ROGERS, Circuit Judge.
This product liability action arose from an oven explosion that seriously injured
plaintiff Nancy Siegel. Siegel sued the oven’s manufacturer, Dynamic Cooking Systems, Inc., who in turn filed an indemnity claim against Burner Systems Inc., the manufacturer of the oven’s temperature regulator. Experts for both sides agreed that a gas leak in the regulator caused the explosion and that Siegel was not eontributorily negligent. However, the experts could not determine whether the leak resulted from a defective oven design attributable to Dynamic, or from a manufacturing defect attributable to Burner Systems. Because of this ambiguity, the district court
dismissed the indemnity claim. This was proper because Dynamic failed to bring forth evidence that a manufacturing defect in the regulator was a probable, rather than merely a possible, cause of the explosion. At the ensuing trial, the district court entered a directed verdict for Dynamic because Siegel could not prove whether Dynamic or Burner Systems was at fault for the explosion. This was improper, as Siegel presented two potential causes for the explosion, both of which placed liability on Dynamic.
I
Siegel bought her Dynamic-made range from a local retailer in 2006, and used it for nearly two years without incident. In February 2008, Siegel was warming food on the stove top of the range when she heard an unusual noise coming from the oven. Siegel opened the oven door to investigate when a fireball came out, knocking her down and seriously burning her.
Siegel and her insurance company, Kentucky Farm Bureau Mutual Insurance Company, sued Dynamic under various theories of product liability. After preliminary discovery indicated that the range’s regulator was the source of a gas leak, Dynamic filed a third-party indemnity claim against Burner Systems, the regulator’s manufacturer. Neither Siegel nor Kentucky Farm asserted claims directly against Burner Systems.
Each party hired experts to help determine the cause of the accident. The experts agreed that a gas leak originated in the regulator, causing the explosion, but differed as to the cause of the leak. The dispute centered on whether Burner Systems had defectively manufactured the regulator or whether Dynamic had defectively designed the oven. The expert for Dynamic, Mark Mulcahy, identified three possible causes of the regulator leak: a manufacturing defect, a break in the regulator’s seal caused by gas pressure, and degradation of the seal.
By contrast, Burner Systems’s expert, Thomas Crane, concluded that the regulator leaked because Dynamic defectively designed the oven. After testing an exemplar range, Crane found that the oven’s design exposed the regulator to heat that exceeded the regulator’s maximum operating temperature. Crane opined that this excess heat caused a degradation of the regulator’s diaphragm seal, making the oven’s design the presumptive cause of the explosion.
Donan Engineering, the expert for Sie-gel and Kentucky Farm, provided an initial report concluding that the cause of the fire was a “manufacturer’s defect.” The report, however, did “not specify the precise nature of the alleged manufacturing defect,” how the expert reached its conclusions, or which manufacturer caused the defect.
Siegel v. Kentucky Farm Bureau
Mut. Ins.,
No. 3:08CV-00429, 2010 WL 3000746, at *2 (W.D.Ky. July 26, 2010). In response to expert reports by Burner Systems and Dynamic experts, Donan Engineering muddied the waters even further. Donan asserted that, if Crane’s measurements could be believed, the explosion could have been caused by a design defect. However, even if Crane’s measurements were disproved, Donan stated that “[i]t can still be concluded to a reasonable degree of engineering certainty that the cause and origin of the loss was internal to the [Dynamic] range.”
Based on Crane’s report, Burner Systems moved for summary judgment on Dynamic’s indemnity claim. The district court granted the motion because the experts could not state that a manufacturing defect was the probable, rather than just a possible, cause of the leak.
Siegel,
2010 WL 3000746, at *5-6. The court noted that it was “unusual” to dismiss an indemnity claim prior to adjudicating the merits, and acknowledged that it had to make “factual determinations ... that would ordinarily be squarely within the province of the jury.”
Id.
at *5. Nonetheless, the court dismissed the indemnity claim because Dynamic “simply [could not] meet its burden.”
Id.
at *6.
With Burner Systems dismissed from the litigation, the parties proceeded on Sie-gel’s direct claim. While considering the parties’ pretrial motions, the district court granted Dynamic’s motion to exclude Crane from testifying as an expert. The court reasoned that Crane’s methodology was unreliable and his findings were “a string” of “speculation.” With Crane excluded, Dynamic moved for summary judgment, which was denied because there was a material issue of fact whether the accident was caused by an intrinsic defect in the range or an external factor. The court cautioned that Siegel had a “small eviden-tiary needle to thread,” but determined that “she ha[d] the right to attempt it nonetheless.”
Before trial, the district court made three additional evidentiary rulings that are challenged on appeal. Dynamic first moved to exclude evidence regarding the cost of potential alternative regulators. The district court withheld ruling pending evaluation of the evidence at trial, and cautioned that evidence of cost would only be allowed if Siegel could show that Dynamic had considered other regulators, and that the regulator was a probable cause of the accident. The district court also withheld ruling on Dynamic’s motion to exclude evidence regarding safety improvements to later model ranges. Finally, Dynamic moved to exclude evidence of other incidents, claims, and lawsuits. The district court granted this motion, noting that there was no record evidence showing that the other incidents were “substantially similar” to the one in this case. In addition, the court noted that the prejudicial effect of this evidence would outweigh its probative value.
Two days prior to trial, Dynamic produced a list of approximately a dozen warranty claims where individuals complained that the kickplate on the outside of the range was abnormally hot. This type of incident would have been consistent with Crane’s theory of defective oven design. The court determined initially that there was not enough evidence to warrant the admission of these prior warranty claims because they were not “substantially similar.”
The district court revisited the issue before the second day of trial. After hearing arguments on both sides, the court amended its prior ruling and held that plaintiff could introduce Crane’s findings regarding the temperature measurements he made and whether the range met applicable
American National Standards Institute (ANSI) standards regarding maximum temperature exposures. However, the court again determined that Crane’s other testimony was unreliable and inadmissible, including his assertions that high temperatures contributed to the degradation of the regulator on his exemplar range and that degradation of the regulator was the cause of the explosion. The court also reaffirmed its exclusion of the prior warranty claims, reiterating that Siegel had still not established that they represented “substantially similar” incidents.
At trial, Siegel proceeded under two liability theories: (1) a design defect in the oven and (2) a manufacturing defect in the regulator. Siegel called two experts from Donan Engineering to testify. Siegel also presented Crane’s testimony regarding ANSI standards and the higher temperatures recorded in his exemplar range. Dynamic moved for judgment as a matter of law under Federal Rule of Civil Procedure 50, which the district court granted. Sie-gel’s claims against Dynamic were dismissed with prejudice.
Siegel v. Dynamic Cooking Sys., Inc.,
No. 3:08CV-00429-S, 2010 WL 4884220, *11 (W.D.Ky. Nov. 24, 2010). The court held that there was not sufficient evidence to support the design defect claim and that Siegel could not rely on circumstantial evidence to prove her manufacturing defect claim, reasoning that Siegel failed to isolate the probable cause of the explosion' and could not succeed under Kentucky products liability law.
II
Dynamic was not entitled to a directed verdict on Siegel’s product liability claim. The court wrongly held that because Sie-gel was attempting to prove her manufacturing defect claim through circumstantial evidence, she had to eliminate all other causes of her injury, and that Siegel undermined this claim by introducing evidence of a design defect.
The court erred because a plaintiff like Siegel can prove a product liability claim using the fact of the malfunction if she eliminates those causes for which the manufacturer would not be liable.
Perkins v, Trailco Mfg. and Sales Co.,
618 S.W.2d 855, 857-58 (Ky.1981). Siegel has done so because Dynamic is liable for her injuries regardless of which of her theories the jury adopts. Under Kentucky law, a manufacturer is liable if it has placed into the market a product that is in a “defective condition” and is “unreasonably dangerous” to the user or consumer or to his property.
See Turpin v. Stanley Schulze and Co., Inc.,
No. 2008-CA-298-MR, 2009 WL 875218, at *7 (Ky.App. Apr. 3, 2009) (citing
Nichols v. Union Underwear Co.,
602 S.W.2d 429, 431 (Ky.1980)). “The product either is or is not unreasonably dangerous to a person who should be expected to use or be exposed to it. If it is, it can make no difference whether it is dangerous by design or by accident.”
Ulrich v. Kasco Abrasives Co.,
532 S.W.2d 197, 200 (Ky.1976). A plaintiff is not required to show precisely
how
a product is defective, but simply must show
whether
it was defective.
See Kentucky Farm Bureau Mut. Ins. v. General Elec. Co.,
No. 3:09-CV-915-JHM, 2011 WL 195615, at *3 (WD.Ky. Jan. 20, 2011).
For example, in
General Electric,
the plaintiffs’ dryer caught fire and burned down their house.
Id.
The plaintiffs’ expert eliminated all external causes of the fire and identified three potential causes internal to the dryer. The defendant argued that because plaintiffs could not winnow down the three potential causes, the balance had not tipped from possible to probable. The
General Electric
court rejected the argument because the probable cause of the fire was the dryer, irrespec
tive of which dryer component failed. The court concluded that “[t]he jury is not left to speculate as to multiple theories of liability because each of the potential causes is attributable to [defendant].”
Id.
at *4. Two other district courts applying Kentucky law have reached similar conclusions.
See Kentucky Farm Bureau Mut. Ins. v. Deere & Co.,
No. 1:07-CV-8, 2008 WL 339622 (W.D.Ky. Feb. 6, 2008);
Kentucky Farm Bureau Mut. Ins. v. Hitachi Home Elec.,
No. 3:08-30-DCR, 2009 WL 2760956 (E.D.Ky. Aug. 26, 2009).
Viewing the evidence in the light most favorable to Siegel, she identified the probable cause of the fire: the Dynamic range. Siegel heard a strange clicking sound coming from the oven, opened the door, and a fireball escaped. Siegel eliminated potential external causes for the explosion: the oven was properly installed, properly used, and properly serviced. Although the experts on which Siegel relies could not identify the precise internal cause of the explosion, they identified the Dynamic range as the probable cause of the explosion. This circumstantial evidence was enough to allow a jury to conclude that Dynamic was liable.
The Restatement (Third) of Torts supports this conclusion. The Restatement allows a jury to impose product liability if the incident “was of a kind that ordinarily occurs as a result of product defect” and was not “solely the result of causes other than product defect.” Restatement (Third) of Torts: Prod. Liability § 3. The court below required Siegel to prove either a design defect or a manufacturing defect. The Restatement, however, provides an example contradicting this requirement:
Mary purchased a new automobile. She drove the car 1,000 miles without incident. One day she stopped the car at a red light and leaned back to rest until the light changed. Suddenly the seat collapsed backward, causing Mary to hit the accelerator and the car to shoot out into oncoming traffic and collide with another car. Mary suffered harm in the ensuing collision. As a result of the collision, Mary’s car was set afire, destroying the seat assembly. The incident resulting in the harm is of a kind that ordinarily occurs as a result of product defect. Mary need not establish whether the seat assembly contained a manufacturing defect or a design defect.
Id.
illus. 2. Siegel’s oven, like Mary’s car, was used without incident for a few years. During ordinary use, the oven suddenly exploded and destroyed the defective components. Under the logic expressed in the Restatement, a plaintiff like Siegel need not establish whether the oven contained a manufacturing defect or design defect.
Our decision in
In re Beverly Hills Fire Lit.,
695 F.2d 207, 218 (6th Cir.1982), does not require a different conclusion. The
Beverly Hills
court explained Kentucky’s rule regarding evidence sufficient to survive a directed verdict:
“[LJegal causation may be established by circumstantial evidence. While reasonable inferences are permissible, a jury verdict must be based on something other than speculation, supposition or surmise. The type of evidence that will support a reasonable inference must indicate the
probable
as distinguished from a
possible
cause.... There must be sufficient proof to tilt the balance from possibility to probability.”
Id.
(quoting
Huffman v. S.S. Mary & Elizabeth Hosp.,
475 S.W.2d 631, 633 (Ky.1972)). Based on this language, Dynamic argues that a plaintiff must show which defect inside a product caused the alleged injuries. Dynamic ignores, however, a key portion of the Kentucky rule:
Where an incident could result from more than one cause, plaintiff tips the
balance from possibility to probability only by ruling out other theories of causation: “[W]here an injury may as reasonably be attributed to a cause that will excuse the defendant as to a cause that will subject it to liability, no recovery can be had.”
Id.
(quoting
Sutton’s Adm’r v. Louisville & N.R Co.,
168 Ky. 81, 181 S.W. 938, 940 (1916)). As this language makes clear, isolating the cause of the injury is required where the alternative cause is one that “will excuse the defendant” from “liability.”
Id.
That is not the case here. Siegel presented two potential causes for her injury, both of which place responsibility on Dynamic. It is therefore necessary to reverse the district court’s judgment in favor of Dynamic.
Ill
Siegel raises several evidentiary issues as alternative bases for reversal. We reject these arguments because the district court did not abuse its discretion in making the challenged determinations, for the reasons given below. On retrial, however, the district court will not be precluded from revisiting these determinations within the scope of its discretion.
A. Exclusion of Thomas Crane’s Testimony
First, the district court did not abuse its discretion by excluding the expert testimony of Thomas Crane. A trial judge’s discretion to exclude expert testimony is guided by Federal Rule of Evidence 702
and
Daubert v. Merrell Dow Pharmaceuticals, Inc.,
509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). The trial court acts as a “gatekeeper” that ensures that “any and all scientific testimony or evidence admitted is not only relevant, but reliable.”
Pluck v. BP Oil Pipeline Co.,
640 F.3d 671, 677 (6th Cir.2011). In this role, the trial court considers “whether the reasoning or methodology underlying the testimony is scientifically valid.”
Id. Daubert
sets forth a nonexclusive list of factors relevant to this inquiry: “(1) whether the theory or technique can be or has been tested; (2) whether it ‘has been subjected to peer review and publication’; (3) whether there is a ‘known or potential rate of error’; and (4) whether the theory or technique enjoys general acceptance in the relevant scientific community.”
Id.
(citing
Daubert,
509 U.S. at 593-94, 113 S.Ct. 2786). We have recognized a fifth common factor: whether the proposed testimony grows of independent research or if the opinions were developed “expressly for the purposes of testifying.”
Smelser v. Norfolk S. Ry. Co.,
105 F.3d 299, 303 (6th Cir.1997),
(abrogated on other grounds by Morales v. American Honda Motor Co.,
151 F.3d 500 (6th Cir.1998)).
Crane agreed with the other experts that the gas leak originated in the range’s regulator and that the regulator was too damaged to reveal whether the leak was caused by a design or manufacturing defect. With this in mind, Crane acquired and tested an “exemplar” Dynamic range of the same model. This range, fueled by natural gas, had been used daily for approximately five years. Siegel’s range, by comparison, was fueled by propane and had been used infrequently for 13 months.
Crane first tested the exemplar range using natural gas, and determined that
“higher than anticipated temperatures were generated” near the regulator, though he did not specify what temperatures were recorded, or what temperatures were anticipated. Crane then examined the regulator under a microscope, observing “a deteriorated surface of the diaphragm and irregular crazing and ‘pock marks’ which would be consistent with the deterioration of the Buna rubber material due to high temperatures.”
Crane converted the exemplar range from natural gas to propane gas and resumed testing. During the subsequent testing, the temperature near the range’s regulator reached 259 degrees Fahrenheit, even though the regulator was designed to operate at temperatures no higher than 225 degrees. Crane also tested a newer model Dynamic range, which had an added piece of sheet metal separating the oven combustion chamber from the regulator. The temperatures around the regulator in the newer range did not exceed 225 degrees. Crane determined it was “reasonable to conclude,” based on these tests, “that temperature degradation of the perimeter diaphragm seal contributed to the leak which created the fugitive gas.”
The district court did not abuse its discretion by excluding this testimony because the court appropriately applied the requisite factors. The district court noted that Crane’s report was “detailed and clear,” but determined that Crane’s opinions were “founded largely on speculation.” The court identified a number of problems with Crane’s methodology that undermined the reliability of his conclusions: Crane had not (1) addressed the differences in historical usage between the exemplar range and Siegel’s range; (2) analyzed the impact of the use of two different fuel types; (3) tested or examined multiple exemplar ranges; (4) gathered historical data regarding leaking regulators; or (5) addressed why the regulator in the exemplar range had not failed during five years of use or during testing. The district court concluded that Crane’s opinion “contains not just one speculation but a string of them.” Therefore, although the hypothesis was plausible, the testing was not reliable enough to warrant admission under the Federal Rules of Evidence.
Underlying the district court’s decision were the obvious differences between the exemplar range and the Siegel range. The exemplar range was used ten times more than Siegel’s oven, was fueled by a different gas, and the regulator did not fail. These differences undermined the reliability of Crane’s experimentation. As noted above, the
Daubert
inquiry is “directed at the science and methodology behind the witness’s testimony.”
Legg v. Chopra,
286 F.3d 286, 291 (6th Cir.2002). The court did not balance expert opinions, evaluate the correctness of any of the data, or determine whether Crane’s opinion was plausible. Instead, the district court focused on the inadequacies of Crane’s methodology, a determination properly within the ambit of
Daubert.
Finally, Siegel takes issue with the trial court’s decision to take “judicial notice” of the fact that propane is used at a higher pressure gas than natural gas. However, we have approved taking judicial notice of commonly known and widely accepted scientific principles. For example, in
Fowler v. Tennessee Val. Auth.,
321 F.2d 566, 570 (6th Cir.1963), we held that “[i]t is common knowledge that, if an electric power line is grounded by a conductor, electric current will flow through the conductor.” Taking notice of the relative pressure exerted by propane verses natural gas is a similarly commonly known and accepted scientific principle. This ruling by the district court was also not an abuse of discretion.
B. Other Evidentiary Issues
The district court also did not, as Siegel claims, abuse its discretion by excluding evidence of prior accidents, which may only be introduced when such accidents are “substantially similar” to the one at issue.
Rye v. Black & Decker Mfg. Co.,
889 F.2d 100, 102 (6th Cir.1989). “Substantial similarity means that the accidents must have occurred under similar circumstances or share the same cause.”
Id.
Siegel sought to introduce evidence of other claims involving gas leaks from the same model range. Siegel argued that these incidents showed that gas leakage was a pervasive problem in Dynamic ranges, but offered no evidence that these leaks occurred under conditions substantially similar to this case. She has not shown, for example, that leaks in the prior incidents originated in the regulator. Sie-gel argues that Dynamic “did nothing to rule out the regulator as the source of the leak in these other incidents.” But Siegel, not Dynamic, has the burden to show substantial similarity.
Similarly, it was within the court’s discretion to deny Siegel’s attempt to admit customer complaints about heat emanating from the oven kickplate. Siegel argues that these customer complaints are consistent with Crane’s theory that the oven was defectively designed. The district court agreed and admitted a portion of Crane’s testimony regarding the kick-plate based on this evidence. However, the district court did not abuse its discretion by excluding evidence of heat complaints because, once again, Siegel did not show similarity—that is, that heat from the kiekplates in the other ovens caused a gas leakage or any other damage.
The district court likewise did not abuse its discretion in excluding evidence regarding design changes that Dynamic made to later model ranges or the cost of such changes. Under Rule 407 of the Federal Rules of Evidence, subsequent design changes are not admissible to prove a design defect, but may be admitted to prove the feasibility of precautionary measures, if disputed.
See Bauman v. Volkswagenwerk Aktiengesellschaft,
621 F.2d 280, 233 (6th Cir.1980). Dynamic conceded the feasability of remedial measures; therefore, the evidence was inadmissible.
IV
As for Dynamic’s common-law indemnity claim, Burner Systems was entitled to summary judgment because there was not sufficient evidence for a jury to conclude on a more-probable-than-not basis that the oven explosion was caused by a manufacturing defect in the regulator. Common-law indemnity has survived Kentucky’s passage of a comparative negligence scheme, and is appropriate where a party is secondarily liable for the acts of another.
See Degener v. Hall Contr. Corp.,
27 S.W.3d 775, 780-81 (2000). To prevail on such an indemnity claim, Dynamic had to show that a manufacturing defect in the regulator existed at the time of sale and that the defect was a probable cause of the injury.
See Calhoun v. Honda Motor Company, Ltd.,
738 F.2d 126, 130 (6th Cir.1984). While causation can be proved “either by direct or circumstantial evidence,”
id.,
reliance on circumstantial evidence “must be sufficient to tilt the balance from possibility to probability.”
Perkins,
613 S.W.2d at 857 (internal quotation marks omitted). Moreover, “where an injury may as reasonably be attributed to a cause that will excuse the defendant as to a cause that -will subject it to liability, no recovery can be had.”
Sutton’s Adm’r,
181 S.W. at 940.
For example, in
Midwestern V.W. Corp. v. Ringley,
503 S.W.2d 745 (Ky.1973), the plaintiff sued the manufacturer of her Volkswagen after she crashed into a telephone pole. The plaintiff claimed that a manufacturing defect in the brake drum caused the car to swerve into the pole. At trial, the plaintiffs witnesses testified that a number of factors could have caused the car to swerve — a brake-drum defect, dirt or water in the brake lining, improper adjustment of the drum, or improper tire pressure. Because the evidence was equivocal, the court reversed a jury verdict in favor of the plaintiff.
Id.
at 748. The court reached a similar conclusion in
Briner v. General Motors Corp.,
461 S.W.2d 99 (Ky.1970). The
Briner
plaintiff sued her car’s manufacturer, alleging that vibrations in her automobile had caused a collision. Her experts offered a variety of possible causes of the vibrations, only some of which were attributable to the manufacturer. The court affirmed a directed verdict in favor of the manufacturer because the plaintiff could not show that the manufacturer’s conduct was anything more than a possible cause of the accident.
Id.
at 101.
Dynamic’s expert — like the experts in
Ringley
and
Briner
— failed to “tilt the balance” because he could not isolate the regulator as the probable cause of Siegel’s injury. That expert, Mark Mulcahy, prepared two reports for Dynamic in anticipation of litigation. In the first report, Mul-cahy concluded that gas leakage occurred at the regulator as a result of the failure of the internal rubber diaphragm. According to Mulcahy, this failure was the result of a tear caused either during the manufacturing process or by excessive gas pressure. Mulcahy found that there was “[ijnsuffi-cient evidence” to “assign[] engineering probability to either scenario.” R. 88-7, Mulcahy Report, at 4. In a subsequent report — prepared in response to Crane’s expert report — Mulcahy modified his opinion and agreed that the probable cause of the explosion was a failure of the perimeter diaphragm. R. 83-8, Mulcahy Supp. Report, at 4. However, Mulcahy was still unable to determine the probable cause of the perimeter diaphragm failure. Instead, Mulcahy presented three different scenarios that could have caused the failure, only one of which was a manufacturing defect attributable to Burner Systems. This is so because “[djamage to the regulator was sufficient ... to preclude identifying a specific scenario for development of leakage at the onset of the event.”
Id.
at 6.
Mulcahy’s later testimony confirmed that the regulator was not defectively designed and that the potential leak scenarios listed in his supplemental report were only possible (not probable) causes of the leak. Nowhere did he opine as to the probable cause of the failure of the perimeter diaphragm.
Summary judgment on the indemnity claim before resolution of the direct claim, moreover, was not an abuse of discretion. Dynamic takes out of context the Kentucky Supreme Court’s statement that “[indemnity is not an issue until fault has been determined.”
See Clark v. Hauck Mfg. Co.,
910 S.W.2d 247, 253 (Ky.1995). That case merely stands for the proposition that “[tjhere can be no indemnity without liability,” and does not purport to preclude categorically the dismissal of an indemnity claim before primary fault has been determined.
Id.
Dynamic correctly notes that, in practice, the liability determination generally precedes a determination on indemnity. Of course, this procedure prevents a premature assignment of indemnity liability when there is no actual liability.
Id.
However, in some instances, summary judgment may be preferable, especially where the third party defendant has less of a connection to the underlying
action. For example, if Tom hit Richard with a car, Richard could sue Tom. If Tom then sued Harry seeking indemnity, a pretrial determination might be appropriate if Harry had absolutely no connection to Tom, the car, or the accident. The court should not force Harry to continue to be a party to the suit where he has no discernible liability for the claim.
In the instant case, there is no evidence that the pretrial dismissal of the indemnity claim was an abuse of discretion. Discovery had been completed, including all expert opinions and reports, and the parties were fully briefed on the summary judgment motions. Most importantly, Dynamic had not produced sufficient evidence to carry its burden on its indemnity claim. As discussed above, no reasonable jury could have found that a manufacturing defect was a probable, rather than a possible, cause of the accident. Accordingly, dismissing the indemnity claim against Burner was not an abuse of discretion.
This result is not fundamentally unfair, as argued by Dynamic, even though Crane was later excluded as an expert. Dynamic relies on the district court’s later statement that “had Mr. Crane’s testimony been excluded prior to the court’s consideration of Burner Systems’ motion, there would have been ... no risk that a jury would have been asked to engage in impermissible speculation about which company’s actions caused the regulator to fail.” Notwithstanding this statement, Dynamic’s argument fails because Crane’s testimony was largely irrelevant to the disposition of the indemnity issue. Dynamic bore the burden of proving that a manufacturing defect was a probable cause of Siegel’s injuries. Dynamic’s own expert could not identify a probable cause. Even if Crane’s testimony had been excluded at the summary judgment stage, the exclusion would not have cured Dynamic’s proof problem. The exclusion of Crane’s opinion is not relevant; what is dispositive was Mul-cahy’s inability to identify the regulator as a probable cause.
This conclusion is not in tension with reversal of the directed verdict against Siegel. There was sufficient evidence to get to the jury on the claim against Dynamic because testimony established some defect in the oven as a probable cause. But there was no such testimony to establish component manufacturer liability. Without sufficient evidence, summary judgment was warranted in favor of the component manufacturer.
Our analysis is not affected by the provision in Kentucky’s comparative negligence statute, Ky. Rev. Stat. § 411.182, which provides for direct recovery by the plaintiff, on an apportioned basis, against multiple parties, including third party defendants, whose culpability each contributed to an injury. In Kentucky, “apportionment of liability arose from statutory provisions permitting contribution and several liability among joint tortfeasors
in pari delicto.” See Degener,
27 S.W.3d at 780. The right of apportionment and contribution only “ ‘exists where there [is] concurring or combined negligence causing the same injury.’ ”
Hall v. MLS Nat. Medical Eval., Inc.,
No. 05-185-JBC, 2007 WL 1385943 (E.D.Ky. May 8, 2007) (quoting
Elpers v. Kimbel,
366 S.W.2d 157, 161 (Ky.1963)). “Because contribution claims often require the apportionment of causation between or among joint tortfeasors on a pro rata basis, the right of contribution has been discussed by at least some Kentucky courts as contribution and apportionment.”
Id.
(internal quotation marks omitted). Such liability does not appear to have been pleaded or argued. In any event, apportioned liability still requires proof of the same elements of liability against the impleaded party, including cau
sation in a product liability case. As the Kentucky Supreme Court has stated, “[f]ault may not be properly allocated to a party, a dismissed party or [a] settling nonparty unless the court or the jury first finds that the party was at fault; otherwise, the party has no fault to allocate.”
Owens Corning Fiberglas Corp. v. Parrish,
58 S.W.3d 467, 471 n. 5 (Ky.2001). As discussed above, Dynamic failed to bring forth sufficient evidence for a jury to find that Burner manufactured a defective part.
Y
Because Burner Systems prevailed against Dynamic’s common-law indemnity claim, a decision we now affirm, the district court properly taxed costs against Dynamic under Federal Rule of Civil Procedure 54(d)(1).
VI
The district court’s decision is affirmed in part, reversed in part, and remanded for further proceedings.