American Home Assurance Company (American Home) prevailed in a bench trial against American President Lines (APL) on its claim that cargo carried by APL arrived at its destination freeze damaged. In APL’s appeal, we consider the amount of proof needed to negate a prima facie case under the Carriage of Goods by Sea Act, 46 U.S.C. § 1300 et seq., (COGSA), the requirements of proving a COGSA defense, and the evidence required to prove damages under COGSA.

I.

APL is a common carrier by sea. In 1990 and 1991, APL delivered four shipments of fruit (cherries and grapes) from the United States to Hong Kong for Primary Export International, Inc. (Primary). The parties stipulated that each shipment was in good condition when received by APL. Primary’s customer, Ever Fresh Trading Company (Ever Fresh), claimed that each shipment was partially freeze damaged on arrival in Hong Kong. Primary’s insurer, American Home, paid Primary for the damage and brought the present action against APL for subrogation.

APL used refrigerated ocean containers for each of the four shipments. Included within each container was a refrigeration unit, the purpose of which was to cool the container and maintain the cargo at the desired temperature. The refrigeration units employed devices that electronically recorded the temperature of the air entering and leaving the refrigeration unit. APL presented evidence that it inspected the refrigeration units both before and after the voyages to ensure that they were in good working order. In addition, some of the containers also contained one or more “Ryan” recorders. A Ryan recorder is a small portable recorder that creates a graphical record of temperatures in the container.

When the shipments were delivered to the Hong Kong fruit market, surveyors arrived to examine and sample the cargo. APL’s surveyor was Mr. Cheung. Mr. Iu was Ever Fresh’s surveyor. During their joint inspection of representative samples of the various cargos, the surveyors prepared descriptions of their findings, which were consistent with freeze damage. They also took color photographs and attached them to the surveys. Experts at trial gave conflicting testimony about whether the photographs demonstrated freeze damage. After determining the proportion and severity of the damage, the surveyors agreed to depreciation allowances for each shipment.

APL attempted to discredit these reports by impheating Messrs. Cheung and Iu in a conspiracy to falsify the surveys. APL also presented extensive evidence of the highest freezing temperatures of the various fruit and the temperature readings recorded during the voyage, all of which were comfortably above the highest freezing levels. However, American Home disputed how complete the records were, and presented evidence at trial that in three out of four shipments, APL’s temperature information was downloaded before the containers were delivered.

Ultimately, the district court concluded that there was “no credible evidence to substantiate fraud, collusion, conspiracy or mistake.” The district court found that the fruit had been delivered freeze damaged, that APL was liable, and that APL was unable to avoid liability under any of the COGSA exceptions. The district court awarded damages based on the depreciation allowances in the survey reports.

II.

To recover under COGSA, American Home bore the burden of establishing that the fruit was damaged in transit. See Complaint of Damodar Bulk Carriers, Ltd., 903 F.2d 675, 683 (9th Cir.1990). To establish a prima facie case, American Home had to prove that the fruit was delivered to APL in good condition and that APL delivered the fruit in a damaged condition. See id.; Quaker Oats Co. v. M/V Torvanger, 734 F.2d 238, 240 (5th Cir.1984), cert. denied, 469 U.S. 1189 , 105 S.Ct. 959 , 83 L.Ed.2d 965 (1985).

The parties stipulated that the cargos were delivered to APL undamaged. Thus, to meet its initial burden under COGSA, American Home needed only to demonstrate that APL delivered freeze-damaged fruit. The district court found that American Home met this burden. The court rested its conclusion on: (1) a factual finding that the surveyors’ reports, photographic evidence, and expert testimony proved the existence of freeze damage; and, (2) a legal conclusion that Mr. Cheung’s survey report was legally binding on APL.

A.

With respect to the district court’s factual findings, we review for clear error. See Havens v. F/T Polar Mist, 996 F.2d 215, 217 (9th Cir.1993). Under this standard, we will not disturb the district court’s findings “[i]f the district court’s account of the evidence is plausible in light of the record viewed in its entirety-” Anderson v. Bessemer City, 470 U.S. 564, 573-74 , 105 S.Ct. 1504, 1511 , 84 L.Ed.2d 518 (1985). “Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.” Id. at 574 , 105 S.Ct. at 1511 .

Applying this standard, we find that the district court’s conclusion was not clearly erroneous. For each shipment, American Home presented evidence of the initial inspections done by surveyors when the fruit arrived in Hong Kong. When each of the four shipments arrived, Mr. Cheung surveyed containers on behalf of APL and wrote descriptions in his reports. The various shipments were described as covered with a film of frost, harder than normal, and softened at ambient temperature. A number of the fruit were cut into halves for inspection and found to be semi-transparent. Additionally, American Home presented evidence of the pulp temperatures of the various cargos and photographs that Mr. Cheung took and attached to his surveys. American Home’s expert testified that such conditions were consistent with freeze damage. Mr. Cheung’s conclusions were corroborated by Mr. Iu, the surveyor acting on behalf of Primary’s customer, Ever Fresh.

By contrast, APL introduced no direct evidence that the shipments were not freeze damaged. Instead, APL attempted to challenge the credibility of the American Home’s witnesses and offered circumstantial evidence in the form of temperature readings, which it argued showed that the fruit could not have frozen.

Based on this evidence, the district court concluded that the fruit arrived in Hong Kong freeze damaged. Given the direct evidence of freeze damage presented by American Home, we cannot say this conclusion was implausible. See Anderson, 470 U.S. at 573-74 , 105 S.Ct. at 1511 . Since there were two possible interpretations of the evidence, the court’s conclusion that American Home established a prima facie case was not clearly erroneous. Id.

B.

Appellant also contests the district court’s conclusion that, absent proof “of fraud, collusion, conspiracy to defraud or mistake ... [APL] must be bound by the report of its own hired representative.” We review the district court’s conclusion that APL was legally bound by its surveyor’s report de novo. See Havens, 996 F.2d at 217 .

Although this Circuit has not specifically addressed the question of surveys as admissions in the admiralty context, the issue has been presented in several other forums. We are persuaded by the reasoning of other courts that have addressed this issue.

When “a survey report .■.. bears the notation Vithout prejudice,’ it is not binding on the [carrier], nor does it estop [it] from contesting the amount and/or damages.” Cameron Offshore Servs., Inc. v. M/V PBR-108, 1980 A.M.C. 527 , 529 (W.D.La.1979) (citing Berns & Koppstein, Inc. v. Orion Ins. Co., 170 F.Supp. 707 (S.D.N.Y.1959) and Establissements Edouard Materne v. S.S. Leerdam, 143 F.Supp. 367, 370 (S.D.N.Y.1956)). Since Mr. Cheung’s survey report bears the notation “without prejudice,” the district court erred by ruling that the report was binding on APL.

We will not reverse, however, if the error was harmless. See Fed.R.Civ.P. 61; McDonough Power Equip., Inc. v. Greenwood, Inc., 464 U.S. 548, 553, 104 . S.Ct. 845, 848, 78 L.Ed.2d 663 (1984). While' the surveyor’s report was not legally binding on APL, the district court could have allowed its introduction as an admission. See Brown & Root v. American Home Assurance Co., 353 F.2d 113, 116 (5th Cir.1965), cert. denied, 384 U.S. 943 , 86 S.Ct. 1465 , 16 L.Ed.2d 541 (1966); Compagnie De Navigation, Etc. v. Mondial United Corp., 316 F.2d 163 (5th Cir.1963); Cox v. Esso Shipping Co., 247 F.2d 629, 632-34 (5th Cir.1957); Cameron, 1980 A.M.C. at 529. If the district court had admitted the report as an admission, it would have evaluated the admission’s reliability in light of APL’s explanation or repudiation. Brown & Root, 353 F.2d at 116 .

During the proceedings, the district court actually allowed APL to explain the surveys. APL introduced temperature records, from which it attempted to raise the inference that the fruit could not have frozen. APL also attempted to prove that the surveys were part of conspiracy to defraud. The district court was ultimately unpersuaded by those explanations. The court found that “having weighed the evidence on both sides ... the testimony and reports of the ‘eye witness’ surveyors must be given more credibility [sic] than that of the temperature monitors[,]” and that “Mr. Cheung’s testimony [was] entitled to greater weight [than the temperature readings].” Thus, for practical purposes, the court treated the survey as an admission. Nothing indicates that a different conclusion would have resulted if the district court had not erred in finding that APL was also legally bound by Mr. Cheung’s survey. We therefore find that the district court’s error was harmless.

III.

APL also maintains that the district court abused its discretion by not allowing APL to introduce evidence of Mr. Iu’s twelve-year-prior criminal conviction in Hong Kong for conspiracy to defraud shipping companies by falsifying surveys in exchange for bribes. The conviction should have been admitted, APL contends, either to impeach Mr. Iu under Federal Rule of Evidence (Rule) 609(b), or as substantive evidence to prove Mr. Iu’s motive or intent under Rule 404(b). APL had the burden of proving the admissibility of Mr. Iu’s conviction. See United States v. Arambula-Ruiz, 987 F.2d 599, 602-03 (9th Cir.1993). We review the district court’s decision not to admit evidence for abuse of discretion. Campbell v. Wood, 18 F.3d 662, 685 (9th Cir.), cert. denied, — U.S. -, 114 S.Ct. 2125 , 128 L.Ed.2d 682 (1994).

A.

Even though the conviction was over ten years old, the district court could have allowed its admission under Rule 609(b) if it determined that, in the interest of justice, the conviction’s probative value substantially outweighed its prejudicial effect. Generally, a conviction of fraud is probative of witness’s credibility. Cf. United States v. Ortega, 561 F.2d 803, 806 (9th Cir.1977) (explaining in the related context of Rule 609(a) that prior convictions involving dishonesty are peculiarly probative of credibility). The question remains: How probative of Mr. Iu’s credibility was this conviction? APL argues that, in determining the probative value of his conviction, we should consider how critical Mr. Iu’s credibility was to their case. This argument misconstrues the yardstick by which probative value is measured. Probative value is determined by how likely the evidence is to prove some fact, not how important proof of that fact is to the proponent’s case. Accordingly, the probative value of Mr. Iu’s conviction is measured by how well it demonstrates his lack of trustworthiness, not how badly APL wants to impeach him. The district court doubted that the conviction would be very probative of Mr. Iu’s credibility, both because of its age and because of Mr. Iu’s claims that the guilty plea was coerced. Consequently, we find that the district court did not abuse its discretion by excluding the conviction.

B.

Alternatively, APL argues that Mr. Iu’s conviction should have been admitted under Rule 404(b) to prove his intent or motive to falsify his surveys. This Court has established the following four-part test to evaluate admissibility of evidence under Rule 404(b):

U.S. v. Bradley, 5 F.3d 1317, 1319-20 (9th Cir.1993). Although the district court did not specifically apply this test, we may affirm the district court’s decision on any grounds supported by the record. See Scott v. Kuhlmann, 746 F.2d 1377, 1378 (9th Cir.1984).

With regard to the first factor', the district court excluded the conviction in part based on Mr. Iu’s claims that his plea was coerced. The district court was uncertain whether the conviction was proper and recognized that inquiry into whether the conviction was proper would require the court to explore the Hong Kong judicial system. The district court’s misgivings demonstrate that it did not believe there was sufficient evidence to determine that Mr. Iu committed the prior offense.

The second factor also weighs against admissibility. Neither motive nor intent is an element of American Home’s claims or APL’s defenses. Rather, APL sought to introduce evidence of Mr. Iu’s motive and intent to refute American Home’s survey evidence. Thus, motive and intent were offered as rebuttal and were not material issues in the case.

The third requirement, that the conviction not be too remote, also weighs against admissibility; Mr. Iu’s conviction is over twelve years old.

The fourth element favors admission because the conviction was substantially similar to the actions APL is attempting to prove here: falsification of a survey. Given that three of the four factors governing weigh against admissibility, the district court was within its discretion in deciding to exclude the conviction.

IV.

APL also contends that if the shipments were freeze damaged during the voyage, APL was entitled to relief under one of the COGSA exceptions. After American Home proved that the fruit was delivered freeze damaged, the burden under COGSA shifted to APL to prove that it was eligible for one of the COGSA exceptions. Complaint of Damodar, 903 F.2d at 683 . To avail itself of one of these defenses, APL had to prove either that injury was from an excepted cause, or that the carrier exercised due diligence to make the ship seaworthy and properly stowed and eared for the cargo. See Michael F. Sturley, 2A Benedict on Admiralty § 56 (Aileen Jenner & Joan Dorfman eds., 1994); Complaint of Damodar, 903 F.2d at 686 (explaining that if unseaworthiness did not cause the damage, carrier had burden to show loss was caused by one of the exemptions before burden shifts back to shipper to prove carrier was at fault); Quaker Oats, 734 at 240-41; Sunkist Growers, Inc. v. Adelaide Shipping Lines, 603 F.2d 1327, 1341 (9th Cir.1979) (holding that carrier cannot invoke fire exemption if it failed to carry burden of showing due diligence to make the ship seaworthy and unseaworthiness caused the damage); States Marine Corp. of Delaware v. Producers Coop. Packing Co., 310 F.2d 206 , 212 (9th Cir.1962) (explaining that if carrier failed to show damage was from an excepted cause, it had the further burden of showing by a preponderance of the evidence that damage was not due to failure to stow or care for cargo properly), cert. denied, 444 U.S. 1012 , 100 S.Ct. 659 , 62 L.Ed.2d 640 (1980).

APL bears the burden of explaining what caused the damage because APL was in control of the cargo and therefore had exclusive knowledge of what happened to it. See Schnell v. The Vallescura, 293 U.S. 296, 304 , 55 S.Ct. 194, 196 , 79 L.Ed. 373 (1934). APL did not meet its burden under any of the COGSA exceptions because it offered no evidence of what caused the freeze damage.

That APL inspected the containers and found no defects in the refrigeration equipment only proves that APL acted diligently to avoid damage which may have been caused by a defect in the machinery. In fact, there are several possible causes of freeze damage, including temperature fluctuations that occur routinely during defrosting cycles, temperature fluctuations during vessel loading or unloading operations when refrigeration units are shut off for safety reasons, excess humidity, uncontrollable temperature changes within the refrigeration unit caused by changing weather conditions, or improper location of the refrigerating unit within the shipping vessel. See Pueblo Int’l, Inc. v. Puerto Rico Marine Management, 1989 A.M.C. 466 , 468-69 (D.P.R.1988). We cannot know whether APL exercised due diligence to prevent this particular loss if the cause of the loss has not been identified.

To relieve itself of liability, APL had the burden “to explain the unexplained or unexplainable loss.” Cf. Quaker Oats, 734 F.2d at 243 (describing the burden in the context of § 1304(2)(q)). APL made no such showing. Since APL did not satisfy that burden, it could not successfully invoke any of the COGSA exceptions.

A.

APL could have been exonerated under COGSA if it had proven the freeze damage resulted from an unseaworthy condition that was not caused by its own lack of due diligence. See COGSA § 1303(1) & § 1304(1). Under the unseaworthy condition exception, APL bore the burden of proving the exercise of due diligence. See Sunkist Growers, 603 F.2d at 1335-36 .

APL offered no direct proof of what unsea-worthy condition caused the freeze damage. Instead, APL sought to rely on a logical inference that the only possible cause of the freeze damage was malfunctioning of refrigeration equipment. This inference was not sufficient to establish that the damage was the result of an unseaworthy condition. Moreover, without knowing what the unsea-worthy condition was, we cannot evaluate whether APL’s evidence of due diligence was sufficient to prevent that condition. For example, if placing the containers in a different part of the ship would have left them less susceptible to weather-related temperature fluctuations, the cause would not have been an unseaworthy condition. See Sunkist Growers, 603 F.2d at 1339 . In that instance, the fact that APL inspected the refrigeration equipment would have been irrelevant. Given APL’s failure to present credible evidence about the actual cause, it is not entitled to exoneration under the unseaworthiness exception.

B.

APL’s latent defect claim fails for a similar reason. COGSA § 1304(2)(p) absolves a carrier from loss or damage arising from latent defects that were not discoverable by due diligence. Again, APL argued by inference that the only possible cause of the freeze damage was malfunctioning of the refrigeration equipment. Since it inspected all the equipment and found no defect, APL speculates that the defect must have been latent. However, APL did not prove precisely what caused the loss. Since the loss could have resulted from any of several causes, only one of which was a latent defect, APL did not meet its burden of showing that the damage was caused by a latent defect which could not have been discovered through due diligence.

C.

Likewise, APL failed to meet its burden of proof to avail itself of the COGSA § 1304(2)(q) exception. Section 1304(2)(q) is a catch-all clause that exonerates carriers from damage that occurs without fault by the carrier. To escape liability under this exception, APL had to prove first what actually caused the freeze damage and, second, that it was free from any fault whatsoever in contributing to the damage. See Quaker Oats, 734 F.2d at 242-43 ; EAC Timberlane v. Pisces, Ltd., 745 F.2d 715, 720 (1st Cir.1984). APL could not satisfy this burden simply by showing that the crew did all it was supposed to do. Section 1304(2)(q) specifically provides that “the burden of proof shall be on the person claiming the benefit of this exception to show that neither the actual fault or privity of the carrier nor the fault or neglect of the agents or servants of the carrier contributed to the loss or damage.” 46 U.S.C. § 1304 (2)(q). In other words, “the carrier must explain what took place or suffer the consequences.” Quaker Oats, 734 F.2d at 243 (quotation omitted). Since APL did not produce any direct evidence or compelling circumstantial evidence of what caused the loss, APL could not escape from liability under § 1304(2)(q).

Y.

APL’s final contention is that American Home’s damage claims are too speculative to allow recovery. The proper measure of damages under COGSA is the difference between the fair market value of the cargo in sound condition and the fair market value of the cargo in its damaged state. Daido Line v. Thomas P. Gonzalez, Corp., 299 F.2d 669 (9th Cir.1962). APL does not dispute that the value of the fruit undamaged was the invoice price. Therefore, the only issue we address is whether the depreciation allowances developed in the surveys constituted sufficient proof of the fair market value of the cargo in its damaged condition.

APL maintains that the only proper proof of the value of the damaged fruit is evidence of the actual price obtained in the Hong Kong market or of destruction certificates. In support of this argument, APL cites Weirton Steel Co. v. Isbrandtsen-Moller Co., 126 F.2d 593 (2d Cir.1942). APL urges that Weirton stands for the proposition that nothing short of evidence of the price actually realized in a resale can properly form the basis for a damages calculation.

In accord with the Second Circuit’s reasoning in Thyssen, Inc. v. S.S. Fortune Star, we reject such an expansive reading of Weirton. See Thyssen, Inc. v. S.S. Fortune Star, 777 F.2d 57 (2d Cir.1985). The Weirton court rejected surveyor’s reports of the resale value of the cargo because the plaintiff had reconditioned the goods at a small expense. The surveyor’s testimony only concerned the price the cargo would have brought without reconditioning. In order to avoid a windfall to the plaintiff, the court demanded evidence that the product manufactured from the reconditioned cargo sold for less than goods manufactured from undamaged cargo. See Thyssen, 777 F.2d at 62 .

By contrast, the fruit in this case was not reconditioned. Furthermore, APL introduced no direct evidence that the depreciation allowances were inaccurate. At trial, APL elicited testimony that the samples upon which the surveyors relied represented less than ten percent of the fruit. From this fact, APL asked the district court to infer that the surveyors’ reports were necessarily inaccurate. The district court declined APL’s invitation. In the absence of contrary evidence, the survey report’s depreciation allowances were adequate to show actual dam age. See Van Der Salm Bulb Farms, Inc. v. Hapag Lloyd, 818 F.2d 699 , 700 (9th Cir.1987). Given the dearth of contrary evidence, we conclude that the district court’s finding that the surveys were a reliable estimation of the depreciation value of the cargo was not clearly erroneous. See id.

The judgment of the district court is AFFIRMED.