk without giving gross signs of their condition,” and that alcohol the patron had earlier consumed may have “contributed to some degree to the elevated BAC that he had at the time of the crash.”
¶ 17. Two significant problems undermine the Board’s findings and conclusions. First, the Board fails to explain, based on the evidence, why any one theory is more plausible than the other two. The preponderance-of-the-evidence standard applicable in Board proceedings is not particularly demanding, but it does require proof that the existence of the contested fact is “more probable than not.” In re N.H., 168 Vt. 508, 512, 724 A.2d 467, 470 (1998). Where two or more inferences from the evidence are reasonable, this requires — at a minimum — that “the inference from the facts proved must be at least the more probable hypothesis, with reference to the possibility of other hypotheses.” Jackson v. True Temper Corp., 151 Vt. 592, 596, 563 A.2d 621, 623 (1989) (quotation omitted). Based on the evidence here, it is just as reasonable to infer that the patron’s high BAC resulted from him being well over the legal limit when he arrived at the bar and imbibing the two drinks he was admittedly served, as it is to infer that the high BAC resulted from him either being served six drinks by the bartender or obtaining additional drinks from others at the bar. Stated differently, the patron might have lied about the number of drinks he consumed that day; the bartender might have lied about the number of drinks he served the patron; or some other unknown person or persons might have purchased additional drinks for the patron. There was simply no evidence to establish one theory as more probable than the others.3
¶ 18. The Board appears to have resolved this evidentiary dilemma by concluding that any one of the theories would constitute a violation of the regulation in question. This is the second flaw in the Board’s reasoning. As noted, GR17a provides that a licensee or its employees *567“shall not serve alcoholic beverages to a person whom it would be reasonable to expect would be under the influence as a result of the amount of alcohol served to that person.” The regulation was plainly designed to prohibit a licensee from serving alcohol in sufficient quantity to result in intoxication, even if the person does not appear to be intoxicated. Thus, it would apply only to the first theory — that the bartender lied and actually served the patron six large drinks — and then only if there were any evidence to support it.
¶ 19. It would just as plainly not apply to the second theory, however, i.e., the scenario where the licensee serves alcohol to one patron who provides it to another. The rule prohibits service “to a person” where it is reasonable to expect that intoxication would result from the alcohol being served “to that person.” (Emphasis added.) While the Board may be correct that licensees have a “responsibility to make sure that an extra drink which may have been served to a patron [does] not fall into the hands of... any other patron who was a heavy drinker,” that responsibility does not arise from 6R17a, which simply does not by its terms apply to this situation.4
¶ 20. Finally, as to the third possibility, the Board concluded that licensee had a duty, under the circumstances, to determine the patron’s sobriety level immediately upon his arrival and to refuse service of even the first drink if it would result in his intoxication. The bartender testified that, about a week or two before the incident, he had refused to serve the patron in question after overhearing him admit that he had been drinking heavily that afternoon. Based on this one incident, the Board concluded that it was therefore reasonable to presume that the patron had a “propensity” to drink to excess before he arrived, and the bartender “should have assumed that [the patron] had been drinking before he arrived and likely to the same substantial extent, more or less, as he did before.” Based on that one prior incident, the Board wrote that “[i]t should further be presumed in such a case that [the patron] would arrive with sufficient alcohol in him so that he may very well be over the limit when he arrived, no matter how well disguised.” The Board concluded that, even if the patron did not appear to be intoxicated, closer examination would have disclosed signs of impairment, such as slurred speech, watery eyes, or degraded “fine motor skills.” This responsibility apparently applies notwithstanding the Board’s conclusion in this case that “there simply [wasn’t] any evidence . . . indicating that [the patron] was displaying any signs of intoxication when he entered.” I find the Board’s presumptions irrational.
¶21. Although it does not admit to creating a novel duty, the Board has in effect made licensees strictly liable — or very close to it — for every patron who has ever been intoxicated in their establishment. Under the Board’s reading of GR17a, it is not sufficient to shutoff a patron after serving a sufficient number of drinks to result in intoxication, as the regulation reasonably requires. Under *568the Board’s ruling, licensees now have a duty to closely examine every patron who enters their establishment — regardless of whether they appear to be intoxicated — if they have ever been previously intoxicated or admitted to having been so. Further, I find no evidence that Regulation 17a was designed for this purpose.
¶ 22. The rule prohibits service to a person “whom it would be reasonable to expect would be under the influence as a result of the amount of alcohol served to that person.” (Emphasis added.) Thus, even in a crowded bar, as this one was, it is certainly reasonable to require that a licensee not serve anyone who is either visibly intoxicated or who has been served enough alcohol to be intoxicated. To require that a licensee exercise hyper vigilance for every person who has ever previously exhibited signs of intoxication or (as here) merely admitted to heavy drinking, however, strikes me as unreasonable and unenforceable. Although this Court generally employs a deferential standard to an agency’s interpretation of its own regulation, we must nevertheless “endeavor to ensure that such deference does not result in ‘unjust, unreasonable or absurd’ consequences.” In re Verburg, 159 Vt. 161, 165, 616 A.2d 237, 239 (1992) (quoting O’Brien v. Island Corp., 157 Vt. 135, 139, 596 A.2d 1295, 1297 (1991)). That is precisely the result, in my view, of the Board’s ruling here, and I would expect it to lead to even more troubling cases in the future. Therefore, I respectfully dissent.
Although the record speaks for itself, I do not find that it supports the majority’s assertion that the patron “denied” having consumed substantial alcohol during the day. Ante, ¶ 9. To conclude otherwise, the majority suggests, is to “reject the Board’s findings,” but the Board itself relied on the investigating officer’s testimony that the patron admitted having consumed “eight or nine drinks” “over the day” and the Board expressly found, based upon its own viewing of the police videotape, that the patron “was questioned about how many drinks he had and he stated that he had some eight or nine drinks before the accident” and had “started drinking between 12:00 and 2:00 in the afternoon” with “some drinks in his car.” The majority further mischaracterizes the argument here in stating that “the dissent proposes that the multiple extra drinks” which the patron may have consumed at the bar “were all supplied by the other patrons of the bar.” Ante, ¶ 10. First, this possibility was raised by the Board,, not the dissent, and second, the dissent does not propose that this was, in fact, what occurred. The point here is simply that there was no more evidence to support or impeach this possibility as there was to support or impeach the possibility that the patron was served extra drinks by the bartender.
In response, the majority indicates that it does not “believe” that GR17a applies only where the licensee serves the person directly, and states its “view” that it is the licensee’s duty under the regulation to “monitor ‘the amount of alcohol served to that person.’ ” Ante, ¶ 10. Beliefs and views are not a substitute for analysis, however, and nothing in the text of the regulation suggests that it applies in circumstances where persons other than the “licensee or his or her employee” serves alcohol to a person whom it would be reasonable to expect would be under the influence as a result of the amount of alcohol “served to that person.” Regulation 17a.