NATHANIEL R. JONES, Circuit Judge.
Plaintiff-Appellant/Cross-Appellee Owens-Illinois, Inc. (“Owens”) appeals and Defendant-Appellee/Cross-Appellant Aet-na Casualty & Surety Co. (“Aetna”) cross-appeals the final judgment in this action for reimbursement of the costs of responding to the presence of toxic materials on real property. We affirm.
I
From the early 1940s to November 20, 1970, the Hilfinger Corp. (“Hilfinger”) operated a manufacturing plant at 1800 West-wood Avenue in Toledo, Ohio. During this time, zinc die casting and plating, including chrome plating, was conducted on the premises. Hilfinger owned the property until the American Smelting and Refinery Co. (“ASARCO”) bought it in 1963. Hil-finger reacquired title in 1971, and transferred ownership to Owens on June 2,1971, pursuant to the terms of a purchase agreement dated December 8, 1970. Owens used the premises solely for warehousing and storage until the buildings on the premises were demolished in 1983.
In December 1983, greenish snow and water were discovered on the property. It was subsequently determined that hexava-lent chromium, a hazardous substance, had been deposited in the soil. After examining several alternatives to dealing with this problem, Owens decided to “cap” the contaminated area with an impermeable surface. The Ohio Environmental Protection Agency approved this method of remedial action in July 1987.
On August 20, 1985, to recover the significant costs of remediation, Owens filed a complaint in the United States District Court for the Northern District of Ohio. The complaint was initially brought against Lehigh Valley Industries, Inc., “the sole shareholder of and successor in interest to Hilfinger.” J.A. at 14 (Complaint at 2). Owens sought a declaratory judgment that Lehigh Valley Industries, Inc. was liable for all costs incurred by Owens relating to the response to the chromium contamination under the Comprehensive Environmen[867]*867tal Response, Compensation and Liability-Act of 1980, 42 U.S.C. §§ 9601-9675 (1988 & Supp. II 1990). Owens also sought, inter alia, $2,000,000 for breach of contract; $2,000,000 for compensatory damages as a result of “fraudulent misrepresentation”; and $5,000,000 in punitive damages. J.A. at 19 (Complaint at 7).
On December 5, 1985, Owens amended its complaint to include Hilfinger and Le-high Castings, Inc. (“the parent company of Hilfinger and a subsidiary of Lehigh Valley [Industries, Inc.],” Owens’ Br. at 1) as Defendants. On August 24, 1987, AS-ARCO was also added as a Defendant. On May 9, 1988, Owens reached a settlement with Lehigh Valley Industries, Inc., Lehigh Castings, Inc., ASARCO, and their insurers.
On April 28, 1988, Aetna was notified of Owens’ action, and was requested to defend and indemnify Hilfinger with respect to Owens’ claims. On October 6, 1988, Aetna responded that it would not defend and would deny coverage.
A trial was held to assess Hilfinger’s liability to Owens. On January 13, 1989, the district court found that the chromium “was introduced into the soil on the Property by a continuous or repeated wetting over a period of several years,” J.A. at 105 (Findings of Fact ¶ 15) (Jan. 1989), “[during the time Hilfinger Corporation was an owner and operator of the Property,” id at 109 (Conclusions of Law ¶ 5) (Jan. 1989), and that this chromium posed a health threat to humans. Finding Owens’ response to be reasonable, necessary and consistent with the National Contingency Plan, the court awarded Owens $1,469,322.20, plus interest and costs.
On July 18, 1989, Owens filed a supplemental petition with the district court, claiming that $769,322.20 of the judgment, plus interest and costs, remained unsatisfied after thirty days from the date of judgment. Proceeding under Section 3929.06 of the Ohio Revised Code Annotated (Baldwin 1989), it sought to recover this amount from Aetna, Hilfinger’s erstwhile insurer.
On November 26, 1990, a trial commenced to determine Aetna’s insurance responsibilities relating to Owens’ response costs. Of chief concern for purposes of this appeal was the evidence (or lack thereof) relating to how and when the chromium got into the soil, and the scope of Aetna’s coverage.
On the issue of how and when the chromium got into the soil, counsel for Owens set the stage, stating:
The testimony in the underlying case was that the chromium was put into the ground over a long period of time. It did not occur as a result of a single instance, but was a continuous progressive wetting. There will be no evidence as to how the chromium got into the ground. We don’t know.
J.A. at 927.
Howard W. Hilfinger III, general manager of Hilfinger for eight to ten years, testified for Owens concerning Hilfinger’s manufacturing operations at the Westwood Avenue premises. He agreed that there was “[n]o question that Hilfinger knew chromium was going into the sewer system, in connection with [its] operation” and that “[tjhat is the way the system was designed in the normal course of its operation....” Id at 961. He testified that spills and leaks of chromium onto the plant floor occurred. When the spills occurred, “[the whole area] was probably diked with four or six inch square wooden members that surrounded the area, and one area at one part of that surrounded area would have been this outlet to the drain.” Id at 959. Mr. Hilfinger maintained, however, that it was never Hilfinger’s intention, nor did the company expect to contaminate its premises with chromium.
Stephen F. Ahern, former counsel for Lehigh Valley Industries, Inc., Lehigh Castings, Inc., and Hilfinger, testified that, at the time Owens purchased the Hilfinger plant, Owens knew of the existence of “cracks in the cement in the area where plating operations were undertaken.” Id at 969.
James Mack, who supervised the investigation of the chromium contamination at [868]*868the Westwood Avenue premises, gave testimony on behalf of Owens. He opined:
... [T]he chromium [on the Hilfinger site] originated due to prolonged period of exposure, probably certain amount of wetting taking place over a period of time, possibly twenty years. This [sic] came to this conclusion because of the depth of penetration in the fairly high concentrations of chromium in the soil. The soil are [sic] decided low permeable soil, and we found chromium in elevated [sic] as deep as 18 feet.
... [0]ne of the actions undertaken at [the Hilfinger manufacturing facility] was plating of parts with chromium. It was my opinion that the plating operations, through some mechanism, are the origins of the chromium in the soil.
Id. at 990-91. When asked to give an opinion as to the time frame of contamination, he responded:
... I believe, again, it took place over a prolonged period of time, its time-frames are difficult to pin down because of the nature of the chromium migration within the soil profile. I believe, possibly, it could have taken place over a twenty year period. As to the exact time period of the specific events, I don’t, really, have an opinion to that effect.
Id. at 991. On direct examination, he concluded that “one single event probably did not provide this level of or degree of contamination within the soil.” Id. at 992. On cross-examination, he admitted that intentional spilling or dumping, or the pouring of vat solutions into a sewer could explain the contamination.
Owens also introduced the deposition testimony of Thomas Gene Naymik, a geology expert, who had reviewed materials relating to the chromium contamination at the Westwood Avenue premises. He admitted that he could not explain to a reasonable degree of certainty how the chromium got into the soil. He stated that “one event could have occurred, you know, 34 years ago [i.e., approximately 1951] or further into the past to give the distribution of chromium that we see today.” Id. at 513. He maintained that it is impossible to prove or disprove that chromium discharge occurred between 1963 and 1971.
Joint Exhibit 9 was also admitted into evidence. This exhibit consisted of several articles from the Toledo Blade which stated that, on at least two separate occasions in 1962, Hilfinger released large quantities of a chrome substance into the Ottawa River.
On the issue of the scope of Aetna’s insurance coverage, Owens produced Jack W. Galliers, an insurance agent for Hilfinger. Galliers testified that Aetna insured Hilfinger from 1957 to 19641 with, inter alia, comprehensive general liability (“CGL”) coverage. Based on “eheck[ing] prior policies for information, that were in what we called our archives,” id. at 936, Galliers suggested that Aetna’s coverage began before 1957, perhaps back to 1954. These policies were for three-year periods, and each had an aggregate limit of $100,-000 for property damage for the three-year period. Galliers’ testimony was based on his recollections and on a letter of coverage he wrote to ASARCO at the request of Hilfinger on August 21, 1963. The letter of coverage stated, in pertinent part:
We also provide Public Liability coverage in the Aetna Casualty and Surety Company with limits of $300,000 each person, $500,000 each accident Bodily Injury and $100,000 each accident Property Damage.
Id. at 63. On cross-examination, Galliers admitted that the conditions of Aetna’s coverage changed from time to time, but he could not say exactly what those changes were nor when they occurred. Neither could he say what the original conditions [869]*869were. The actual policies were not introduced.
Joint Exhibit 4, a CGL standard policy form generally issued by Aetna prior to 1955, was admitted into evidence. There was no stipulation and there was no testimony relating to how many years prior to 1955 this form was issued (though the form itself reads: “Second Revision — December 1, 1947”), or to whether the policy form was ever issued to Hilfinger.
An Aetna comprehensive liability insur-' anee standard policy form which was generally issued from approximately 1955 to 1966 was submitted as Joint Exhibit 5. Again, there is no indication from the record that this policy form was ever issued to Hilfinger by Aetna.
Both Joint Exhibit 4 and Joint Exhibit 5 contained the following language relating to property damage coverage:
[The Insurer agrees with the Insured, subject to the limits of liability, exclusions, conditions and other policy terms,] [t]o pay on behalf of the Insured all sums which the Insured shall become legally obligated to pay as damages because of injury to or destruction of property, including the loss of use thereof, caused by accident.
Id. at 194, 205. And Section IV of these Joint Exhibits reads:
This Policy applies only to accidents which occur during the policy period within the United States of America, its territories or possessions, [or] Canada [or Newfoundland].
Id. at 202, 207.
Joint Exhibit 6 consisted of “Binding Advices” which showed Hilfinger to have had property damage insurance with Aetna in the amount of $100,000 for each accident and $100,000 aggregate with regard to each of two three-year policies covering August 18, 1949 to August 18, 1955. Joint Exhibit 8, an “Exhibit of Reinsurance” prepared by Aetna, reflected these facts as well.
At the close of Owens’ case-in-chief, Aet-na moved for an involuntary dismissal pursuant to Rule 41(b) of the Federal Rules of Civil Procedure. The district court denied this motion.
Aetna then called Dr. Tola B. Moffet, a geologist and civil engineer, to the stand. At one point, the following colloquy took place:
THE COURT: Normally, the solids and rinse water, or plating, would that be able to permeate through concrete floor?
WITNESS: In the descriptions that I recall from depositions, or testimony, the floor, most likely, was composed of concrete. That concrete was, probably, cracked, or there may have been blocks, or wooden panels, in the floor. It is not clear, there really is insufficient evidence, to conclude it. My opinion is it was cracked and permeable, and over the span of the normal use of this facility, that chromium solution would have been spilled and would have leaked through the cracks in the, beneath the vats, basically.
J.A. at 1019-20.
Then, on the topic of the scope of insurance coverage, Mr. Robert E. Keenan, a general adjuster for Aetna, was asked, “Do any of the documents [referred to above, among others] that [counsel for Owens] has shown you, regarding Aetna insurance, indicate the terms of any policy of insurance?” Id. at 1054. He responded, “No they do not.” Id.
After the trial concluded, the parties filed post-trial briefs. Aetna again requested dismissal under Rule 41(b). This request was again denied.
On November 1, 1991, the district court ordered that judgment be entered in favor of Aetna. The court noted that “[n]one of the experts at the 1988 trial or the 1990 trial could testify with reasonable certainty that the hexavalent chromium was introduced by accidental means,” J.A. at 67 (Findings of Fact ¶ 22) (Nov. 1991), and stated that “[determining the cause of the loss cannot be left to mere conjecture or by proving the existence of hypothetical or problematical conditions which themselves cannot be proven and are not known to exist,” id. at 68 (Conclusions of Law If 7) (Nov. 1991). Given this, the court conclud[870]*870ed that “[Owens] has not demonstrated by a preponderance of the evidence that the chromium pollution found at the Hilfinger site was placed in the soil by accident during the period of coverage by Aetna.” Id. at 68 (Conclusions of Law If 10) (Nov. 1991).
The court determined that this period of coverage spanned from 1949 to 1964. Regarding the terms of this coverage, it held that the “proffered standard policy forms are evidence of general contents of the CGL policies issued by Aetna to Hilfinger,” id. at 67 (Conclusions of Law 114) (Nov. 1991), and that “the term ‘accident’ is to be construed to the ordinary understanding and common usage and speech of people generally” and “is narrower in scope than an occurrence,” id. at 68 (Conclusions of Law 11118-9) (Nov. 1991).
On November 27, 1991, Owens filed its Notice of Appeal. It contends that the district court erred when it concluded that Owens did not prove by a preponderance of the evidence that the chromium contamination on the Westwood Avenue property was caused by accident during the period of time that Aetna provided coverage for Hilfinger. Aetna cross-appealed on December 11, 1991. It argues that, en route to deciding in its favor, the district court made several specific rulings adverse to Aetna’s case. Should we agree with Owens' arguments that the district court erred in its judgment, Aetna requests that we consider the merits of its contentions on cross-appeal.
II
Rule 52(a) of the Federal Rules of Civil Procedure describes the deference to be afforded a district court’s findings of fact upon the conclusion of a bench trial: “Findings of fact, whether based on oral or documentary evidence, shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses.” The Supreme Court expounded on this, writing:
... If the district court’s account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently. Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous....
This is so even when the district court’s findings do not rest on credibility determinations, but are based instead on physical or documentary evidence or inferences from other facts....
When findings are based on determinations regarding the credibility of witnesses, Rule 52(a) demands even greater deference to the trial court’s findings....
Anderson v. City of Bessemer City, 470 U.S. 564, 573-75, 105 S.Ct. 1504, 1511-12, 84 L.Ed.2d 518 (1985). The Sixth Circuit has also commented:
[T]he factual conclusions rendered by a district court sitting without a jury are binding on appeal unless this Court is left with a definite and firm conviction that a mistake has been made. Thompson v. National Railroad Passenger Corp., 621 F.2d 814 (6th Cir.1980), cert. denied, 449 U.S. 1035, 101 S.Ct. 611, 66 L.Ed.2d 497. It is the appellant who must shoulder the burden of proving such a mistake, Godley v. Kentucky Resources Corp., 640 F.2d 831 (6th Cir.1981), and this burden is not met merely be demonstrating a conflict in the testimony, Walling v. General Industries Corp,, 330 U.S. 545, 67 S.Ct. 883, 91 L.Ed. 1088 (1947); Godley v. Kentucky Resources Corp., supra, nor by seeking to redetermine the credibility of witnesses. NLRB v. Nichols of Ohio, Inc., 472 F.2d 1228 (6th Cir.1972). Moreover, the appellate court must review the facts in the light most favorable to the present appellee. Drayton v. Jiffee Chemical Corp., 591 F.2d 352 (6th Cir.1978); United States v. Summit Fidelity & Surety Co., 408 F.2d 46 (6th Cir.1969).
Sawyer v. Arum, 690 F.2d 590, 591-92 (6th Cir.1982).
[871]*871Legal determinations made by the court below are not accorded the same deference, and are reviewed de novo.
Ill
Aetna argues that the district court’s judgment should be affirmed because, inter alia: (1) Owens did not prove by a preponderance of the evidence that the chromium got into the soil by “accident”; and (2) Owens did not prove by a preponderance of the evidence that the “continuous wetting” through which the chromium got into the soil occurred at any point during the time that Aetna insured Hilfinger.'
A
Regarding how the contamination occurred, the district court found only that the chromium “ ‘was introduced into the soil on the Property by a continuous or repeated wetting over a period of several years.’ ” J.A. at 65 (Findings of Fact ¶ 17) (Nov. 1991) (quoting Findings of Fact ¶ 15 (Jan. 1989)). Evaluating the testimony of the experts on this score, the district court found that “[n]one of the experts at the 1988 trial or the 1990 trial could testify with reasonable certainty that the hexava-lent chromium was introduced by accidental means.” Id. at 67 (Findings of Fact ¶ 22) (Nov. 1991).
Owens takes issue with the district court’s understanding of the term “accident.” Owens suggests that the district court proceeded to rule on the merits of the case with a misunderstanding of what “accident” means under the relevant Ohio law. Specifically, Owens insists, inter alia, that the court’s understanding of “accident” was limited to a happening identifiable in time and place, whereas a proper understanding of “accident” encompasses a continuous or repeated series of events. Given a proper understanding of “accident,” Owens contends, the district court would have found that Owens met its burden.of proving that, regardless of the specific cause, the chromium was introduced into the soil by accidental means.
We are not convinced that the district court limited its understanding of “accident” to exclude those happenings caused by a continuous or repeated series of events. We also find that the district court’s understanding pf “accident” was otherwise in accordance with the relevant Ohio law.
Though the district court concluded that “accident’’ “is narrower in scope than an occurrence,” id. at 68 (Conclusions of Law 119) (Nov. 1991), this does not necessarily misstate Ohio law. The state appellate court in Grand River Lime Co. v. Ohio Casualty Ins. Co., 32 Ohio App.2d 178, 289 N.E.2d 360, 365 (1972), wrote:
We adopt the argument ... that the word “occurrence” is much broader than the term “accident.” Such proposition is well.stated in Aerial Agricultural Service v. Till (N.D.Miss.1962), 207 F.Supp. 50, 57:
“To begin with, the word “occurrence,” to the lay mind, as well as to the judicial mind, has a meaning much broader than the word “accident.” As these words are generally understood, accident means something that must have come about or happened in a certain way, while occurrence means something that happened or came about in any way. Thus accident is a special type of occurrence, but occurrence goes beyond such special confines and, while including accident, it encompasses many other situations as well.
See also Buckeye Union Ins. Co. v. Liberty Solvents & Chemicals Co., 17 Ohio App.3d 127, 477 N.E.2d 1227, 1232-33 (Ohio Ct.App.1984). As Owens itself notes in its Brief, the fact that “accident” is understood to be narrower in scope than “occurrence” does not necessarily preclude a continuous or repeated series of events from falling within its definitional confines. Nor do we find that the district court reached an improper conclusion in this regard.
The district court determined that “accident” “is to be construed to the ordinary understanding and common usage and speech of people generally.” J.A. at 68 [872]*872(Conclusions of Law ¶ 8) (Nov. 1991). Under Ohio law, where, as in this case, a term in an insurance contract is not defined, it is to be given its ordinary meaning: “The law in this state is well-established with respect to the interpretation of insurance contracts. A court has an obligation to give plain language its ordinary meaning_” Miller v. Marrocco, 28 Ohio St.3d 438, 504 N.E.2d 67, 69 (1986); see also Tomlinson v. Skolnik, 44 Ohio St.3d 11, 540 N.E.2d 716, 717-18 (1989) (“Thus, in reviewing an insurance policy, words and phrases used therein ‘must be given their natural and commonly accepted meaning, where they in fact possess such meaning, to the end that a reasonable interpretation of the insurance contract consistent with the apparent object and plain intent of the parties may be determined.’ ”) (quoting Gomolka v. State Auto Mut. Ins. Co., 70 Ohio St.2d 166, 436 N.E.2d 1347, 1348 (1982)); Johnson v. Claycraft Co., No. 16-91-37, 1992 WL 180097, at *3, 1992 U.S.App. LEXIS 4120, at *9 (Ohio Ct.App. July 23, 1992) (“Additionally, words and phrases used in a policy of insurance are to be given their general, ordinary, usual and popular meaning.”). Thus, insofar as the district court held that the ordinary meaning of the term “accident” should be applied, it was not legally in error.
Ohio courts have consistently described the ordinary meaning of “accident” to refer to unintended and unexpected happenings. See Hybud Equip. Corp. v. Sphere Drake Ins. Co., 64 Ohio St.3d 657, 597 N.E.2d 1096, 1102 (1992) (“In its common, ordinary use, the word ‘accidental’ means unexpected, as well as unintended.”), cert. denied, — U.S. -, 113 S.Ct. 1585, 123 L.Ed.2d 152 (U.S. Mar. 23, 1993) (No. 92-1260); Munchick v. Fidelity & Casualty Co., 2 Ohio St.2d 303, 209 N.E.2d 167, 170 (1965) (“The term, ‘accidental,’ as used in an insurance policy means ‘an unexpected happening without intention or design.’ ”); Burns v. Employers’ Liability Assurance Corp., 134 Ohio St. 222, 16 N.E.2d 316, 317 (1938) (“Generally an accident is considered as an event proceeding from an unexpected happening or unknown cause without design, and not in the usual course of things.”); Sanborn Plastics Corp. v. St. Paul Fire & Marine Ins. Co., No. 91-G-1668, 1992 WL 398460, at *6, 1992 Ohio App. LEXIS 6585, at *14 (Ohio Ct.App. Dec. 31, 1992) (“Neither of these policies defined [‘accidental event’]. However, as it is commonly used, the word ‘accidental’ typically refers to something which was not expected or intended.”).
Having concluded that the district court did not err in its understanding of “accident,” and having reviewed what Ohio courts have understood the ordinary meaning of “accident” to be, we proceed to determine whether the district court erred in its evaluation of the evidence in light of the applicable legal standards. Owens contends that the district court erred, as a legal matter, by making Owens’ burden too great. The district court asserted that “determining the cause of the loss cannot be left to mere conjecture or by proving the existence of hypothetical or problematical conditions which themselves cannot be proven and are not known to exist.” J.A. at 68 (Conclusions of Law 117) (Nov. 1991). Owens argues that this statement reflects the fact that the district court implicitly required Owens to prove the actual or specific cause of the contamination, as opposed to proving merely the probable cause. It cites Inland Rivers Serv. Corp. v. Hartford Fire Ins. Co., 66 Ohio St.2d 32, 418 N.E.2d 1381 (1981) to support the proposition that mere probability will suffice.
In Inland Rivers, a barge sank for causes unknown. The owner of the barge sought compensation from its insurer, claiming that, though it could not specify the cause of the sinking, the cause was probably a “peril of the sea.” The insurer argued that the barge sank as a result of bad bottomry and neglect, and denied coverage. The trial court found for the insured. The appellate court held that the insurer was entitled to judgment in its favor due to the insured's failure “to present evidence of probative force sufficient to prove by a preponderance that the loss was caused by a peril or risk covered by the policy.” 418 N.E.2d at 1382. The appellate court concluded that the insured’s evi[873]*873dence “utterly failed to prove what or who caused the puncture in the hull, the most likely point of entry of the water.” Id.
The Ohio Supreme Court reviewed the matter and reversed. It held that, when a ship sinks without apparent cause, a presumption of unseaworthiness arises. This presumption may be rebutted by the insured by producing evidence that the probable cause of the sinking was external to the vessel. The court wrote:
The record reveals that [the insured] offered expert evidence that a hole found in the side plate of the barge, below the water line, had been “punched in” by an instrumentality external to the vessel, and that the hole was the cause of the sinking. We find this evidence to be sufficient to support the trial court’s finding that [the insured] overcame the presumption of unseaworthiness and carried its burden of persuading the finder of fact that the loss resulted from a covered “peril of the sea.”
Id. at 1383.
Inland Rivers supports the proposition that the rebuttable presumption of unseaworthiness may be overcome by evidence that the probable cause of a sinking was a peril of the sea. It also supports the proposition that an insured may carry its burden of persuasion by evidence that the probable cause was a peril of the sea; it is not required to prove, by a preponderance of the evidence, the actual cause. Inland Rivers did not relieve the insured of its continual burden of proof by a preponderance of the evidence, as pointed out in the dissenting opinion in that case:
It is a universal principle of insurance and maritime law that the insured has the burden of proving, by a preponderance of the evidence, that the damage to the vessel was caused by an insured risk, i.e., a peril of the sea.
While the presumptions ... may shift the burden of producing evidence back and forth between the parties ... the burden of proof, also known as the burden of persuasion, remains with the insured throughout the trial. The burden of proof does not shift to the insurer to prove that the loss was caused by something other than a peril of the sea.
Id. at 1384 (Celebrezze, C.J., dissenting) (citations omitted). In other words, evidence that the probable cause of a sinking was a peril of the sea, while rebutting the presumption of unseaworthiness, may or may not constitute sufficient evidence to further persuade the factfinder, by a preponderance of the evidence, that the cause of the sinking was some peril of the sea— whatever the specific cause may be. In Inland Rivers, the majority of the court decided that the rebuttal evidence also constituted sufficient evidence to carry the insured’s burden of persuasion. The dissent did not, writing:
It seems to me that if the finder of fact could not conclude what caused the hole or the sinking, then the insured, the party with the burden of proof, did not succeed in demonstrating, by a preponderance of the reliable, probative and credible evidence, the probability, or even the possibility, that such sinking was caused by a peril of the sea, a covered risk....
My reading of the trial judge’s comments illustrates that he not only believed that the barge owner had not demonstrated that the sinking was caused by a peril of the sea, but also, that, in his mind, it was equally plausible that the sinking was caused by a leak within the barge itself, by bad bottomry or by some other non-covered risk.
Thus, construing the trial judge’s comments about the conflicting expert testimony and all the evidence in a light most favorable to the barge owner, we have, at most, a case that is “in equipoise.” Much like in baseball where “a tie goes to the runner,” an insurance case in equilibrium “goes to the insurer” when the insured fails to meet its burden of proof.
I am fully cognizant of the fact that the insured, in cases like this, does not have to prove the exact cause of the sinking or offer into evidence the precise submerged object that might have caused the sinking. I dissent on the basis that the finder of fact and a unani[874]*874mous Court of Appeals, though rendering different judgments, concluded that the evidence did not preponderate in favor of attributing the sinking to a covered risk. Since the trial judge was in a better position than we are to evaluate the conflicting evidence and credibility of the witnesses, I submit that it is more appropriate to defer to his analysis of the competing inferences.
The record in this case is voluminous. The pleadings, transcript, exhibits and appendices occupy the better part of a file cabinet. After a long trial and a fully-developed appeal, the finder of fact and a unanimous Court of Appeals concluded, in essence, that the barge sank in some unknown manner, by some unknown object, at some unknown time. We sit as an appellate court with a limited scope of review; we do not sit as a roving bank of claims adjusters with Ouija boards. How this court can, eight years after the sinking of this barge, substitute its judgment for the finder of fact and conclude that the sinking was, indeed, caused by a covered risk, remains, to me, an existential mystery.
Id. at 1384-85 (Celebrezze, C.J., dissenting) (citations omitted).
In the instant case, we conclude that the district court did not necessarily require, as a matter of law, that Owens prove the actual or specific cause of the contamination by a preponderance of the evidence. We view the district court’s statement on the insufficiency of conjecture and hypothesis to reflect the court’s view of the causation evidence. In attempting to meet its “burden of determining by a preponderance of the evidence that an accident occurred within the property damage liability policy,” J.A. at 67 (Conclusions of Law ¶ 6) (Nov. 1991), according to the district court, Owens did not present more than conjecture and hypothesis, see id. at 67 (Findings of Fact ¶ 22) (Nov. 1991) (“None of the experts at the 1988 or the 1990 trial could testify with reasonable certainty that the hexavalent chromium was introduced by accidental means.”), which the court found to be insufficient: “[Owens] has not demonstrated by a preponderance of the evidence that the chromium pollution found at the Hil-finger site was placed in the soil by accident during the period of coverage by Aet-na.” Id. at 68 (Conclusions of Law II10) (Nov. 1991).
A review of the evidence supports the district court’s assessment. Opinions on causation ran the gamut of potential causes. Owens’ own counsel suggested that the issue of causation would be left chiefly to conjecture and hypothesis: “There will be no evidence as to how the chromium got into the ground. We don’t know.” Id. at 927. The range of potential causes did not exclude expected/intended ones. For example, Owens’ own expert witness, Mr. Mack, admitted that the chromium contamination could be explained by intentional spilling or dumping. Id. at 994. Mr. Hilfinger’s self-serving declaration that it was not Hilfinger’s intention, nor did the company expect to contaminate its premises could be properly discounted by the district court, which is entitled to special deference in judging the credibility of witnesses. Owens argues that “there is no evidence to show that Hilfinger intended to cause harm to the property or that it foresaw that its process for plating automobile parts would result in damage to the property.” Owens’ Br. at 27. But Aetna did not have the burden of proving intentional/expected causation; Owens had the burden of proving probable unintentional/unexpected causation to bring the chromium contamination within the alleged terms of the alleged insurance policy. See Inland Rivers, 418 N.E.2d at 1383 (“It is undisputed that one seeking to recover on an insurance policy has the burden of proving a loss and demonstrating coverage under the policy.”). In the end, the district court determined that Owens did not meet its burden. We agree.
B
Regarding when the contamination occurred, the district court found only that the chromium was introduced into the soil “ ‘[djuring the time Hilfinger Corporation was an owner and operator of the Proper[875]*875ty.’ ” Id. at 67 (Conclusions of Law 11 5) (Nov. 1991) (quoting Conclusions of Law ¶ 5 (Jan. 1989)). Aetna insured Hilfinger from 1949 to 1964. Plant operations which might have caused the chromium contamination on the plant premises included time periods both before and after the time period of Aetna’s coverage, from 1945 to 1949 and from 1964 to 1970. See id. at 955 (testimony of Mr. Hilfinger to the effect that, at no time from 1945 to 1970 did the Hilfinger plant not conduct chrome plating operations). Experts testified that contamination might well have occurred before or after the policy period in question. Though one expert, Mr. Mack, testified that the time period of contamination might necessarily have included the policy period, he specifically qualified his opinion: “I believe, possibly, [the contamination] could have taken place over a twenty year period. As to the exact time period of the specific events, I don’t, really, have an opinion to that effect.” Id. at 991. All told, it was not erroneous for the district court to find that Owens did not meet its burden of showing, by a preponderance of the evidence, that the contamination occurred during the policy period. Since, according to the alleged policies, a covered accident had to occur within the policy period, this finding was enough to warrant judgment for Aetna. Cf. Fireman’s Fund Ins. Cos. v. Ex-Cell-O Corp., 750 F.Supp. 1340, 1349 (E.D.Mich.1990) (“Because it is just as likely that the contamination occurred after the expiration of the last policy in 1979 or that the contamination resulted from the day-to-day operations of the ... plant, I find that policyholders have not carried their burden of proving an occurrence.”).
IV
Aetna has filed a cross-appeal presenting alternative grounds for affirming the judgment below. In light of our disposition of Owens’ appeal, we need not consider Aetna’s cross-appeal. Cf. Brunet v. City of Columbus, Nos. 86-3557, 86-3603, 1987 WL 44518, at *3, 1987 U.S.App. LEXIS 11363, at *7-*8 (6th Cir. Aug. 25, 1987) (“appellate courts review judgments, not statements in opinions”), cert. denied, 485 U.S. 1034, 108 S.Ct. 1593, 99 L.Ed.2d 908 (1988).
The judgment below is AFFIRMED. Aetna’s cross-appeal is DISMISSED.