This action arises out of the collapse on two different days of two separate sections of the roof of a cooper concentrator mill building located at the Granduc minesite in Tide Lake, British Columbia, Canada, and the refusal of two of the building’s four insurers, Hanover Insurance Company (“Hanover”), and Utica Mutual Insurance Company (“Utica”),1 to indemnify the building’s owner, Newmont Mines Limited (“Newmont”), for the resulting losses. Hanover and Utica appeal from a final judgment in the amount of $2,500,000 (Canadian), plus prejudgment interest computed from February 20, 1980, entered after a jury trial in the United States District Court for the Southern District of New York, 609 F.Supp. 295 (Sprizzo, J.). The two insurers were found liable for failure to pay, under the terms of property damage insurance policies issued by them, for the expenses incurred by Newmont in the repair of the damaged building. The insurers’ primary contentions on this appeal are that they are not liable because the insured failed to inform them of a material increase in the risk and that, even if they are liable, the two collapses constituted one occurrence and thus the maximum amounts payable under the policies “for any one occurrence” should not have been calculated separately for each collapse. For the reasons set forth below, we affirm.
I. BACKGROUND
The Granduc minesite is located near the Alaskan border in an area of British Columbia subject to heavy snowfalls and substantial snow accumulation. The centerpiece of the minesite and the focus of this lawsuit is the concentrator facility: a huge wedge-shaped, multi-function structure with a sloping roof nearly fifteen stories high at its peak. The concentrator is 442 feet long by 240 feet wide and has a large roof with a slope length of approximately 500 feet. Due to the size of the concentrator, structural design principles required that an expansion joint be installed in the middle of the building to permit expansion and contraction to occur without causing damage. Consequently, the concentrator was designed and constructed as two adjacent structures, each approximately 220 feet long and each capable of supporting vertical and horizontal loads independent of the other. The two structures are separated in the middle by a double set of columns, located at what has been designated line G.
Because this region receives the highest annual snowfall in the world — as much as 1200 inches per year — the concentrator facility was specially designed to act as a giant snow slide, which would deposit snow from the roof into a catch basin at the base of the building. The roof was slanted at a thirty-degree angle, with the belief that this slope, together with the natural convection of heat generated in the building from ordinary operations, would cause the *130snow to slide off the roof.2 Indeed, this system of removing snow from the concentrator’s roof was successful until the 1978-1979 winter.
Newmont owned and operated the concentrator from 1970 until mid-1978, when it ceased operations at the Granduc mine due to depressed economic conditions in the copper industry. Prior to closing the Granduc mine, and because the concentrator would no longer be in operation or heated, Newmont sought advice regarding appropriate snow clearance procedures from Swan Wooster Engineering (“Swan Wooster”), one of the engineering firms that participated in the design of the concentrator. Swan Wooster recommended periodic inspections of the concentrator and removal of snow accumulated in the catch basin at the base of the building. Newmont also discussed the cessation of operations with Allendale, one of the companies insuring the concentrator facility. By letter dated August 8, 1978, Newmont informed Allen-dale of its plans to permanently close the Granduc mine and its intended inspection and snow clearance program. No similar letter was sent to Hanover or Utica.
Thereafter, in accordance with Swan Wooster’s recommendations, Newmont personnel regularly inspected the concentrator and, when needed, removed accumulated snow from the base of the building. Beginning in mid-December 1978 and continuing thereafter, Newmont personnel observed that snow was not sliding from the roof. On March 1,1979, Kurt Dahlke, an employee of Newmont, inspected the concentrator, observed approximately eight to ten feet of snow on the roof, and reported: “Everything ok. No sliding, no build-up.”3
On the next inspection trip, on March 14th, Newmont personnel discovered that a section of the concentrator roof, measuring approximately 55 by 200 feet and located just below the midline of the concentrator (line G), had collapsed sometime between March 1st and March 14th. The collapse was due to the weight of ice and snow. On March 17th, a second section of the roof, measuring approximately eighty by seventy feet and located in the upper half of the building some seventy-five feet above the area of the first hole, collapsed due to the weight of ice and snow. Both collapses occurred in portions of the roof having a load capacity of 140 pounds per square foot.
Subsequently, on May 31, 1979, Newmont conveyed its title to the Granduc mine property, including the damaged concentrator, to Esso Resources Canada Limited (“Esso”) pursuant to a written agreement of purchase and sale. In a supplementary letter agreement, Newmont agreed to pursue its rights under its insurance policies and to hold any proceeds in trust for Esso. Thereafter, Esso repaired the damaged concentrator building at a total cost of approximately $6.6 million (Canadian).
The insurance policies issued by Hanover and Utica (as well as those issued by Adri*131atic and Allendale) insured Newmont’s interest in various properties located in twelve different states of the United States and in British Columbia. The four insurance companies provided “layered” insurance coverage against a variety of perils, including that of collapse caused by the weight of ice and snow. The first layer of insurance was provided by the Hanover and Adriatic policies, each of which covered fifty percent of the first $500,000 (in excess of a $25,000 deductible) for any one occurrence. The Utica policy provided intermediate coverage of $1,000,000 in excess of $525,000 for any one occurrence. The Allendale policy provided a third layer of coverage with a $1,525,000 deductible for any one occurrence and a limit of liability of $9,189,000 for the Granduc mine property-
The terms and conditions of the Hanover and Utica policies were set forth in identical typewritten “manuscript” policies prepared by Newmont’s insurance broker, Johnson & Higgins. Representatives of Hanover and Utica reviewed and approved the terms of the manuscript policies and appended certain standard-form statutory conditions; the companies then issued the policies. In so doing, neither Hanover nor Utica inspected, either personally or through a representative, any of Newmont’s properties covered by the policies. Neither company made any inquiry of Newmont or its brokers regarding the properties, any special climatic conditions at any of the sites, or the design specifications of the buildings. Instead, Hanover and Utica were content in their knowledge that Allendale, the excess carrier, provided fire prevention and engineering services to Newmont and otherwise monitored the Granduc properties.
Newmont filed proof of claim with the four insurance companies in January of 1980. All four companies refused to pay their respective apportioned shares of the claimed loss by the date called for in the policies, February 20, 1980, and thus Newmont and Esso commenced this suit on February 29, 1980. Adriatic and Allendale reached settlements with Newmont and Esso and were dismissed from the action prior to trial. Hanover and Utica denied liability for the loss and asserted several affirmative defenses. They contended at trial — as they do on appeal — that Newmont’s cessation of operations at the Granduc mine in 1978 and the consequent absence of heat in the concentrator building constituted a material change in the risk insured by them. They further maintained that under British Columbia’s Insurance Act they were entitled to notice of a material change in the risk and that the failure of Newmont to give such notice to them resulted in the suspension of insurance coverage. Hanover and Utica also asserted that the collapses of the two sections of the roof constituted one occurrence rather than two because both were caused by the same accumulation of snow and ice on the roof.4 Finally, both companies argued that Newmont and Esso failed entirely to prove the amount of damages attributable to the alleged second occurrence.
In response to special interrogatories, the jury rejected each of these contentions and concluded that although there was a change in the risk it was not material and that, in any event, Hanover and Utica had knowingly waived any obligation on the part of Newmont to notify them of a material change in risk. The jury also found that the collapses of the two sections of the concentrator roof constituted two separate occurrences and that the cost of repairing the damage caused by the first collapse was $1,936,606 (Canadian) and the cost of repairing the damage caused by the second collapse was $4,663,394 (Canadian).
Hanover and Utica moved for judgment notwithstanding the verdict or, alternative*132ly, for a new trial. By order dated December 11, 1984, Judge Sprizzo denied their motion on the ground that sufficient evidence was introduced at trial to support the jury’s findings, and requested additional briefing as to the appropriate date and exchange rate for converting the jury’s award from Canadian dollars into United States dollars for purposes of entering judgment. After the additional briefing was completed, the district court held that the “breach-day” rule applied and that the exchange rate prevailing on February 20, 1980 (the day Utica and Hanover became obligated under their policies to pay the claims) should be used for converting the award into United States dollars. Newmont Mines Limited v. Adriatic Insurance Co., 609 F.Supp. 295, 298 (S.D.N.Y. 1985). Judgment ultimately was entered in favor of Newmont and Esso for $2,000,000 (Canadian) as against Utica and $500,000 (Canadian) as against Hanover, together with prejudgment interest on both amounts from February 20, 1980.5
Hanover and Utica appeal, contending: (1) Newmont’s cessation of operations and heat at Granduc constituted a material change in the insured risk, which released them from all liability for the losses; (2) the two collapses constituted a single “occurrence” and thus the maximum amounts payable under the policies “for any one occurrence” should not have been calculated separately for each collapse; (3) Newmont and Esso failed to prove the amount of damages attributable to each of the alleged two occurrences; and (4) the amount of the judgment should have been converted from Canadian dollars to United States dollars at the exchange rate prevailing on the date of judgment rather than the rate in effect on the date of the alleged breach.
II. DISCUSSION
A. Material Change in the Risk
Hanover’s and Utica’s pivotal argument regarding their liability is that the cessation of heat at Granduc was a change material to the risk insured by them, which released them from all liability for the losses. Under British Columbian law, which is agreed to govern the substantive issues in this action, “[a]ny change material to the risk and within the control and knowledge of the insured shall avoid the contract as to the part affected thereby, unless the change is promptly notified in writing to the insuror.” Insurance Act, B.C.Rev.Stat. ch. 200 § 220, Statutory Condition No. 4 (1979). In response to special interrogatories four and five, the jury stated that while the cessation of operations and heat at the Granduc mine was a change in the insured risk, it was not a material change.
Hanover and Utica take sharp issue with that finding. In order to overcome the jury’s finding, however, Hanover and Utica bear a heavy burden. The evidence presented to the jury must be viewed in the light most favorable to Newmont and all inferences must be drawn in its favor. As we declared in Mallis v. Bankers Trust Co., 717 F.2d 683, 688-89 (2d Cir.1983):
Simply stated, judgment n.o.v. should be granted only when
(1) There is such a complete absence of evidence supporting the verdict that the jury’s findings could only have been the result of sheer surmise and conjecture, or
(2) There is such an overwhelming amount of evidence in favor of the movant that reasonable and fair minded men could not arrive at a verdict against him.
Similarly, for Hanover and Utica to prevail on their motion for a new trial, the trial court must be convinced that “the jury has reached ‘a seriously erroneous result’ or the verdict is a ‘miscarriage of justice.’ ” Mallis, 717 F.2d at 691 (quoting Bevevino *133v. Saydjari, 574 F.2d 676, 684 (2d Cir. 1978)). Finally, our own review of a denial of a motion for a new trial is limited: We will reverse the district court’s decision “only if the trial judge clearly abused his discretion.” Philip v. Mayer, Rothkopf Industries, Inc., 635 F.2d 1056, 1063 (2d Cir. 1980) (citing Diapulse Corp. of America v. Birtcher Corp., 362 F.2d 736, 744 (2d Cir.), cert. dismissed, 385 U.S. 801, 87 S.Ct. 9,17 L.Ed.2d 48 (1966)). To the extent that a new trial was sought on the ground that the verdict was against the weight of the evidence, we have disclaimed the authority to review a ruling on such a motion. Portman v. American Home Products Corp., 201 F.2d 847, 848 (2d Cir.1953); see Mallis v. Bankers Trust Co., supra, 717 F.2d at 690 n. 11; Compton v. Luckenbach Overseas Corp., 425 F.2d 1130, 1132-33 & n. 2 (2d Cir.), cert. denied, 400 U.S. 916, 91 5. Ct. 175, 27 L.Ed.2d 155 (1970).
Hanover and Utica simply have been unable to overcome this heavy burden or to contradict the substantial evidence tending to show that the change was not material to them. The parties agree that the question of materiality was properly left to the jury. See Insurance Act, B.C. Rev.Stat. ch. 200 § 15(2) (1979) (“The question of materiality is one of fact.”). In addition, the precise issue framed for the jury in the charge was not merely whether the change increased the risk of loss, but rather whether Utica or Hanover reasonably would have attached importance to the cessation of operations and heat at the concentrator facility “in determining whether to cancel their policies or to require payment of additional premiums or require changes or modifications of the risk in order for them to continue their coverage.”
We believe that there are four facts which clearly are sufficient to sustain the jury’s verdict. First, neither Hanover nor Utica showed any inclination at the time the policies were issued to inspect the Granduc property; to inquire about any special climatic conditions there; to ascertain the procedures for removing snow from the concentrator roof; or to examine the concentrator building or its design specifications. Accordingly, the jury was entitled to infer that Hanover and Utica agreed to insure the building without regard to the snow removal procedures being used or whether the building was heated.
Second, the insurance policies themselves expressly permitted Newmont to cease operations at its various locations, including Granduc. The only condition placed on Newmont before it could cease operations was that normal levels of fire protection and watch service be maintained. Newmont concedes that normal levels of these services were not maintained; nevertheless, this contractual clause is significant because it contains no reference to maintaining temperatures or continuing snow removal procedures. Consequently, the jury could have inferred that neither Hanover nor Utica was concerned, in the event of a shut-down, with whether heat was continued at the Granduc minesite.
Third, the insurance policies covering the Granduc property were blanket policies covering twenty-five different locations with widely varying climatic conditions; yet, no distinctions were made with respect to particular sites. It is reasonable, therefore, to conclude that Hanover and Utica were concerned with an “overall assessment” of the risk, rather than with a particularized assessment of the increased risk at Granduc due to the cessation of heat.
Finally, evidence was presented that Allendale, the excess insurer of the Granduc property, continued its insurance in force, without any increase in premium or change in terms, after being notified by Newmont that operations at the Granduc Mine were being discontinued. This evidence is undeniably probative on the question of what significance a reasonable insurer would attach to the cessation of operations and consequent absence of heat at the Granduc facility.6
*134In sum, the evidence presented to the jury was sufficient to support its verdict that the change was not material. It has long been settled that “[cjourts are not free to reweigh the evidence and set aside the jury verdict merely because the jury could have drawn different inferences or conclusions or because judges feel that other results are more reasonable.” Tennant v. Peoria & Pekin Union Railway, 321 U.S. 29, 35, 64 S.Ct. 409, 412, 88 L.Ed. 520 (1944). It is fair to say, in retrospect, that the absence of heat in the concentrator increased the risk of roof failure. Nevertheless, it was for the jury to decide whether or not that change to the risk was material. We will not upset their determination that it was not. In light of this conclusion, we do not reach the question whether the evidence supported the jury’s conclusion that Hanover and Utica waived whatever rights they had to be informed of the cessation of operations and heat at Granduc.7
B. Number of Occurrences
Since the liability of both Hanover and Utica is limited to a dollar amount per occurrence (or for “any one loss, disaster, or casualty”), the characterization of the two collapses as either one or two occurrences is significant to any assessment of the insurers’ total exposure under the policies. Hanover and Utica contend that since the two collapses of the roof were caused by the same condition (excessive snow accumulating on the roof of the concentrator), both collapses should be considered to constitute only one occurrence. They claim that “[t]he legal test for determining whether there is one or more than one occurrence turns upon the cause of the damage and not the items damaged.” Judge Sprizzo instructed the jury, without objection by counsel for Hanover or Utica, that
it is for you to decide whether or not the losses which are alleged to have occurred or the loss that’s alleged to have occurred in this case was the result of a single, continuous event or incident, or whether or not it was the result of two separate incidents.
If you find that the collapse of the two sections of the roof was a single, continuous event or incident, then the collapse constituted a single occurrence — and there would be only one loss.
If, on the other hand, you find that the collapse of the two sections of the roof constituted separate events or incidents that were not causally related, then of course you would have two separate losses.
In response to special interrogatory seven, the jury found that there were two occur*135rences within the meaning of the policies. We find that the above charge was an accurate construction of the terms of the insurance policies and that the evidence developed at trial fully supported a finding that the collapse on two different days of two separate sections of the concentrator roof constituted two occurrences within the guidelines of the foregoing instruction.
1. Construction of the Term “Occurrence”
The cardinal principle for the construction and interpretation of insurance contracts — as with all contracts — is that the intentions of the parties should control. See, e.g., 29 N.Y.Jur. Insurance §§ 593-594 (1963). Unless otherwise indicated, words should be given the meanings ordinarily ascribed to them and absurd results should be avoided. As we have stated before, the meaning of particular language found in insurance policies should be examined “in light of the business purposes sought to be achieved by the parties and the plain meaning of the words chosen by them to effect those purposes.” Champion International Corporation v. Continental Casualty Co., 546 F.2d 502, 505 (2d Cir.1976), cert. denied, 434 U.S. 819, 98 S.Ct. 59, 54 L.Ed.2d 75 (1977).
Both policies at issue here contain monetary limitations on recovery on a per occurrence basis. The Hanover policy states that Hanover “shall not be liable for more than its proportion of $250,000 P/O $500,000 for any one loss, disaster, or casualty by any of the perils insured hereunder.” (emphasis added). The Utica policy states that Utica “shall not be liable for more than its proportion of $1,000,000 excess of loss of $500,000 for any one occurrence in addition to the deductible under the primary insurance [$25,000] ... by any of the perils insured hereunder.” (emphasis added). No party to this litigation has suggested that the difference between the italicized language in each of the two policies requires dissimilar analysis, and the parties have focused their arguments on the term “occurrence.” Accordingly, we direct our attention to that term.
Neither policy provides any further definition of “occurrence” (or of “loss, disaster, or casualty”). The term “occurrence” ordinarily is understood to denote “something that takes place,” especially “something that happens unexpectedly and without design.” Webster’s Third New International Dictionary 1561 (unabridged ed. 1981); accord Black’s Law Dictionary 974 (5th ed. 1979) (“A coming or happening. Any incident or event, especially one that happens without being designed or expected.”). In his charge, Judge Sprizzo used event and incident interchangeably with occurrence, and indeed those words are synonymous with occurrence. Thus, if we were to go no further, it would be clear that Judge Sprizzo’s charge represented an accurate construction of the plain meaning of the language used in these insurance policies.
Hanover and Utica, however, argue that the term “occurrence” has taken on a specialized and particular meaning in the insurance business and that this meaning should be recognized as controlling in this case. If the term “occurrence,” as used in property damage policies, in fact had acquired a gloss on its meaning through prior business usage and judicial construction, we would be compelled to consider that understanding in ascertaining the meaning intended by the parties in using it. See Standard Oil Company v. United States, 340 U.S. 54, 60 & n. 17, 71 S.Ct. 135, 138 & n. 17, 95 L.Ed. 68 (1950); Maryland Insurance Co. v. Woods, 10 U.S. (6 Cranch) 29, 45, 3 L.Ed. 143 (1810). We are not persuaded, however, that the term “occurrence” has obtained any such specialized or singular meaning in the context of property insurance.
In construing the term “occurrence” as used in liability insurance policies, courts generally “have concluded ... the number of occurrences for purposes of applying coverage limitations is determined by reference to the cause or causes of the damage and not the number of injuries or claims.” Michigan Chemical Corp. v. American Home Assurance Co., 728 F.2d 374, 379 (6th Cir.1984); accord Home Indemnity *136Co. v. Mobile, 749 F.2d 659 (llth Cir.1984); Appalachian Insurance Co. v. Liberty Mutual Insurance Co., 676 F.2d 56, 61 (3d Cir.1982); Cargill Inc. v. Liberty Mutual Insurance Co., 488 F.Supp. 49 (D.Minn. 1979), aff'd, 621 F.2d 275 (8th Cir.1980) (per curiam); Owens-Illinois Inc. v. Aetna Casualty & Surety Co., 597 F.Supp. 1515 (D.D.C.1984).8 But see Arthur A. Johnson Corp. v. Indemnity Insurance Co. of North America, 7 N.Y.2d 222, 227-30, 196 N.Y.S.2d 678, 164 N.E.2d 704 (1959) (rejecting any single definition of occurrence); Hartford Accident & Indemnity Co. v. Wesolowski, 33 N.Y.2d 169, 173, 350 N.Y. S.2d 895, 305 N.E.2d 907 (1973) (same).
In the liability context, we have found that the selection of a “per occurrence” basis and the corresponding rejection of a “per claim” basis indicated that the liability policy was not intended to gauge coverage on the basis of individual accidents giving rise to claims, but rather on the underlying circumstances which resulted in the claims for damages. Champion International, 546 F.2d at 505-06. A liability policy is intended to protect an individual or a business from liability for their tortious conduct. Consequently, since that is the “business purpose sought to be achieved by the parties,” it is eminently reasonable to look to the underlying conduct or cause of that liability.
On the other hand, when construing a property damage policy, as we are here, the business purpose sought to be achieved by the parties is considerably different. The goal of such a policy, simply stated, is to provide financial protection against damage to property. In accordance with this purpose, the parties here must have intended to provide coverage for property damage each time it occurred unexpectedly and without design, unless the damage occurring at one point in time was merely part of a single, continuous event that already had caused other damage.9 If we looked only *137to the cause of the damage, as Hanover and Utica would have us do, it would lead to the absurd result that the collapse of the roof of one building caused on one day by a snowstorm would be considered the same occurrence as an entirely unrelated collapse days later of the roof of another building located several miles away simply because the same snowstorm was the cause of both collapses. Indeed, even if the damage caused by the first collapse was repaired completely and a claim filed with the insurer before the second collapse occurred, they both would be considered one occurrence by Hanover and Utica. Accordingly, we hold that Judge Sprizzo’s charge faithfully interpreted the term occurrence in light of the business purpose sought to be achieved by the parties, the plain and ordinary meaning of the language used, and prior judicial constructions of similar terms in other policies. The question remaining, therefore, is whether the evidence presented at trial supported the jury’s finding that the collapse of two sections of the concentrator roof constituted two occurrences.
2. Evidence of Two Occurrences
Several facts support the jury’s finding that there were two occurrences. To begin with, the two collapses occurred at least three, and perhaps as many as seventeen, days apart. In addition, David Calder, Newmont’s structural engineering expert, testified that the concentrator was designed as two independent structures. He stated that the concentrator was separated in the middle by a double set of columns, located at line G. He further testified that none of the columns, trusses, purlins, or even the roof itself above line G was connected in any way to the structure below line G. In addition, he testified that the collapse of the portion of the roof north of line G, which occurred sometime between March 1st and March 14th, did not cause or contribute to the collapse of the roof south of line G, which occurred on March 17th. Moreover, he testified that the failure in the structure which resulted in the second collapse was of a type that occurs instantaneously, and thus the second collapse in all likelihood occurred entirely on March 17th and was not merely a continuation of the earlier collapse. Calder also testified that temperature changes, which necessarily took place, may have been a contributing cause to the second collapse. Finally, Calder opined that the two collapses were separate events.
Contrary to the assertions of Hanover and Utica, this testimony was not intended, and we have not accepted it, as proof that the concentrator actually consisted of two separate buildings. Rather, this testimony simply indicates that the second collapse was not a consequence of the first because each half of the concentrator was structurally distinct and capable of supporting a weight of ice and snow independent of the other. The jury had before it all the evidence on this issue; it heard the arguments presented by both sides; and Judge Sprizzo’s charge was accurate. In light of the substantial deference which we afford a jury’s findings, Mallis, 717 F.2d at 688-89, 691, we will not overturn its conclusion here that the two collapses were two occurrences within the meaning of the policies.
C. Allocation of Damages
This issue requires little discussion. Hanover and Utica contend that, even if there were two occurrences, Newmont failed to prove the amount of damages that should be allocated to each occurrence. Of course, due to the arrangement of the layered insurance coverage provided here, if Newmont and Esso were to receive full recovery on a two-occurrence basis, they had to show that at least $1,525,000 in repair costs was attributable to each collapse.
Because the two collapses were discovered only three days apart, they were repaired simultaneously and no effort was made at that time to allocate repair costs between the two collapses. The parties stipulated that the aggregate cost of repairing the building was $6,600,000 (Canadian). In addition, Newmont introduced *138the estímate of its engineering expert, Calder, that the cost of repairing the hole in the lower portion of the roof was $1,936,-606; the balance of the aggregate $6,600,-000 (Canadian) cost of repair, it followed, could be attributed to the second collapse. Hence, the jury’s finding of damages in the amount of $4,663,394 (Canadian) for the second collapse was arrived at not by speculation, as Hanover and Utica contend, but by simple subtraction ($6,600,000 — $1,936,-606). The jury plainly was entitled to base its conclusions on Calder’s estimate, see W.L. Hailey & Co. v. County of Niagara, 388 F.2d 746, 753 (2d Cir.1967), and we find that the subtraction method described above provided a rational basis for computing damages attributable to the second occurrence with an acceptable degree of certainty, see Lexington Products, Ltd. v. B.D. Communications, Inc., 677 F.2d 251, 253 (2d Cir.1982).
D. Conversion of Canadian to United States Dollars
Under the insurance policies at issue here, claims arising in Canada were to be paid in Canadian currency. The jury’s verdict therefore was in Canadian dollars. Since monetary exchange rates fluctuate over time, the District Court, in deciding how to convert the jury’s award into United States dollars, had to determine whether the appropriate rate for conversion was the rate prevailing on the date judgment was entered (“judgment-day rule”) or the one in effect on the date the breach of contract occurred (“breach-day rule”). Newmont and Esso argued in the district court that the law of New York, which follows the breach-day rule, governed the currency conversion issue. Hanover and Utica, on the other hand, contended that British Columbian law governed and required application of the rate of exchange in effect on the date of judgment. Judge Sprizzo determined that he did not have to decide whether New York or British Columbian law applied, because both would require application of the rule that the jury’s award should be converted into United States currency as of the date of the breach.
We have held that a federal court sitting in diversity must apply the currency conversion rule employed by the courts of the state in which the action was brought. Vishipco Line v. Chase Manhattan Bank, N.A., 660 F.2d 854, 865 (2d Cir.1981), cert. denied, 459 U.S. 976, 103 S.Ct. 313, 74 L.Ed.2d 291 (1982). Here, the action was commenced in the State of New York, and thus we look to the law of that state for the appropriate rule.
As a general rule, New York follows the breach-day rule. Id. at 866; Dougherty v. Equitable Life Assurance Society, 266 N.Y. 71, 97-98, 193 N.E. 897 (1934); Parker v. Hoppe, 257 N.Y. 333, 339-40, 178 N.E. 550 (1931). More important, New York apparently follows this rule even in those cases where foreign law is applied to determine the substantive rights of the parties. See Parker, 257 N.Y. at 339-40, 178 N.E. 550; Brill v. Chase Manhattan Bank, 14 A.D.2d 852, 852, 220 N.Y.S.2d 903, 904 (1st Dep’t 1961) (per curiam); Taubenfeld v. Taubenfeld, 198 Misc. 108, 110, 97 N.Y.S.2d 158, 160-61 (Sup.Ct.N.Y. County 1950). As we stated in Vishipco, “[ujntil the New York courts choose to change their position on this question, we are bound to apply the breach-day rule, and do so here.” 660 F.2d at 866-67.
We therefore conclude that Judge Sprizzo correctly ordered the jury’s verdict in Canadian dollars to be converted into United States dollars at the exchange rate prevailing on the date Hanover and Utica became obligated under the policies to pay the claims.
III. CONCLUSION
Accordingly, the judgment appealed from is affirmed.
. The other two insurers, Adriatic Insurance Company (“Adriatic”) and Allendale Mutual Insurance Company ("Allendale”), reached settlements with the building’s owner and were dismissed from the action prior to trial.
. At a thirty degree slope, snow generally will slide down. In order to facilitate and thereby ensure such sliding, however, heat is applied to the roof surface.
. Despite the eight to ten feet of snow which had accumulated on the roof, there was not necessarily any cause for alarm on March 1st. The roof had been designed with a load capacity of 140 pounds per square foot in the upper two-thirds of the roof and 280 pounds per square foot in the lower third. In addition, the roof had a safety factor of fifty to sixty-seven percent above the stated load capacity.
New fallen snow has a ten percent density weight, or six pounds per cubic foot. As snow rests on the ground (or on a roof), it compresses, ultimately increasing its density weight to 25 pounds per cubic foot. A snow expert consulted by Newmont in 1966 estimated that snow accumulating on the roof would have a 25% density weight, or 15 pounds per cubic foot. Newmont, in a 1977 document entitled "Snow Clearance Program Granduc Concentrator" estimated that, at an average weight of 12‘/2 pounds per cubic foot, 11.2 feet of snow could accumulate safely on the 140 pounds per square foot portion of the roof. Moreover, this figure did not take into account the safety factor. Finally, although it was determined after the second collapse that blocks of snow, which apparently had come from the roof, weighed 25'/2 pounds per cubic foot, Dahlke had reported before either collapse that the snow on the roof was dry and powdery, and thus presumably not very dense.
. Hanover and Utica predicate this argument on their claim that no new snow fell at the minesite between the first and second collapses. Although the record indicates that no new snow fell at Granduc between March 14th and March 17th, no similar evidentiary support is available for the period preceding March 14th. Since it is impossible to determine precisely when between the 1st and the 14th the first collapse occurred, it cannot be said that no new snow fell between the first collapse and the second collapse.
. The exchange rate for Canadian dollars posted by the Federal Reserve Bank on February 20, 1980 was one Canadian dollar to 86.5652 United States cents. Thus, Hanover was ordered to pay Newmont $432,826 (U.S.) plus interest, and Utica was ordered to pay $1,731,304 (U.S.) together with interest.
. Although Hanover and Utica concede that this evidence supports the jury’s conclusion that the change in risk was not material, they contend that it was irrelevant and prejudicial and thus *134should not have been presented to the jury. We disagree.
This evidence is indisputably relevant to the materiality issue, not only because Allendale was an insurer of the same property, but also because Hanover and Utica admittedly relied on Allendale to assess the risk and monitor the property. Although Hanover and Utica are correct in noting that Allendale, as the excess carrier, was in a different position with regard to the risk than they were, this distinction goes to the strength of the evidence, not its admissibility. In addition, any possible prejudice from the admission of this evidence could have been mitigated by highlighting this distinction for the jury.
Judge Sprizzo, after weighing the probative value of this evidence against any possible prejudice from its admission, decided to admit it into evidence. In so doing, he certainly was well within the "wide discretion given to trial , judges in making evidentiary rulings.” United States v. McGrath, 558 F.2d 1102, 1107 (2d Cir. 1977), cert, denied, 434 U.S. 1064, 98 S.Ct. 1239, 55 L.Ed.2d 765 (1978).
. If we were to reach the waiver issue, we would have little trouble upholding the jury’s finding that Hanover and Utica knowingly waived their right to be notified of any changes in the risks insured. Indeed, their failure to have statutory condition number four printed on the policies, by itself, is sufficient to support the jury’s verdict. Section 220(1) of the Insurance Act requires that the statutory conditions “shall be printed on every policy with the heading ‘Statutory Conditions'" and further provides that "no variation or omission of ... any statutory condition shall be binding on the insured." Insurance Act, B.C.Rev.Stat. Ch. 200 § 220(1) (1979); see also Canadian Acceptance Corp. Ltd. v. Indemnity Marine Assurance Co., Ltd., 15 W.W.R. 322, 325-26 (B.C.Ct.App.1955) (failure to print statutory conditions on policy barred insurer from relying upon those conditions to insured’s detriment).
. In each of the cases in which a court looked to the cause or causes of the damages in interpreting "occurrence," the issue was the meaning of specific policy definitions of "occurrence" in the context of the particular facts presented in each case. For instance, the liability insurance policies at issue in Michigan Chemical stated that “[a]ll damages arising out of such exposure to substantially the same general conditions shall be considered as arising out of one occurrence” and “[a]ll such exposure to substantially the same general conditions existing at or emanating from one premises location shall be deemed one occurrence.” 728 F.2d at 378.
In contrast, the Hanover and Utica policies do not contain any definition of "occurrence.” Instead, under the heading “Perils Excluded,” they contain a definition of "collapse,” which is defined as the “caving in or falling inwards or outwards of the building or structure ... or material part thereof____” While this definition was not drafted in order to answer the question faced here, we nevertheless believe it is pertinent to an assessment of whether the parties, at the time the policies were issued, would have construed two collapses to constitute only one occurrence. Indeed, this definition is strong evidence that the parties considered each collapse to be an independent event or occurrence.
. Hanover and Utica have cited the Ninth Circuit’s decision in Royal Indemnity Co. v. Kaiser Aluminum and Chemical Corp., 516 F.2d 1067 (9th Cir.1975), as support for their claim that we should define "occurrence" by looking to the cause of the collapses. In Royal, the court was faced with a property damage insurance policy, but the precise question before the Ninth Circuit was whether an insurer could consider all losses caused by one hurricane to be separate accidents for purposes of adjusting the insured's premium. The policy provided that the maximum “Incurred Losses" which could be considered in premium adjustment for any one accident were $50,000. During the term of the policy, the insured had suffered nearly four million dollars in losses as a result of hurricane damage to four of its plants in the New Orleans area.
The policy defined “one accident” as including “all resultant or con-commitant [sic] Accidents, whether to one or more than one Object, or to part of an Object.” Evidence on the question of what the parties intended by the “one accident” clause was received by a magistrate, and this evidence supported the insured’s claim that the hurricane, with all of the losses it caused, was one accident. Relying on this factual determination, and on the principle that any ambiguity should be resolved by a construction favoring the insured, the court held that all of the losses caused by the single hurricane resulted from one accident. Thus, the insurer was permitted to include only $50,000 as an incurred loss in determining the amount of the adjusted premium. This case, if pertinent at all, supports our determination that the meaning of “occurrence” must be interpreted in the context of the specific policy and facts of this case.