I. BACKGROUND

On February 21, 2011, Paul Gemmink and his daughter, Christine, visited the[*48] Jay Peak ski resort in Jay, Vermont. The two skied down the Northwest Passage trail, with Christine preceding her father as she turned onto the Kokomo trail, which intersected the Northwest Passage trail. When Christine reached the base of the ski lift at the end of the trail, she noticed that her father had failed to follow her descent. Instead, a Jay Peak ski patroller would find Gemmink “combative and in obvious pain,” lying on his back by a tree on the left side of the Kokomo trail, near the Kokomo-Northwest Passage intersection. App’x at 31. Gemmink had been rendered unconscious and, though argumentative, could not recall or provide an account of the incident. Christine, however, had observed a ski jump situated near the trees on the right side of the intersection, leading her and her father to surmise that another patron “fl[ew] of[f] the jump” and collided with Gemmink. Id. at 30, 32. Gemmink suffered fractures to his left ribs and left transverse processes in the incident, injuries that, according to Gemmink, are usually attributable to a significant impact coming from right to left, and are therefore at least consistent with the theory that a skier jumped from the right of the intersection into Gemmink.

Proceeding pro se, Gemmink brought this action against Jay Peak to recover for injuries that he claims were sustained as a result of Jay Peak’s negligence. Specifically, Gemmink asserts that Jay Peak negligently permitted dangerous jumps on its ski trails and that, in consequence of such a constructed jump at the Kokomo-North-west Passage intersection, Gemmink suffered a collision with another skier resulting in harm to his left side. The District Court (Murtha, J.) granted Jay Peak’s motion 1 on for summary judgment, finding that Gemmink had failed to establish that Jay Peak’s alleged negligence was the cause of his injuries. Gemmink now appeals.

II. DISCUSSION

A.

This Court reviews a grant of summary judgment de novo. Amerex Group, Inc. v. Lexington Ins. Co., 678 F.3d 193, 199 (2d Cir.2012). Where, as here, the party opposing summary judgment bears the burden of proof at trial, summary judgment should be granted if the moving party can “point to an absence of evidence to support an essential element of the nonmoving party’s claim.” Goenaga v. March of Dimes Birth Defects Found., 51 F.3d 14, 18 (2d Cir. 1995). The court draws all inferences in favor of the nonmoving party, but the opposing party “must come forward with specific evidence demonstrating the existence of a genuine dispute of material fact.” Robinson v. Concentra Health Servs., Inc., 781 F.3d 42, 44 (2d Cir.2015).

Before the district court, it was assumed that the negligence of Jay Peak in the structure and maintenance of the jumps was sufficiently made out to survive summary judgment. Accordingly, for purposes of this opinion, we will assume ar-guendo that Gemmink has established such a potential basis for liability on the part of Jay Peak. The issue before us, then, is a not-infrequent one in torts cases: whether the plaintiff proffered sufficient evidence for a jury to find, more probably than not, that the ground for liability (here, the assumed negligence) was the cause of the plaintiffs injury.

As Professor Abraham has demonstrated in his recent article, a showing of cause-in-fact almost always involves circumstantial evidence. See Kenneth S. Abraham, Self-Proving Causation, 99 Va. L.Rev. 1811, 1815-16 (2013). Thus, in considering whether a plaintiff has proven causation, a trier of fact asks whether it is[*49] likely that the harm that occurred resulted from the negligence (or from another basis of liability) attributed to the defendant. In other words, is the reason that the defendant’s behavior is deemed risky, and the defendant deemed potentially hable, the harm that in fact occurred?

In such circumstances, as then-Chief Judge Cardozo set out in Martin v. Herzog, 228 N.Y. 164, 126 N.E. 814 (1920), a jury can assume that the injury occurred as the expected or ordinary result of the defendant’s conduct. Id. at 816. If for some reason it was not the ordinary result of the defendant’s conduct, that the “extraordinary” had occurred must be shown by the party wishing to counter causation. For example, if a defendant proprietor has failed to install lights on its stairways after dark, and a person coming down the stairs in the dark of night falls and injures himself, one can fairly assume that the failure to illuminate the stairs caused the injury. And it will be up to the defendant to show that something extraordinary happened, say, that an animal scampered up the stairs and tripped the injured person instead. In essence, the greater the risk that the defendant’s conduct will result in the harm the plaintiff suffered, the more likely that a jury will be allowed to find that such conduct was the cause of that harm.

In addition to considering the strength of the circumstantial evidence linking injury and harm, however, the cases dealing with questions of causation take into account two other factors. First, where one party has knowledge or access to information that renders that party better able than his adversary to explain what actually transpired, courts have tended to put the onus on that party to do so. This principle — that the party with superior knowledge bears the burden of coming forward with evidence — has always served as a basis of finding negligence under the doctrine of res ipsa loquitur. See, e.g., Griffen v. Manice, 166 N.Y. 188, 194-96, 59 N.E. 925 (1901). But it also serves as a basis for finding causation. See Williams v. Utica Coll. Of Syracuse Univ., 453 F.3d 112, 120-21 (2d Cir.2006); Williams v. KFC Nat. Mgmt. Co., 391 F.3d 411, 431-32 (2d Cir.2004) (Calabresi, J., concurring). Thus, the requirement that the plaintiff be able circumstantially to show a link between the expected risk of defendant’s conduct and what actually occurred tends to be greater when the plaintiff is better able to explain what happened, and is considerably less when, instead, it is the defendant who can better or more easily proffer evidence of what, in fact, occurred.

But cases of this sort also involve a third factor. Thus, in deciding whether sufficient proof of causation has been proffered to get to a jury, courts consider whether the law of the jurisdiction is indifferent as to error in one direction or the other. If an erroneous finding of causation is, in the law of the jurisdiction, more harmful than an erroneous finding of no causation, the requirements of circumstantial evidence and knowledge grow stronger. Conversely, where the law of the jurisdiction makes clear that an erroneous finding of no causation is more harmful, the requirements are diminished. Compare Williams v. Utica Coll. Of Syracuse Univ., 453 F.3d at 121 (finding summary judgment against plaintiff appropriate because, inter alia, New York courts placed only a minimal duty on the defendant to avert the type of harm incurred, which is “close to saying that if an error is to be made in this context, it is better made in favor of the defendant ”) (emphasis added), with Williams v. KFC Nat. Mgmt. Co., 391 F.3d at 432 (finding summary judgment against plaintiff inappropriate because, inter alia, of “the absence of any reason to[*50] prefer erring in favor of [the defendant] rather than the plaintiff’).

B.

With these three factors in mind, we turn to the case before us. The first factor favors the defendant. The causal link between Jay Peak’s assumed negligence in its maintenance of ski jumps and the injury incurred by the plaintiff is far too attenuated to sustain Gemmink’s claim. Our common experience does not tell us that this kind of lack of maintenance results in accidents of this sort with any frequency. And plaintiff has failed to proffer expert testimony suggesting a stronger link. See Human Rights Comm’n v. LaBrie, Inc., 164 Vt. 237, 668 A.2d 659, 667 (1995) (“Generally, expert ... testimony is required to support a finding of causation where the link is obscure and abstruse such that a layperson can have no well founded knowledge and can do no more than indulge in mere speculation.”) (internal quotation marks omitted).

The second is, at most, neutral. Neither Gemmink nor Jay Peak has greater knowledge or access to information concerning what actually happened on the Kokomo trail.

We turn, then, to the third factor: Is this an area where, in Vermont, liability of ski operators to skiers is close to strict, so that whether negligence was the cause of the alleged injury is a matter that, in uncertainty, should be decided in favor of the skier? Or is this an area where the risk of injury, even in the presence of negligence on the part of the ski operator, is assumed primarily by the skier, so that the requirement of causation is fairly placed on the skier (unless either (a) the evidentiary link between the evidence of negligence and causation of the kind of harm that occurred is particularly strong, or (b) the defendant is in a distinctly better position to tell us what happened)? Or, finally, is Vermont relatively indifferent to error in one direction or the other, offering no reason to favor either the plaintiff or defendant?

A review of Vermont law suggests that it follows the approach of symmetrical indifference. By statute, although assumption of risk has generally been subsumed in comparative negligence, 12 V.S.A. § 1036, it has been expressly retained as to sporting events, 12 V.S.A. § 1037. This would suggest that Vermont prefers to err on the side of finding no causation with respect to sport injuries like the one that here occurred. At the same time, however, the decision of whether the risk borne by the plaintiff in the sporting event was sufficiently “obvious and necessary” as to be assumed generally forms a jury question under Vermont law. See Estate of Frant v. Haystack Grp., Inc., 162 Vt. 11, 641 A.2d 765, 770-71 (1994) (rejecting the conclusion that “by enacting § 1037, the legislature intended to provide more protection from liability for ski areas” and stating that “ § 1037 is broad enough ... [that sjkiers should be deemed to assume only those skiing risks that the skiing industry is not reasonably required to prevent,” as determined by “a jury [applying] a contemporary sense of what constitutes an obvious or necessary risk”). Vermont’s approach stands in notable contrast both to Connecticut, where participants in sporting events rarely assume the risk of that participation, see, e.g., Jagger v. Mohawk Mtn. Ski Area, Inc., 269 Conn. 672, 849 A.2d 813, 827 (2004), and to New York, where assumption of risk is powerfully applied by courts to bar recovery by participants in sporting events, see, e.g., Martin v. New York, 64 A.D.3d 62, 878 N.Y.S.2d 823, 825-26 (3rd Dept.2009); N.Y. Gen. Obl. Law § 18-106. This contrast rein[*51] forces our conclusion that Vermont wants us to treat errors in this area pretty much symmetrically.

Consequently, we are left to infer causation, then, from only the placement of the ski jumps and the nature of Gemmink’s injuries. We cannot infer a causal link between Jay Peak’s assumed negligence in its maintenance of ski jumps and the injury incurred on the facts presented, and the plaintiff does not provide sufficient evidence to support a link between his injuries and alleged theory of causation. Under these circumstances, the district court was clearly correct in its holding that the evidence adduced by Gemmink was not sufficient to raise a question for the jury.

III. CONCLUSION

The judgment of the District Court is, therefore, AFFIRMED.

Judge POOLER joins only Parts I and 11(B) of the opinion.