OPINION
SIMON, District Judge:
After a flood occurred in the warehouse of a business that purchased and resold electronic parts, a dispute arose between the business and its insurer. The insured sued, alleging express breach of contract and breach of the implied covenant of good faith. The insurer moved for summary judgment. Without holding a Dau-bert hearing,1 the district court excluded the insured’s expert witnesses and granted summary judgment to the insurer, finding insufficient evidence that the flood caused damage to the insured’s inventory. [811]*811Because the district court abused its discretion by not allowing a jury to resolve contested but otherwise admissible expert testimony, we reverse and remand for trial.
FACTS
Pyramid Technologies, Inc. (“Pyramid”) purchased an insurance policy (the “Policy”) from Hartford Casualty Insurance Company (“Hartford”). The Policy provides coverage limits of $1 million for building replacement costs, $5.5 million for business personal property replacement costs, and $3 million for lost business income and additional expenses due to the inteiTuption of business operations. To trigger coverage for building or business personal property replacement costs, the Policy requires damage to property or its direct physical loss.
Pyramid purchased and resold electronic parts, many of which were out-of-date or not state-of-the-art. It did not test the inventory unless required to do so by a customer or prospective customer. Pyramid stored its inventory on shelves in a warehouse that did not have air conditioning or humidity control. Pyramid had approximately 52 million items in its warehouse at the time of the flood.
In the morning hours of August 11, 2005, Pyramid employees arrived at work to find the warehouse and certain other building locations flooded with one to two inches of water. Although the flood water did not reach the shelves on which inventory items were located, several employees saw visible condensation on packages in the lower three to four shelves. ServPro, a professional cleanup company, performed cleanup operations from August 11 through August 16, 2005.
After discovering the flood, Pyramid was concerned about the humidity level in the warehouse and the condensation found on its packages. Pyramid asked Hartford to test the inventory. Hartford’s expert, Peter Helms from Belfor USA Technical Services, visited the site after cleanup and, relying on humidity tests conducted after most of the water had been removed and drying equipment had been in place for more than 24 hours, determined that the humidity did not reach a level that could have caused damage to any of the inventory. Hartford refused to test the inventory, which would have cost more than $13 million to test every item. Hartford based its decision largely on Helms’ conclusion that the inventory was not damaged by the flood.
While Hartford was visiting the site after the flood, a potential Pyramid customer, WMS Gaming, Inc. (“WMS”), was conducting a quality control site visit before approving Pyramid as a parts supplier. The Hartford representative told this potential customer that the water intrusion was “no big deal” and that Hartford would not test the parts. The customer replied that the flood was a “big deal” to WMS. Shortly thereafter, WMS declined to approve Pyramid as a parts supplier.
One month after the flood, Pyramid hired Allied Public Adjusters, Inc. to assist in pursuing an insurance claim. Pyramid also hired its own expert, David Spiegel, to determine what the humidity levels were at the time of the water intrusion. Spiegel opined that the humidity level in the warehouse rose to more than 90% and that the conditions caused by the flood exceeded the protection levels of the moisture-proof packaging. During routine inventory checks after the flood, Pyramid employees quarantined more than 250,000 items, looking for visible signs of corrosion, tarnish, or discoloration. In August 2007, Hartford finally agreed to conduct limited test[812]*812ing of a small subset of parts identified by Pyramid as being damaged.
Hartford retained Dr. Arum Kumar of SEAL Laboratories to conduct tests on 374 items out of Pyramid’s inventory. These parts were selected by Pyramid as exhibiting signs of water damage. Dr. Kumar determined that 147 of those items exhibited corrosion, tarnish, or discoloration. Dr. Kumar conducted additional tests on those 147 items. He found that two parts failed the additional testing, and they were deemed unsuitable for commercial applications. Dr. Kumar stated that corrosion, tarnish, and discoloration are always caused by moisture, but he concluded that the August 11, 2005 flood was not the cause of the corrosion damage to the parts he examined.
Pyramid hired two additional experts, Del Mortenson and Ken Pytlewski, to evaluate the validity of Dr. Kumar’s report. Mortenson questioned Dr. Kumar’s opinion on the grounds that Dr. Kumar used “military” standards of suitability instead of “commercial” standards. Pytlewski challenged other opinions of Dr. Kumar’s, including his opinions that any corrosion caused by the flood would necessarily have been uniform and that visible corrosion is not a failure criteria under military standards. Pytlewski also noted the internal inconsistency in Dr. Kumar’s report between his statement that the cause of the moisture-related corrosion cannot be determined and his conclusion that the flood was not the cause of any moisture-related damage. Pytlewski testified at a deposition that in his opinion, some of the corrosion to Pyramid’s inventory occurred as a result of the high humidity caused by the August 11, 2005 flood.
By May 2010, approximately 17 million of Pyramid’s 52 million parts in inventory at the time of the flood had been sold, and approximately 35 million parts remained in Pyramid’s inventory. In October 2010, Pyramid sold most of its remaining inventory at a distress sale price of $125,000.
PROCEDURAL BACKGROUND
Pyramid filed this civil action in California state court. Hartford removed the lawsuit to federal court. On March 21, 2011, Hartford moved for summary judgment. In opposition to Hartford’s motion, Pyramid offered the expert reports of Spiegel, Mortenson, and Pytlewski, among other evidence. In reply to Pyramid’s opposition, Hartford argued that the expert reports of Spiegel, Mortenson, and Pytlew-ski, and much of the testimony of Tony Mavusi, the president of Pyramid, was inadmissible. The district court did not hold a Daubert hearing.
The district court also did not hold oral argument on Hartford’s motion for summary judgment. Instead, on June 1, 2011, the district court granted summary judgment in favor of Hartford, sustained many of Hartford’s objections to the testimony of Mavusi, and excluded the expert reports of Spiegel, Mortenson, and Pytlewski. The district court excluded the reports of Mortenson and Pytlewski as being “illegible,” and the court excluded the Spiegel report on the grounds that Spiegel was not a qualified expert and that his report was not based on sufficient facts or data and was not the product of reliable principles and methods.
Pyramid moved for reconsideration and submitted enlarged and more legible versions of the Mortenson and Pytlewski reports. The district court accepted the enlarged reports as sufficiently readable, but then excluded them as unreliable and not based on sufficient fact's or data. The district court also concluded that even if these reports were admissible, they fail to raise a genuine dispute of material fact [813]*813because they do not sufficiently address causation. The district court denied Pyramid’s motion for reconsideration.
STANDARD OF REVIEW
We review a district court’s order granting summary judgment de novo. Ford v. City of Yakima, 706 F.3d 1188, 1192 (9th Cir.2013) (per curiam). We review eviden-tiary rulings for abuse of discretion and reverse if the exercise of discretion is both erroneous and prejudicial. Nev. Dep’t of Corr. v. Greene, 648 F.3d 1014, 1018 (9th Cir.2011). We review underlying factual determinations for clear error. United States v. Lukashov, 694 F.3d 1107, 1114 (9th Cir.2012).
DISCUSSION
A. Exclusion of Pyramid’s Expert Witnesses
1. Legal Standards
Rule 702 of the Federal Rules of Evidence provides that expert opinion evidence is admissible if: (1) the witness is sufficiently qualified as an expert by knowledge, skill, experience, training, or education; (2) the scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (3) the testimony is based on sufficient facts or data; (4) the testimony is the product of reliable principles and methods; and (5) the expert has reliably applied the relevant principles and methods to the facts of the case. Fed. R.Evid. 702.
Under Daubert, 509 U.S. at 579, 113 S.Ct. 2786, and its progeny, including Daubert v. Merrell Dow Pharm., Inc., 43 F.3d 1311 (9th Cir.1995), a district court’s inquiry into admissibility “is a flexible one.” Alaska Rent-A-Car, Inc. v. Avis Budget Grp., Inc., 738 F.3d 960, 969 (9th Cir.2013) (citation omitted), cert. denied, - U.S. • — , 134 S.Ct. 644, 187 L.Ed.2d 420 (2013). In evaluating proffered expert testimony, the trial court is “a gatekeeper, not a fact finder.” Primiano v. Cook, 598 F.3d 558, 565 (9th Cir.2010) (citation and quotation marks omitted).
“[T]he trial court must assure that the expert testimony ‘both rests on a reliable foundation and is relevant to the task at hand.’ ” Id. at 564 (quoting Daubert, 509 U.S. at 597, 113 S.Ct. 2786). “Expert opinion testimony is relevant if the knowledge underlying it has a valid connection to the pertinent inquiry. And it is reliable if the knowledge underlying it has a reliable basis in the knowledge and experience of the relevant discipline.” Id. at 565 (citation and quotation marks omitted). “Shaky but admissible evidence is to be attacked by cross examination, contrary evidence, and attention to the burden of proof, not exclusion.” Id. at 564 (citation omitted). The judge is “supposed to screen the jury from unreliable nonsense opinions, but not exclude opinions merely because they are impeachable.” Alaska Rent-A-Car, 738 F.3d at 969. Simply put, “[t]he district court is not tasked with deciding whether the expert is right or wrong, just whether his testimony has substance such that it would be helpful to a jury.” Id. at 969-70.
Like the test for admissibility in general, the test of reliability is also flexible. Estate of Barabin v. AstenJohnson, Inc., 740 F.3d 457, 463 (9th Cir.2014) (en banc). To determine reliability, the Supreme Court has suggested several factors: “1) whether a theory or technique can be tested; 2) whether it has been subjected to peer review and publication; 3) the known or potential error rate of the theory or technique; and 4) whether the theory or technique enjoys general acceptance within the relevant scientific commu[814]*814nity.” Id. (quoting United States v. Hankey, 203 F.3d 1160, 1167 (9th Cir.2000)); see also Primiano, 598 F.3d at 564. These factors are “meant to be helpful, not definitive, and the trial court has discretion to decide how to test an expert’s reliability as well as whether the testimony is reliable, based on the particular circumstances of the particular case.” Primiano, 598 F.3d at 564 (citations and quotation marks omitted); see also Barabin, 740 F.3d at 463. The test “is not the correctness of the expert’s conclusions but the soundness of his methodology,” and when an expert meets the threshold established by Rule 702, the expert may testify and the fact finder decides how much weight to give that testimony. Primiano, 598 F.3d at 564-65.
After an expert establishes admissibility to the judge’s satisfaction, challenges that go to the weight of the evidence are within the province of a fact finder, not a trial court judge. A district court should not make credibility determinations that are reserved for the jury.
2. Spiegel
Spiegel is a certified restorer with the National Institute of Disaster Restoration. He is also a certified: (1) master restorer and water, fire and odor control journeyman with the Institute of Inspection, Cleaning, and Restoration; (2) indoor environmentalist and mold remediator with the Indoor Air Quality Association; and (3) Level I thermographer with the Infrared Training Center. Spiegel is also a general and specialty licensed contractor with the state of California and has 38 years of experience in property damage repair and more than 15 years of experience in construction defect investigation.
In two conelusory sentences and without analysis or explanation, the district court held that Spiegel was not a qualified expert “on the scientific, technical, or specialized data on which he purports to opine” and that his opinion regarding relative humidity was not based on sufficient facts or data and was not the product of reliable principles and methods. The district court abused its discretion in reaching these conclusions.
Spiegel used weather data from the time of the incident, thermo-hygrome-ter and infrared data, and ambient condition data to opine on the level of humidity in the warehouse at the time of the flood. As noted, Spiegel is a certified thermogra-pher, certified indoor environmentalist, certified master restorer, certified water control journeyman, and certified mold remediator, with decades of experience. This expertise and experience is relevant to the issues on which Spiegel opined. Because the district court provided no explanation or analysis for rejecting these qualifications, the district court abused its discretion in summarily determining that Spiegel was not qualified as an expert. See Barabin, 740 F.3d at 464 (holding the district court “failed to assume its role as gatekeeper” when it excluded expert testimony for “dubious credentials” without conducting a Daubert hearing or assessing expert’s findings). Spiegel’s many relevant certifications and decades of relevant experience render him qualified to issue his expert opinion.
In addition, in preparing his report Spiegel conducted two site visits to Pyramid’s warehouse, interviewed several Pyramid employees who saw the water intrusion and its immediate aftermath, reviewed ServPro’s ambient condition measurements, and recorded his own ambient data during his visits, including data from a 5:30 a.m. visit designed to compare the difference in indoor and outdoor conditions when the warehouse was closed and locked [815]*815(as it was during the flood). Spiegel also took digital photographs and electronic thermo-hygrometer readings, performed infrared imaging, reviewed www.weather underground.com to determine the weather conditions at a nearby airport at the time of the flood, and reviewed the Helms report. These facts and data constitute a sufficient basis for Spiegel’s expert report.
Although not discussed by the district court, Spiegel relies on more facts and data in reaching Ms expert conclusions than did Hartford’s expert witness. Hartford’s expert Helms spent approximately two hours conducting a visual inspection of the warehouse but did not take any measurements, thermographic readings, infrared images, or other data. Although Helms contacted ServPro for its readings, Helms did not learn the locations from which ServPro obtained its readings, what ServPro did to get its readings, or what kind of detection machine ServPro used. The day after his two-hour site visit, Helms completed his report and concluded that based on the humidity levels measured by ServPro, no damage occurred to the components in question.
Spiegel also adequately explained his methodology in reaching his opinion. Spiegel described how the data he collected and reviewed helped him determine the conditions of the warehouse at the time of the event and at the time of ServPro’s measurements (which were relied on by Helms to determine that no damage from humidity could have occurred). Unlike Helms, however, Spiegel took into consideration the fact that the warehouse doors were closed and locked during the flood but open during ServPro’s measurements, and Spiegel calculated and considered the difference between the indoor and outdoor conditions during Ms 5:30 a.m. visit. Spiegel applied that difference to the weather data at the nearby airport on the night of the incident to extrapolate the indoor conditions on the night of the flood. Spiegel also used the infrared temperature readings of the packages on the shelves in comparison to the air temperature taken during his visit to determine the likely temperature of the packages on the night of the incident.
The record shows that the knowledge underlying Spiegel’s report “has a reliable basis in the knowledge and experience of the relevant discipline,” rendering his report reliable. Primiano, 598 F.3d at 565 (citation and quotation marks omitted). The record also shows that Spiegel’s reliance on the nearby airport weather information from www.weatherunderground. com is acceptable in the industry, for Helms testified that he relied on the same data. Spiegel also explained how he applied the data to reach his conclusions and how Helms failed to rely on proper data to reach Ms conclusion. In short, Spiegel’s principles and methods were reliable and his report is not one of the “unreliable nonsense opinions” that should be screened from use. Alaska Rent-A-Car, 738 F.3d at 969. Thus, the district court abused its discretion in excluding this evidence.
Excluding the Spiegel report was both erroneous and prejudicial. Spiegel’s expert report provides evidence that: (1) the Helms report relied on improper data to conclude that no damage was caused by humidity following the flood; (2) during the flood, the humidity exceeded 90% and was above the dew point; (3) the applicable standards for moisture-proof packaging require a one-year shelf life and humidity below 90% to prevent failure; (4) the vast majority of the parts stored in moisture-proof packaging was well beyond the one-year warranty protection; and (5) the conditions during the flood “without question” put the affected moisture-proof [816]*816packages outside the packing standards. Thus, as Spiegel opines, the humidity and condensation caused by the flood may have compromised the packaging and possibly the components themselves. This is admissible evidence from which causation and damage reasonably may be inferred. Although Spiegel did not say with certainty that the humidity from the flood caused damage to Pyramid’s inventory, a jury could reasonably infer causation from Spiegel’s report and Pyramid’s other evidence. It is not necessary for Spiegel’s report to establish every element of Pyramid’s claim in order for it to be admissible in evidence. See Primiano, 598 F.3d at 564 (“Reliable expert testimony need only be relevant, and need not establish every element that the plaintiff must prove, in order to be admissible.”) (citation omitted).
If Spiegel’s report had been admitted, the district court would have been required to view it in the light most favorable to Pyramid when considering Hartford’s motion for summary judgment. Because the report could assist a trier of fact in inferring that the flood caused sufficiently high humidity to damage Pyramid’s parts and that Helms’ contrary conclusion was not reliable, the exclusion of Spiegel’s report is prejudicial to Pyramid. See Messick v. Novartis Pharm. Corp., No. 13-15433, 747 F.3d 1193, 1197-99, 2014 WL 1328182, at *3-5 (9th Cir. Apr. 4, 2014) (reversing grant of summary judgment where expert’s testimony, which would have created a genuine issue of material fact, was excluded because it was erroneously deemed unreliable and irrelevant).
3. Pytlewski
The district court also excluded the Pyt-lewski report under Rule 702, stating that it was not the product of reliable principles and methods, and under Rule 701,2 stating that the report does not explain how the opinions are rationally based on the perceptions of the witness. Because Pytlew-ski was proffered as an expert witness, however, only Rule 702 applies.
Pytlewski was retained to review and comment on the report prepared by Hartford’s expert Dr. Kumar of SEAL Laboratories. The district court appears to have accepted Pytlewski as a qualified expert, excluding his report solely based on reliability. The district court concluded that Pytlewski’s report was not the product of reliable principles and methods.
An expert opinion is reliable “if the knowledge underlying it has a reliable basis in the knowledge and experience of the relevant discipline.” Alaska Rent-A-Car, 738 F.3d at 969 (quoting Primiano, 598 F.3d at 565). Pytlewski’s opinions are based on his knowledge and experience as a professional engineer and metallurgist. For example, Pytlewski countered Dr. Kumar’s statement that the damage to Pyramid’s inventory could not have been caused by the flood because the corrosion was not uniform by explaining that metal exposed to moisture would not oxidize uniformly. This opinion is within the knowledge and experience of a metallurgist. See, e.g., Stilwell v. Smith & Nephew, Inc., 482 F.3d 1187, 1192-93 (9th Cir.2007) (holding admissible metallurgist’s expert testimony that nails were poorly manufactured and could have been designed to last longer); White v. Ford [817]*817Motor Co., 312 F.3d 998, 1008 (9th Cir.2002) opinion amended on denial of retig, 335 F.3d 833 (9th Cir.2003) (describing metallurgist’s testimony that was “well within his metallurgical expertise” when he “identified wear on the ratchet wheel of the brake that showed repeated tip-on-tip engagement rather than the proper engagement”); see also 6 Am.Jur. Trials 555 (describing common use of expert metallurgists to describe, among other things, causes and effects of corrosion). The “reliability” test is flexible and should be applied based on the circumstances of the case. Given the subject matter and type of opinions that Pytlewski rendered and his knowledge and experience, Pytlewski’s opinion is reliable. See Messick, 747 F.3d at 1197-98, 2014 WL 1328182, at *3.
The exclusion of Pytlewski’s report is prejudicial because his report provides evidence from which a fact finder could disregard the opinion of Dr. Kumar and reasonably infer damages and causation relating to the flood. Further, Pytlewski’s deposition testimony links the flood as at least a partial cause of the damage to Pyramid’s inventory. Thus, the district court abused its discretion in excluding Pytlewski’s expert report and testimony.
4. Mortenson
The district court also excluded the Mortenson report under Rule 702, stating it was not reliable, and under Rule 703,3 stating it was not based on facts or data known to Mortenson. Mortenson opined that it was improper for Hartford’s expert Dr. Kumar to use military standards of suitability instead of commercial standards in testing the selected parts from Pyramid’s inventory. Mortenson testified at deposition, however, that he did not know what standards should have been used and that he is not aware of the governing commercial standards because that is not his field of expertise. Thus, Mortenson’s testimony that Dr. Kumar should not have used military standards and should have used commercial standards was not based on facts or data known to Mortenson and is inadmissible under Rule 703. Additionally, Morten-son’s testimony is not reliable because he did not have the knowledge or experience required under Rule 702 to permit him to give expert testimony in this matter. The exclusion of Mortenson’s report by the district court was not an abuse of discretion. See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 153-54, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999) (holding trial court did not abuse its discretion to exclude expert testimony on grounds that expert’s methodology in analyzing relevant data was unreliable, even though the expert was qualified, where there was no evidence that other experts in the industry used the expert’s particular approach).
B. Grant of Summary Judgment Against Pyramid’s Claims
In its complaint, Pyramid alleges that Hartford breached its insurance contract and the implied covenant of good faith and fair dealing by: (1) failing properly to investigate and then improperly denying Pyramid’s claim for an alleged loss of inventory, including refusing to test the in[818]*818ventory; (2) failing properly to respond to the building restoration claim, including making a “low-ball” estimate for damage that was one-fourteenth of the value ultimately paid by Hartford and unreasonably delaying final payment on the restoration claim; and (3) refusing to pay for an alleged business interruption. Pyramid appeals the district court’s grant of summary judgment in favor of Hartford and against both of Pyramid’s claims.
A party is entitled to summary judgment if the “movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed.R.Civ.P. 56(a). The moving party has the burden of establishing the absence of a genuine dispute of material fact. Celotex Corp. v. Catrett, 477 U.S. 317, 323, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986).
In this diversity action, the substantive law governing Pyramid’s breach of contract claim is California law. See Neely v. St. Paul Fire & Marine Ins., 584 F.2d 341, 345 (9th Cir.1978). Whether evidence on a particular issue is sufficient to raise a question of fact for the jury, however, is governed by federal law. Id. The federal test is whether a “reasonable jury viewing the summary judgment record could find by a preponderance of the evidence that the plaintiff is entitled to a favorable verdict.” Narayan v. EGL, Inc., 616 F.3d 895, 899 (9th Cir.2010) (citation omitted). The judge is not to weigh the evidence, draw legitimate inferences, or make credibility determinations. Id. “The evidence of the non-movant is to be believed and all justifiable inferences drawn in his favor.” Id. (citation and quotation marks omitted). An inference is justifiable if it is rational or reasonable — it does not need to be the most likely or most persuasive inference. Id. The inferences must have a sufficient evidentiary basis. Neely, 584 F.2d at 345-46. ‘Where conflicting inferences may be drawn from the facts, the case must go to the jury.” Munger v. City of Glasgow Police Dep’t, 227 F.3d 1082, 1087 (9th Cir.2000) (citation and quotation marks omitted).
2. Pyramid’s Claim of Loss of Inventory
Pyramid argues that Hartford breached the insurance contract by both failing properly to investigate and denying Pyramid’s claim under the Policy for loss of inventory. Under the Policy, the burden is on Pyramid to initiate and support its claim. See 1231 Euclid Homeowners Ass’n v. State Farm Fire & Cas. Co., 135 Cal.App.4th 1008, 37 Cal.Rptr.3d 795, 802 (2006). To succeed on its breach of contract claim, Pyramid must establish a contract, Pyramid’s performance or excuse for nonperformance, Hartford’s breach, and resulting damages to Pyramid. Abdelhamid v. Fire Ins. Exch., 182 Cal.App.4th 990, 106 Cal.Rptr.3d 26, 32-33 (2010). The disputed issues in this case are whether Pyramid suffered damage to any of its inventory and, if so, whether that damage was caused by the flood.
a. Damages
It is undisputed that at least some of Pyramid’s inventory had visible corrosion, tarnish, or discoloration. Hartford’s expert Dr. Kumar identified visible corrosion, tarnish, or discoloration on more than 40% of the items he tested. Pyramid quarantined more than 250,000 items as showing visible signs of corrosion, tarnish, or discoloration. Dr. Kumar testified that corrosion constituted actual physical damage. Thus, at least some of Pyramid’s inventory had actual, physical damage. When there is actual, physical damage, [819]*819then the diminution of market value may be a proper measure of damages. See State Farm Fire & Cas. Co. v. Superior Court, 215 Cal.App.3d 1435, 264 Cal.Rptr. 269, 274-75 (1989) (diminution of value is not a cause of a loss but a measure of damages).
The parties dispute whether visible corrosion, tarnish, or discoloration of a part is sufficient to constitute a “failure” under “military standards” of suitability. Dr. Kumar determined that only two out of the 147 parts failed under that standard, whereas Pytlewski noted that under the Department of Defense Test Method Standards of Microcircuits visible corrosion is included as a failure criterion. Whether visible corrosion constitutes a “failure” of a part under military standards is a factual dispute for the jury to resolve.
In addition, and even more importantly, regardless of this dispute over whether visible corrosion, tarnish, or discoloration constitutes a failure under military standards, two of the 374 parts (0.535 percent) actually tested by Dr. .Kumar failed, even under Dr. Kumar’s standards. Extrapolating this failure percentage to 52 million parts is evidence of at least some failure (approximately 278,200 parts). There is also evidence in the record that some customers returned a few parts because of corrosion. The failure of some parts and the return of other parts are evidence from which a jury reasonably could infer that Pyramid was harmed by the presence of corrosion on at least some of its inventory.
Additionally, if the inventory items were damaged by the flood, which Hartford admits was a covered event, the inability to sell the items due to the physical damage, regardless of whether those items would fail a Department of Defense test under military standards, would constitute a covered loss. See, e.g., MRI Healthcare Ctr. of Glendale, Inc. v. State Farm Gen. Ins. Co., 187 Cal.App.4th 766, 115 Cal.Rptr.3d 27, 37 (2010) (“In modern [insurance] policies, ‘physical loss or damage’ is typically the trigger for coverage. Clearly, this threshold is met when an item of tangible property has been ‘physically altered’ by perils such as fire or water.” (citation omitted)); see also Allstate Ins. Co. v. Smith, 929 F.2d 447, 450 (9th Cir.1991) (describing an “easy to imagine” situation where a “leaky roof could lead to water damage to [someone’s] property. Presumably, water damage would be an ensuing loss covered by the policy but repairing the roof would not be covered.”); Meridian Textiles, Inc. v. Indem. Ins. Co. of N. Am., No. CV 06-4766 CAS, 2008 WL 3009889 at, *4-6 (C.D.Cal. Mar. 20, 2008) (yarn that was water-damaged, had a tangible change such as odor, mold or mildew, or had a detectable change such that the yarn was likely to develop odor, mold or mildew and was, therefore, unable to be sold, is a covered loss); Columbiaknit, Inc. v. Affiliated FM Ins. Co., No. Civ. 98-434-HU, 1999 WL 619100 at, *5-6 (D.Or. Aug. 4, 1999) (fabric with mold, odor, or with increased microbial counts that will develop mold or odor and unable to be sold, is a covered loss). Drawing all reasonable inferences in favor of Pyramid, a reasonable fact finder could find that some of the inventory items had moisture-related damage that diminished their market value. That diminution in market value is a recoverable measure of damages. See State Farm, 264 Cal.Rptr. at 274-75.
Because there are at least some parts that have actual physical damage and some parts that failed testing or were returned by customers, there is evidence from which a jury could determine that Pyramid suffered harm to its inventory. The fact that some, but not all, of the inventory was damaged does not support the granting of [820]*820summary judgment against Pyramid’s loss of inventory claim. Determining the amount of harm suffered is for the jury.
b. Causation
Hartford argues that Pyramid failed to produce any evidence that the components’ corrosion, tarnish, or discoloration was caused by the flood of August 11, 2005, as opposed to the age of the parts, the lack of climate control in the warehouse, or other potential causes. Because California provides the substantive law in this case, we follow California’s law on causation in an insurance coverage claim.
Under California law, the “efficient proximate cause” doctrine is “the preferred method for resolving first party insurance disputes involving losses caused by multiple risks or perils, at least one of which is covered by insurance and one of which is not.” Julian v. Hartford Underwriters Ins. Co., 35 Cal.4th 747, 27 Cal.Rptr.3d 648, 110 P.3d 903, 906 (2005) (citations omitted); see also Brown v. Mid-Century Ins. Co., 215 Cal.App.4th 841, 156 Cal.Rptr.3d 56, 67 (2013) (noting that “the efficient proximate cause doctrine applies when a loss is caused by a combination of a covered and specifically excluded risks”) (citation and quotation marks omitted).
The “ ‘efficient proximate cause’ of a loss is the predominant, or most important cause of a loss.” Julian, 27 Cal.Rptr.3d 648, 110 P.3d at 907 (citation omitted). Coverage would not exist “if the covered risk was simply a remote cause of the loss” or if an excluded risk was the efficient proximate cause of the loss. Id.; see also California Ins.Code. § 530 (“An insurer is liable for a loss of which a peril insured against was the proximate cause, although a peril not contemplated by the contract may have been a remote cause of the loss; but he is not liable for a loss of which the peril insured against was only a remote cause.”). “If more than one peril contributes to a loss, the question which is the efficient proximate cause generally is a factual matter for the jury to resolve.” Julian v. Hartford Underwriters Ins. Co., 123 Cal.Rptr.2d 767, 770 (Cal.Ct.App.2002), review granted and opinion superseded sub nom. Julian v. Hartford Underwriters, 126 Cal.Rptr.2d 896, 57 P.3d 362 (Cal. 2002), and aff'd, 35 Cal.4th 747, 27 Cal.Rptr.3d 648, 110 P.3d 903 (2005); see also Garvey v. State Farm Fire & Cas. Co., 48 Cal.3d 395, 257 Cal.Rptr. 292, 770 P.2d 704, 714 (1989) (en banc) (“Coverage should be determined by a jury under an efficient proximate cause analysis.”).
The evidence in the record, including both the Spiegel report and the testimony by employees that they saw condensation on the packaging of the parts stored on the lower shelves, supports an inference that the humidity reached a high enough level during the flood to cause significant condensation on the packaging of the parts kept on the lower three or four shelves. There is also evidence that although Pyramid’s moisture-sensitive inventory was generally stored in moisture-proof packaging, most of that packaging was either unsealed or had been compromised by age or the high humidity levels, allowing moisture to reach the components. Pyramid’s employees testified that many of the moisture-proof packages were unsealed and some of those had been folded down and held closed only with a paperclip. The Spiegel report noted that the vast majority of Pyramid’s moisture-proof packages were much older than the industry standard packaging shelf-life recommendation of 12 months and that the moisture-proof package standards require lower than dew-point humidity. Thus, there is evidence from which a jury could infer that moisture from the flood may have reached moisture-sensitive components because the [821]*821packaging was not fully sealed, failed from age, or failed because the humidity caused by the flood reached above the dew point.
Pytlewski, Pyramid’s expert metallurgist, acknowledged in his report the difficulty in determining causation, but testified at a deposition that he believed that at least some of the damage was caused by the water intrusion that occurred during the flood. Although he could not state definitively that the water intrusion caused all of the harm, “[l]aek of certainty is not, for a qualified expert, the same thing as guesswork.” Primiano, 598 F.3d at 565. Additionally, as discussed above, there is evidence supporting an inference that the condensation may have breached the packaging and reached the components. Further, the fact that more than 250,000 items were quarantined because they showed visible signs of moisture-related damage after the flood is evidence supporting an inference of causation.
The existence of material factual issues relating to causation is further evident in reviewing the reports of Hartford’s experts. Although Helms opined that the humidity caused by the flood could not have caused any damage to Pyramid’s inventory, there is evidence in the record from which a reasonable fact finder could discredit Helms’ conclusion. Helms spent only two to three hours conducting a visual inspection of Pyramid’s warehouse and did not conduct any tests, take any building measurements, or otherwise investigate the inventory. It took Helms one day to complete his report, and he relied exclusively on ServPro’s readings and assumed they were accurate. Helms also does not appear to have considered that ServPro’s measurements were taken more than 24 hours after the drying operation began and with open warehouse doors or that the moisture-proof packaging was compromised on many items. It is the jury’s province to determine how much weight, if any, to give the conclusions reached by Helms or any of the experts at trial.
Dr. Kumar did not provide a definitive statement on causation and gave contradictory statements. His declaration and part of his report state that the water intrusion did not cause the corrosion found in his testing, but his deposition testimony and another portion of his report state that the cause cannot be determined because it could have been the water intrusion or it could have been age or some other exposure to moisture. Under California’s efficient proximate cause doctrine, whether the damage found by Dr. Kumar was caused by the flood or by some other cause is an issue for the jury. See Julian, 123 Cal.Rptr.2d at 770.
Analyzing the causal chain is “necessarily speculative” and “[cjhoosing between the speculations is ordinarily a question for the trier of fact, who must determine the balance of probabilities.” Shawmut Bank, N.A. v. Kress Assocs., 33 F.3d 1477, 1496 (9th Cir.1994) (citation omitted). This is not a case where the party with the burden of proof at trial submitted “no evidence” from which causation could be inferred, thereby requiring summary judgment. Id. at 1497-98. There is sufficient evidence in the record, drawing all reasonable inferences in Pyramid’s favor, from which a reasonable jury “could find by a preponderance of the evidence that the plaintiff is entitled to a favorable verdict.” Narayan, 616 F.3d at 899 (citation omitted). Whether those inferences should be drawn in favor of Pyramid after considering and weighing all of the evidence is for a jury to decide. Thus, summary judgment is inappropriate against Pyramid’s claim of loss of inventory.
3. Pyramid’s Claim of Business Interruption
Pyramid submitted, withdrew, re-submitted, and then modified its business in[822]*822terruption claim. The only lost business currently claimed by Pyramid is from WMS, a potential customer. The burden is on Pyramid to initiate and support this claim under the Policy. See 1231 Euclid, 37 Cal.Rptr.3d at 802. Thus, to defeat summary judgment against its business interruption claim against Hartford, Pyramid must show that there are material factual issues about whether its loss of potential WMS business is a covered loss under the Policy.
Business interruption is covered under the Policy, through an endorsement entitled “Gross Earnings” that deleted and replaced the Policy’s original “Business Income” and “Extra Expense” provisions. The Gross Earnings endorsement provides:
We will pay for the actual loss of Business Income you sustain and the actual, necessary and reasonable Extra Expense you incur due to the necessary interruption of your business “operations” during the “Period of Restoration” due to the direct physical loss of or direct physical damage caused by or resulting from a Covered Cause of Loss to property at “Scheduled Premises.”
Pyramid must, therefore, show that it actually lost business income from WMS because of the flood in order to prevail on its claim that Hartford breached the insurance contract by failing to cover Pyramid’s alleged business interruption.
Pyramid fails to show that there are material issues of fact that it actually lost WMS business as a result of the flood. Whether WMS would have contracted "with Pyramid if there was no flood, and for what amount, is too speculative to support Pyramid’s claim. Although Carmine Greco, the senior buyer at WMS, testified that he had “committed” to buy at least $1 million dollars of inventory from Pyramid, he also explained that any purchase by WMS from Pyramid was subject to the approval of WMS quality control people and subject to additional negotiation. Pyramid and WMS, thus, were still negotiating and had not yet entered into a binding and enforceable contract.
Greco further testified that he and Pyramid had discussed contract terms for only a few parts, although he “anticipated” more parts would be purchased from Pyramid. He added that after learning of the flood, he needed the inventory tested before he would consider purchasing any of it. He also stated that if the product had been tested, he “probably” would have bought it.
WMS’s quality control manager, Nick Savich, also participated in the site visit at Pyramid. Savich sent a letter to Pyramid after the visit, stating that WMS did not approve Pyramid as a supplier based on water intrusion. Savich testified at deposition, however, that if he had known that Pyramid did not have humidity control in its warehouse, that fact alone would have disqualified Pyramid as a supplier.
To find for Pyramid on its claim of business interruption, a jury would need to speculate that Savich would not have discovered the fact that the warehouse did not have humidity control, that WMS quality control people would have signed off on Pyramid as a supplier, that Greco and Pyramid would have successfully negotiated pricing and other terms for many additional parts, and that Greco would have followed through with a large purchase. The record evidence does not support— and the law does not permit — such speculation. See Neely, 584 F.2d at 346 (“Parties are entitled to have the determination of their rights rest on more than speculation and guesswork. Here, the connection between the proffered evidence and the conclusions is too tenuous to permit a jury to make it.”). Accordingly, we affirm the [823]*823district court’s grant of the motion for summary judgment against Pyramid on the claim of business interruption.
4. Pyramid’s Claim of Good Faith and Fair Dealing
“The covenant of good faith and fair dealing has ‘particular application’ to insurers because they are ‘invested with a discretionary power affecting the rights of another.’ ” Amadeo v. Principal Mut. Life Ins. Co., 290 F.3d 1152, 1161 (9th Cir.2002) (quoting Carma Developers (Cal.), Inc. v. Marathon Dev. Cal., Inc., 2 Cal.4th 342, 6 Cal.Rptr.2d 467, 826 P.2d 710, 726 (1992)). Under California law, to establish a breach of the implied covenant of good faith and fair dealing, “a plaintiff must show: (1) benefits due under the policy were withheld; and (2) the reason for withholding benefits was unreasonable or without proper cause.” Guebara v. Allstate Ins. Co., 237 F.3d 987, 992 (9th Cir.2001) (citing Love v. Fire Ins. Exch., 221 Cal.App.3d 1136, 271 Cal.Rptr. 246, 255 (1990)). The reasonableness of an insurer’s conduct is ordinarily a question of fact. Amadeo, 290 F.3d at 1161.
An insured’s claim of breach of the implied covenant of good faith and fair dealing may be dismissed on summary judgment if the defendant insurer can show that there was a “genuine dispute” as to liability. Guebara, 237 F.3d at 992. This genuine dispute doctrine should be applied on a case-by-ease basis and does not protect allegedly biased investigations. Id. at 994, 996. Biased investigation claims include circumstances where: (1) the insurer misrepresents the nature of the investigatory proceedings; (2) the insurer’s employees lie during depositions or to an insured; (3) the insurer dishonestly selects experts; (4) the insurer’s experts were unreasonable; or (5) the insurer fails to conduct a thorough investigation. Id. at 996.
Pyramid argues that Hartford breached the implied covenant of good faith and fair dealing by: (1) improperly refusing to test Pyramid’s inventory for more than two years and conducting inadequate testing; (2) denying coverage on Pyramid’s claim of inventory loss; (3) making a “low-ball” building restoration estimate and delaying payment on the agreed-upon supplemental building restoration amount for four months; and (4) conducting a biased investigation through dishonestly selecting unreasonable experts and an inadequate investigation. Hartford argues that it did not act unreasonably because there was a genuine dispute as to coverage, its investigation was proper, its coverage denial was based on expert opinions, and it ultimately paid all of the building restoration amounts making any delay harmless.
There is evidence in the record supporting an inference that Hartford acted unreasonably or with bias. Pyramid submitted testimony stating that a Hartford representative, while visiting Pyramid’s premises to assess damages, said that it is Pyramid’s job to try to collect under the insurance policy and Hartford’s job to make sure Pyramid does not collect. Pyramid also provided evidence that Hartford’s adjuster Todd Klingaman “downplayed” the flood damage during his site visit by comparing it to a “bucket of water” and suggesting to Pyramid that it did not need to tell its customers about it. Pyramid’s employees also testified that at the conclusion of Hartford’s site visit, before Helms had completed his report, Hartford jumped to an early conclusion that no damage had occurred and thus refused to test any inventory. Additionally, Hartford relied on Helms’ conclusion that humidity could not have caused damage, even though Helms conducted only a [824]*824cursory investigation, relied on readings taken after the drying operation had largely concluded and under different conditions, and was not familiar with many of the electronic components or how moisture affected them.4 Moreover, Hartford did not test any of Pyramid’s inventory until two years after the flood and only after Pyramid had engaged its own expert and after Helms supplemented his report to suggest testing a representative sample might be appropriate. Finally, Hartford’s initial estimate of building damage was very low, $6,640.34, and Hartford did not agree to additional sums for eight months and then delayed for another four months before paying the additional $88,480.01. To the extent a jury finds coverage was required for the inventory loss claim, this evidence further supports an inference that Hartford’s conduct was unreasonable and not entitled to protection under the genuine dispute doctrine.
Further, the admitted portions of the declaration and report of Gene Irizarry, Pyramid’s insurance claim expert, support an inference that Hartford did not handle the claim in good faith. Irizarry opined that Hartford could not have reasonably concluded that Pyramid’s inventory was not damaged in the absence of any testing or investigation and should have conducted testing under the adjustment expense that accompanies every claim. Irizarry also concluded that Hartford was not responsive to the needs of Pyramid and failed to assist Pyramid as Hartford was obligated to do and that Hartford forced Pyramid to engage experts and conduct an investigation that should have been done by Hartford. Irizarry further stated that Hartford made an unreasonable, “low-ball” estimate for the building repairs and unreasonably delayed in investigating and finalizing the building restoration claim.
Hartford also relied on Dr. Kumar’s report, issued two years after the flood, to deny Pyramid’s claim. An expert report alone, however, does not demonstrate a “genuine dispute.” See Guebara, 237 F.3d at 996. Further, even if Dr. Kumar’s report provides Hartford with a “genuine dispute” as of August 2007, the two-year delay and the lack of a thorough investigation during the time period between the August 2005 flood and the August 2007 Kumar report supports Pyramid’s claim for breach of the implied covenant of good faith and fair dealing. See, e.g., Amadeo, 290 F.3d at 1163 (good faith decisions based on an inadequate investigation can support a claim for breach of good faith and fair dealing).
“[Wjhether an insurer’s denial of a claim is unreasonable is dependent upon the facts in each case. The issue remains a question of fact unless only one inference may be drawn from the evidence.” Paulfrey v. Blue Chip Stamps, 150 Cal.App.3d 187, 197 Cal.Rptr. 501, 504 (1983) (emphasis in original) (citations omitted). Summary judgment cannot be granted under the genuine dispute doctrine in a bad faith claim unless “it is undisputed or indisputable that the basis for the insurer’s denial of [825]*825benefits was reasonable — for example, where even under the plaintiffs version of the facts there is a genuine issue as to the insurer’s liability under California law.” Amadeo, 290 F.3d at 1161 (citation omitted). Pyramid produced evidence from which a reasonable jury could draw more. than one inference concerning Hartford’s conduct. Thus, summary judgment against Pyramid’s claim for breach of the implied covenant of good faith is inappropriate.
CONCLUSION
The district court abused its discretion in excluding the expert evidence of David Spiegel and Ken Pytlewski. Such evidence is admissible. The district court did not abuse its discretion in excluding the expert evidence of Del Mortenson. The district court erred in granting summary judgment against Pyramid’s claims because genuine disputes of material fact exist as to whether Hartford breached its contract with Pyramid and breached the implied covenant of good faith. To the extent such claims are premised on Pyramid’s business interruption theory, however, no material issues of fact exist, and the district court did not err in granting summary judgment against that theory of liability.
AFFIRMED IN PART, REVERSED IN PART, and REMANDED FOR TRIAL. The parties shall bear their own costs on appeal.