OPINION

JULIA SMITH GIBBONS, Circuit Judge.

Gordon Auto Body Parts ("Gordon"), a Taiwanese company, was one of several early entrants into the United States market for replacement truck hoods. For a while, it faced competition solely from a handful of other Taiwanese manufacturers, at least some of whom participated in joint ventures with Gordon. Superior Production Partnership (trade name "PBSI") eventually entered the market for certain replacement hoods but found that it could not match the prices of Gordon and the Taiwanese firms. Believing that Gordon and the other firms were conspiring to drive it out of business with predatory prices, PBSI brought antitrust claims against Gordon.

The district court granted Gordon's motion for summary judgment after ruling on several discovery disputes. PBSI appeals both the resolution of the discovery disputes and the grant of summary judgment. Because PBSI has failed to make any showing that Gordon's prices were below an appropriate measure of cost, we affirm.

I.

Gordon and a handful of other Taiwanese manufacturers produce an array of replacement parts for automobiles that have been in collisions. Before the mid-2000s, they were the only producers of replacement hoods for certain models of trucks in the American market. Gordon and at least two of the other Taiwanese manufacturers had formal agreements or joint ventures related to shared investments in plant equipment and production of parts.

PBSI started making replacement bumpers as early as the 1980s. In the early 2000s, its management decided to expand into the market for replacement hoods, eventually beginning production on its first hood in 2006. The company soon began manufacturing replacement hoods for four different trucks. Sometime before or shortly after PBSI's entry into these markets, Gordon and other Taiwanese manufacturers reduced their prices on the same hoods. Though PBSI had planned to expand into production of other replacement parts, it found the price competition in these first four markets sufficiently discouraging to nix those plans. PBSI's management now claims that it may be forced to exit the replacement hood market altogether.

Believing its competitors' price changes to be illegal, PBSI filed a suit in federal district court with several antitrust claims, including violations of Sections 1 and 2 of the Sherman Act and Section 13(a) of the Clayton Act as amended by the Robinson-Patman Act. The core allegations of the complaint described a predatory pricing scheme among Gordon and other Taiwanese manufacturers to force PBSI to cease production of truck hoods. The complaint defined the relevant markets as those for the four replacement truck hood models in the United States.

After discovery commenced, PBSI served requests for production on Gordon. Of particular relevance here, PBSI sought (i) "[a]ll documents and communications relating to the cost of production and manufacture of" the four types of hoods for several years in each market and (ii) "[a]ll documents which evidence, relate to, or refer to communications between or among Gordon" and other Taiwanese manufacturers or "any agreement regarding pricing, sharing or use of tooling or refraining from competition" in the United States or elsewhere. Gordon responded to the cost documentation request with what were apparently summaries of its manufacturing costs and contracts with suppliers; PBSI rejected this production as inadequate to determine Gordon's cost-calculation methods and verify its actual costs incurred. PBSI therefore moved to compel production of all cost documentation — including invoices and accounts payable records — underlying the summaries. Gordon objected that this scale of production would be unduly burdensome.

Gordon did produce tooling1 and production agreements among the Taiwanese firms covering two of the four truck hoods but objected to the requests for agreements covering other parts as irrelevant. Gordon claimed that any such agreements did not concern the "relevant markets." Gordon also insisted that the remaining agreements "only involve[d] the equal sharing of the cost to produce the manufacturing tool/die for [other parts], and do not involve pricing of the part[s]." PBSI moved to compel production of these agreements as well.

The magistrate judge granted PBSI's motion in part. As to cost documentation, the judge agreed that PBSI was "entitled to do some comparison of the documents from which the summaries ha[d] been drawn ... to identify the frequency with which the summaries do not reflect information correctly[,]" but he also expressed concern at the burden of producing "every single piece of paper generated by Gordon over the relevant time period which relates to the costs of the products at issue." Faced with significant uncertainty about the relative merits of these claims, the judge settled on an "auditing" procedure by which the parties would agree on a subset of the underlying cost data necessary to "allow [PBSI] to verify or dispute the accuracy of the summary information[.]" If the "results ... reveal[ed] substantial discrepancies between the summary information and the backup documents," the judge would be willing to compel further production. As to the agreements among the Taiwanese competitors, the magistrate essentially agreed with Gordon that any agreements relating to joint tooling or die production were irrelevant to predatory pricing of the truck hoods, but the judge admitted that he "c[ould not] tell" whether any parts "production agreements" outside of tooling arrangements, namely, "pricing agreements and [other] agreements which might limit competition," had been turned over. The magistrate "assume[d] either that any such agreements [had] already been produced, or that they do not exist."

With respect to the cost data, there is no indication in the record that PBSI objected to this order or asked for reconsideration by the district court until approximately 18 months later when it filed a second, separate motion to compel. In that motion, PBSI reasserted its position that Gordon should produce "all [b]ackup [c]ost [d]ocumentation." By this time, Gordon had produced cost data for five fiscal quarters under an agreement with PBSI negotiated pursuant to the magistrate judge's initial decision. When the magistrate judge denied this second motion to compel, PBSI filed an objection with the district court. The district court construed the second motion to compel as an attack on the conclusion of the magistrate's order denying in part the first motion to compel, found the "auditing procedure" in the magistrate's first order to be a "reasonable compromise between [the parties'] competing interests," and affirmed denial of the second motion to compel.

With respect to the competitor agreements, PBSI filed a motion to "Compel[] Compliance with, or for Clarification of," the magistrate's order on the initial motion to compel. PBSI specifically sought clarification that Gordon should produce any competitor agreements relating to production of parts other than the four truck hoods, as evidence of a "proclivity" for collusion. The magistrate admitted that the previous order "did not address directly the relevance of production agreements for other parts." But he nonetheless refused to compel production of any agreements for other parts. The judge thought it "extremely unlikely that PBSI can make a convincing argument that such evidence is admissible under either Fed. R. Evid. 403 or 406, so that the request is not reasonably likely to lead to the discovery of relevant evidence based" on PBSI's "proclivity" theory. PBSI filed a timely objection with the district court, but the district court affirmed the decision for essentially the same reasons articulated by the magistrate.

PBSI and Gordon also submitted expert reports from economists Henry McFarland and Robert Hall, respectively. Both experts submitted second reports addressing each other's initial findings. We explain these reports in more detail below. Each party moved to exclude the expert testimony of the other. The district court denied PBSI's motion to exclude Hall's reports, but granted in part Gordon's motion to exclude McFarland's reports. The court specifically excluded McFarland's opinions on "predatory intent" and recoupment, the latter of which, in the court's view, was intertwined with McFarland's comparison of prices and costs.

Gordon then moved for summary judgment on all antitrust claims, claiming there was no genuine dispute of material fact on either the issue of below-cost pricing or recoupment of losses. In the same opinion in which it resolved the motions to exclude expert testimony, the district court granted Gordon summary judgment. The court stated that its conclusion rested on a lack of evidence that Gordon could recoup its costs and that it would "not analyze and resolve whether PBSI presents evidence... [of] below-cost pricing[.]" However, as with the motions to exclude, the district court essentially subsumed its analysis of below-cost pricing into its examination of recoupment.

PBSI timely appealed the grant of summary judgment as well as the district court's rulings on discovery and the expert reports.

II.

We begin with the framework of antitrust law. Sections 1 and 2 of the Sherman Act, 15 U.S.C. §§ 1-2, address "concerted" and "independent" action, respectively. Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 767-68, 104 S.Ct. 2731, 81 L.Ed.2d 628 (1984). Because Congress perceived different risks from multilateral and unilateral conduct, it set different standards for liability under each section. Id. at 768-69, 104 S.Ct. 2731.

A.

The law on predatory pricing claims under § 2 is significantly clearer, so we begin there.2 Section 2 prohibits three types of actions: monopolization, attempted monopolization, and a conspiracy to monopolize. See 15 U.S.C. § 2. A monopolization claim has two elements: "(1) the possession of monopoly power in a relevant market; and (2) the willful acquisition, maintenance, or use of that power by anti-competitive or exclusionary means as opposed to `growth or development resulting from a superior product, business acumen, or historic accident.'" Conwood Co. v. U.S. Tobacco Co., 290 F.3d 768, 782 (6th Cir. 2002) (quoting Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 595-96, 105 S.Ct. 2847, 86 L.Ed.2d 467 (1985)). "An attempted monopolization [under § 2] occurs when a competitor, with a dangerous probability of success, engages in anti-competitive practices the specific design of which are, to build a monopoly or exclude or destroy competition." Id. at 782, 290 F.3d 768 (alteration in original) (quoting Smith v. N. Mich. Hosps., Inc., 703 F.2d 942, 954 (6th Cir. 1983)). Conspiracy to monopolize entails "proof of concerted activity," but, like the other two § 2 claims, requires an initial identification of the relevant markets.3 Potters Med. Ctr. v. City Hosp. Ass'n, 800 F.2d 568, 574 (6th Cir.1986). Finally, "for a completed monopolization claim to succeed, the plaintiff must prove a general intent on the part of the monopolist to exclude; while by contrast, to prevail on a mere attempt claim, the plaintiff must prove a specific intent to destroy competition or build a monopoly." Conwood Co., 290 F.3d at 782 (internal quotation marks omitted). But since "`no monopolist monopolizes unconscious of what he is doing... [i]mproper exclusion ... is always deliberately intended.'" Id. (alteration in original) (quoting Aspen Skiing Co., 472 U.S. at 602-03, 105 S.Ct. 2847).

Helpfully, the Supreme Court has operationalized § 2 in this context to focus on the economic "essence" of predatory pricing: that "[a] business rival has priced its products in an unfair manner with an object to eliminate or retard competition and thereby gain and exercise control over prices in the relevant market." Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222, 113 S.Ct. 2578, 125 L.Ed.2d 168 (1993). The Court established two prerequisites to liability. See id. at 222-24, 113 S.Ct. 2578. The first of these requires that the defendant's prices be shown to be "below an appropriate measure of [its] costs." Id. at 222, 113 S.Ct. 2578. This circuit applies a burden-shifting framework to determine the "appropriate measure of a rival's cost." Spirit Airlines, Inc. v. Nw. Airlines, Inc., 431 F.3d 917, 938 (6th Cir.2005). Specifically:

Id. (internal quotation marks omitted). Notably, this court has also required that predatory prices be at least below average total cost. Arthur S. Langenderfer, Inc. v. S.E. Johnson Co., 729 F.2d 1050, 1055-56 (6th Cir.1984).

The Supreme Court's second prerequisite to liability is "a demonstration that the competitor had a reasonable prospect, [under the Robinson-Patman Act], or, under § 2 of the Sherman Act, a dangerous probability, of recouping its investment in below-cost prices." Brooke Group Ltd., 509 U.S. at 224, 113 S.Ct. 2578. This recoupment element requires at least two showings. Recoupment first implies that the "below-cost pricing must be capable... of producing the intended effects on the firm's rivals, whether driving them from the market, or ... causing them to raise their prices to supracompetitive levels within a disciplined oligopoly." Id. at 225, 113 S.Ct. 2578. Additionally, even if the target is forced to exit or is disciplined, "[d]etermining whether recoupment of predatory losses is likely requires an estimate of the cost of the alleged predation and a close analysis of both the scheme... and the structure and conditions of the relevant market." Id. at 226, 113 S.Ct. 2578. Most notable among these "structure[s] and conditions" of the market are barriers to entry that protect the monopolistic profits of a successful predator. See Spirit Airlines, Inc., 431 F.3d at 946.

B.

The text of § 1 of the Sherman Act declares illegal "[e]very contract, combination... or conspiracy, in restraint of trade[.]" 15 U.S.C. § 1. Where, as here, the plaintiff relies on circumstantial evidence of illicit collusion, that evidence must "tend[] to exclude the possibility that the alleged conspirators acted independently" — which assumes that the inferences of conspiracy from that evidence are plausible as a matter of economic theory. Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 588, 593-94, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986) (internal quotation marks omitted). In other words, courts are more likely to infer an illegal agreement as the economic plausibility of anticompetitive conduct grows. See In re Publ'n Paper Antitrust Litig., 690 F.3d 51, 62-63 (2d Cir.2012) ("[B]roader inferences [of an agreement] are permitted, and the `tends to exclude' standard is more easily satisfied, when the conspiracy is economically sensible for the alleged conspirators to undertake and the challenged activities could not reasonably be perceived as procompetitive.") (internal quotation marks omitted); In re High Fructose Corn Syrup Antitrust Litig., 295 F.3d 651, 661 (7th Cir.2002) (Posner, J.) ("More evidence is required the less plausible the charge of collusive conduct."). Additionally, the illegal conspiracy must produce antitrust injury, namely, an injury to competition itself. In re Se. Milk Antitrust Litig., 739 F.3d 262, 284 (6th Cir.2014).

Courts have rarely addressed the application of § 1 to the scenario before us. Matsushita, undoubtedly the leading case, declared single-firm predatory pricing generally implausible and a conspiracy to charge predatory prices even more so. 475 U.S. at 589-91, 106 S.Ct. 1348. But Matsushita did not make it entirely clear if § 1 liability required a showing of prices below some "appropriate" level of cost, as in Brooke Group. See Matsushita, 475 U.S. at 584 n. 8, 106 S.Ct. 1348 (noting that, for that particular case, the Court would view the threshold as either "(i) pricing below the level necessary to sell [plaintiff's] products, or (ii) pricing below some appropriate measure of cost" without deciding if those tests applied to future cases). Nor did Matsushita expressly lay out a recoupment requirement.

Nonetheless, we think it best to infer these same elements in a § 1 predatory pricing claim. Matsushita clearly considered both below-cost pricing and the likelihood of recoupment to impact, if not fully determine, the plausibility of a predatory pricing conspiracy. See id. at 590-91, 106 S.Ct. 1348. And whether predation and recoupment are merely considerations that inform the plausibility of inferring an agreement, as PBSI suggests, or are formally independent requirements beyond a § 1 agreement and antitrust injury, as Gordon argues, is a technicality we need not address here. Either way, a plaintiff must grapple with these concepts in proving its case under § 1. See R.J. Reynolds Tobacco Co. v. Cigarettes Cheaper!, 462 F.3d 690, 695-96 (7th Cir.2006) (Easterbrook, J.).

As explained below, neither PBSI's § 1 claim, § 2 claim, nor its Robinson-Patman Act claim can survive summary judgment on the issue of below-cost pricing. We therefore analyze each assignment of error solely for its illumination of the relationship between Gordon's prices for the hoods and its costs of production.

III.

We turn first to the district court's discovery orders, as any significant errors here would offer a clear reason to reverse summary judgment. See, e.g., Tenn. Valley Auth. v. United States, 60 Fed.Cl. 665, 675-76 (Fed.Cl.2004) ("A party that has not had the opportunity to obtain evidence in support of its opposition to a motion for summary judgment should not have that inability held against it."). We review a trial court ruling limiting or denying discovery for an abuse of discretion. Robertson v. Lucas, 753 F.3d 606, 623 (6th Cir.2014); see also Louzon v. Ford Motor Co., 718 F.3d 556, 560 (6th Cir.2013) (reviewing denial of a motion to compel for abuse of discretion). "An abuse of discretion occurs if the district court relies on clearly erroneous findings of fact, applies the wrong legal standard, misapplies the correct legal standard when reaching a conclusion, or makes a clear error of judgment." Louzon, 718 F.3d at 560 (internal quotation marks omitted).

The parties may demand "any nonprivileged material that is relevant to any party's claim or defense," so long as that material "appears reasonably calculated to lead to the discovery of admissible evidence." Fed. R. Civ. P. 26(b)(1). But the trial court may limit discovery where "the burden or expense of the proposed discovery outweighs its likely benefit, considering the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the action, and the importance of the discovery in resolving the issues." Fed. R. Civ. P. 26(b)(2)(C)(iii). In sum, "[a]lthough a plaintiff should not be denied access to information necessary to establish her claim, neither may a plaintiff be permitted to `go fishing' and a trial court retains discretion to determine that a discovery request is too broad and oppressive." Surles ex rel. Johnson v. Greyhound Lines, Inc., 474 F.3d 288, 305 (6th Cir. 2007) (internal quotation marks omitted).

A.

PBSI argues that the district court should have compelled production "of all documents showing Gordon's costs to produce the four hoods at issue, including the cost of materials, labor, and administrative overhead." Gordon does not dispute the relevance of this cost data; it responds instead that PBSI waived this argument below and that the district court reasonably exercised its discretion by requiring limited production of cost documents.

We agree with Gordon that PBSI waived its right to appeal this issue for failure to object to the magistrate judge's order in a timely fashion. On a non-dispositive issue, a party must object to a magistrate's order within 14 days of receiving a copy of the order. Fed. R. Civ. P. 72(a). Failure to do so risks losing the right to appeal the underlying issue to this court. Stemler v. City of Florence, 126 F.3d 856, 866 n. 9 (6th Cir.1997).

Here, the magistrate judge entered an order in December 2008 granting in part PBSI's motion to compel. That order quite clearly set up a scheme for initially limited production of cost documentation based on the reasonable compromises of both parties, but PBSI did not raise this issue again until a year and a half later when it moved to compel production of the same documents. PBSI's later motion might have been permissible had it "alleged failure to comply with the prior order of the Court," or had it been based on materials reviewed as a result of Gordon's compliance with the first order, but that is not the tack it took. Instead — and as the magistrate judge and district court recognized — the second motion to compel effectively sought full reconsideration of the initial order. In the words of its own motion, PBSI argued that "[t]he information provided by Gordon does not allow PBSI [to] fully evaluate every element of Gordon's costs over the relevant time period.... PBSI must be allowed to look at the entire period of time for which Gordon's pricing was extraordinarily low." This is far beyond the limits of the magistrate judge's original decision.

PBSI did timely object to the denial of its second motion to compel. But it would be inappropriate to permit PBSI to sidestep the 2008 order merely by filing a similar motion and forcing the district court to respond anew. See Daley v. Marriott Int'l, Inc., 415 F.3d 889, 892-93, 893 n. 9 (8th Cir.2005) (denying appellate review where plaintiff failed to object to a magistrate's ruling in time but later filed a motion to reconsider); Key Energy Res., Inc. v. Merrill (In re Key Energy Res., Inc.), 230 F.3d 1197, 1200-01, 1201 n. 3 (10th Cir.2000) (finding waiver even where the district court reached the merits of the magistrate judge's ruling on a motion to reconsider filed after the window for objecting).

Failure to object is not a jurisdictional bar on appellate review, however, so we may "excuse the default in the interest of justice[.]" Alspaugh v. McConnell, 643 F.3d 162, 166 (6th Cir.2011) (internal quotation marks omitted). But we do not think PBSI has shown an injustice or abuse of discretion regarding the production of cost data. The magistrate judge faced plausible, competing claims about the need to verify the summary data already produced, and the potential for "extreme hardship" of fully depicting manufacturing costs in four markets across several years. So the magistrate judge sensibly allowed the burden of discovery to escalate only as a greater benefit became apparent. In this case (and as agreed to by the parties), that meant initial production of cost documentation for five fiscal quarters to allow PBSI to "verify or dispute the accuracy of the summary information[.]" If this limited production suggested flaws in Gordon's summaries, PBSI could return to court to compel further production.

Although the magistrate judge and district court referred to this scheme as an "auditing procedure," it might better be described as sampling4 — a discovery method often thought well within the discretion of a trial court. See, e.g., Larson v. AT & T Mobility LLC, 687 F.3d 109, 130 n. 33 (3d Cir.2012) ("[T]he parties may need some focused discovery, which may include sampling of the sources, to learn more about what burdens and costs are involved in accessing the information, what the information consists of, and how valuable it is for the litigation in light of information that can be obtained by exhausting other opportunities for discovery." (alteration in original) (quoting Fed. R. Civ. P. 26(b)(2) advisory committee notes)); see also Fed. R. Civ. P. 34(a) (permitting sampling of documents or electronically stored information). In fact, sampling may have been a permissible approach to discovery of cost information even if Gordon had never produced any cost summaries whatsoever. See Soto v. Castlerock Farming & Transp., Inc., 282 F.R.D. 492, 503-04 (E.D.Cal.2012) (collecting cases approving sampling as a discovery compromise). And even if this sampling did not produce PBSI's desired results ex post, this does not make it an impermissible method for the trial court to manage discovery ex ante.

B.

PBSI also argues that the district court should have compelled production of the competitor agreements relating to automobile parts other than the four truck hoods. PBSI points to deposition testimony of Gordon employees that appears to confirm the existence of production agreements the Taiwanese firms made to establish artificially high prices in other product and geographic markets. But PBSI contends only that these agreements are relevant to recoupment — not the relationship between the costs and prices of the truck hoods. We simply see no indication — and PBSI gives us no reason — that the agreements among Taiwanese firms about addressing joint production of other parts meaningfully influenced Gordon's cost of producing, or revenue from selling, the four truck hoods at the heart of this case. Those agreements that did concern the truck hoods were produced during initial discovery, and, notably, PBSI does not rely on them to defeat summary judgment on below-cost pricing. Consequently, any error in the discovery orders covering these agreements does not warrant remand.

IV.

We turn next to the expert reports. We review the "district court's decision to admit expert testimony, including its `decisions about how to determine reliability' and `its ultimate conclusion,' for abuse of discretion." Surles ex rel. Johnson, 474 F.3d at 293 (quoting Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999)). We have summarized the substantive test for admissibility as follows:

Burgett v. Troy-Bilt LLC, 579 Fed.Appx. 372, 376 (6th Cir.2014) (quoting In re Scrap Metal Antitrust Litig., 527 F.3d 517, 528-29 (6th Cir.2008)). "The Rule 702 inquiry is a flexible one and its focus must be solely on principles and methodology, not on the conclusions they generate." Lee v. Smith & Wesson Corp., 760 F.3d 523, 526 (6th Cir.2014) (quoting Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 595, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)) (internal alterations and quotation marks omitted). Similarly, Rule 702 directs courts to focus on the reliability of expert testimony, rather than the "credibility and accuracy" of that testimony. In re Scrap Metal Antitrust Litig., 527 F.3d 517, 529 (6th Cir.2008).

Before applying these standards, we describe in more detail the expert testimony at issue. Hall's report found, inter alia, that Gordon's prices never fell below average avoidable cost and that the risk of future entry would undermine any attempt to recoup predatory losses in the markets for the hoods. McFarland attempted to calculate average variable costs in his first report but found this "very difficult to measure" given gaps and ambiguities in the data; ultimately, though, he concluded that Gordon's "prices are often above the estimated level of [average variable] costs, but the difference ... is often very small." He found this "aggressive pricing" "disturbing" and potentially suggestive of a predatory scheme. McFarland also addressed the prospects of a predator recouping its losses in his first report. In his analysis, Gordon might recoup its losses either through higher prices for the hoods after PBSI's exit, or "if [the scheme] discouraged PBSI from entering the production of other hoods," Gordon might enjoy "higher sales and prices[.]"

In his second report, McFarland revamped his theory of predation. Rather than focus on average variable or average avoidable cost, he opted instead for what he called the "no economic sense" test, under which the losses from predation are measured as any downward deviation from a profit-maximizing price. McFarland's second analysis also addressed the likelihood that PBSI would be forced out of the four markets for replacement hoods, and it provided more quantitative explanation of Gordon's recoupment prospects, even comparing those to the "losses" under the "no economic sense" test.

A.

As relevant here, the district court excluded as unreliable McFarland's opinion on two issues: (i) predatory intent and (ii) below-cost pricing. PBSI contests both exclusions. We find no abuse of discretion.

1.

McFarland opined in both reports that Gordon's pricing strategy reflected an intent to force PBSI out of the markets for replacement hoods. The district court excluded this part of the report under Rule 702(a) because, in its view, specialized knowledge would not aid the factfinder in determining intent. We see no meaningful error on this particular ruling.

If the district court used "intent" to mean either the "general" intent required for "completed" monopolization or the "specific" intent required for attempted monopolization, see Conwood Co., 290 F.3d at 782, there is certainly no need for expert assistance. As noted above, the Supreme Court has virtually eliminated the substance of this element, or at least folded it into the other elements of a predatory pricing claim.5 See Aspen Skiing Co., 472 U.S. at 602-03, 105 S.Ct. 2847 ("[T]here is agreement on the proposition that no monopolist monopolizes unconscious of what he is doing.... Improper exclusion (exclusion not the result of superior efficiency) is always deliberately intended." (internal quotation marks omitted)). And if the district court used "predatory intent" as shorthand for injury to competition, see, e.g., D.E. Rogers Assocs., Inc. v. Gardner-Denver Co., 718 F.2d 1431, 1439 (6th Cir.1983), that question collapses into the predation and recoupment analyses in McFarland's report. Post-Brooke Group, a firm only acts as a "predator" to the extent that it sells at prices below an appropriate measure of cost and stands to recoup its losses through later supracompetitive profits. 509 U.S. at 222-24, 113 S.Ct. 2578. And as discussed below, McFarland's report has serious infirmities relating to the cost-price relationship.

Thus, no expert testimony is truly needed on the question of intent, at least assuming that all other elements have been established, and the court therefore did not abuse its discretion in barring McFarland from testifying on this narrow question.

2.

The district court's second quarrel with McFarland's report also has merit. The court framed the problem as a flaw in McFarland's recoupment analysis, but we think the court's reasoning is more applicable to below-cost pricing. We conclude that McFarland's opinion does not apply a permissible test for this issue.

In his initial report, McFarland concluded that he could only say that Gordon's "prices are often above the estimated level of costs, but the difference between price and cost[] is often very small." Though he found this proximity "disturbing," he admitted in his second report that Gordon's "aggressive pricing strategy" could not be shown to fall below average variable or average avoidable costs. He therefore analyzed Gordon's short-term losses using what he termed the "no economic sense test," which calculates losses as possible profits foregone by not charging a higher price. Under this test, a predator does not necessarily price below average variable, average avoidable, or average total cost; rather, it simply charges something less than the profit-maximizing price. McFarland then compared his estimate of these foregone profits with his calculations of what Gordon and its alleged conspirators stood to gain by forcing PBSI out of the market for replacement truck hoods to reach his conclusion about the likelihood of recoupment.

The district court and Gordon are correct that admitting an expert opinion on predation relying on either a "disturbing" proximity test or a "no economic sense" test is flatly contrary to law. See Brooke Group Ltd., 509 U.S. at 222-23, 113 S.Ct. 2578 ("[A] plaintiff seeking to establish competitive injury resulting from a rival's low prices must prove that the prices complained of are below an appropriate measure of its rival's costs."). While the appropriate measure of cost has some flexibility (as, say, between average variable and average total cost), see Spirit Airlines, Inc., 431 F.3d at 938, this flexibility extends only to costs directly incurred, not all conceivable opportunity costs.6 See Arthur S. Langenderfer, Inc., 729 F.2d at 1058 (rejecting a "profit maximization" theory of predation similar to that used here); see also United States v. AMR Corp., 335 F.3d 1109, 1118 (10th Cir.2003) ("Courts and scholars have observed that such a sacrifice test would necessarily involve a great deal of speculation and often result in injury to the consumer and a chilling of competition."). Consequently, McFarland's analysis of putative losses from predation is inadmissible, and the district court did not abuse its discretion in rejecting the "no economic sense" test. Nor are prices close to, but above, some measure of cost permissible evidence of anticompetitive conduct. Rather, they are the essence of competition. Brooke Group Ltd., 509 U.S. at 223-24, 113 S.Ct. 2578.

B.

As it did in the district court proceedings, PBSI now contends that Hall's opinion is contrary to law, and therefore inadmissible, because he compared Gordon's prices only to average variable cost. Relying on a few stray remarks in the 200 or so pages of material from Hall, PBSI claims that his entire analysis is tainted by the lack of references to average total cost, another metric for predation. The district court, reading Hall's report as a whole, disagreed. In its view, the remarks identified by PBSI referred to common practice among economists, not lawyers, and, more importantly, did not purport to "find that predatory pricing may never exist if a firm's pricing is above" average variable cost.

The district court was well within its discretion to admit Hall's testimony. We agree that no statement identified by PBSI categorically excludes average total cost as a legal criterion for predation. And since average variable cost is clearly relevant to and consistent with this court's case law on predation, see Spirit Airlines, Inc., 431 F.3d at 938, any supposed risk of confusion to the factfinder would be easily outweighed by the utility of Hall's (concededly reliable) analysis.

V.

Having addressed and approved the procedural and evidentiary rulings by the district court, we now consider the ultimate issue in this case: whether Gordon should have received summary judgment based on the record before the district court.

We review a grant of summary judgment de novo. Payne v. Novartis Pharm. Corp., 767 F.3d 526, 530 (6th Cir.2014). Summary judgment is appropriate where the evidence on record shows 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), (c). The moving party must first demonstrate to the court that an essential element of the nonmoving party's case is absent; upon that showing, the nonmoving party must then present sufficient evidence to demonstrate that there is a factual controversy as to that element, or at least explain why such evidence is not available. See Celotex Corp. v. Catrett, 477 U.S. 317, 322-23, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). We grant all reasonable inferences in favor of the non-moving party. Matsushita, 475 U.S. at 587-88, 106 S.Ct. 1348. But "[t]he mere existence of a scintilla of evidence in support of the plaintiff's position will be insufficient" to defeat summary judgment. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 252, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986).

Our conclusions on the discovery and expert testimony issues go a long way toward resolving PBSI's appeal of the grant of summary judgment. PBSI, without the benefit of McFarland's "no economic sense" analysis, has little to support an essential part of its case. Rather than pointing to evidence in the record, PBSI primarily argues that vague inferences about "predatory intent" or a "scheme's predatory nature" can replace a cost-price comparison under its reading of Spirit Airlines.

While Spirit Airlines did not lay out a cost test that must be used to show predation, 431 F.3d at 938-39, it is untenable to suggest that a plaintiff can skate by without any method for comparing prices to cost. See Arthur S. Langenderfer, Inc., 729 F.2d at 1056; see also R.J. Reynolds Tobacco Co., 462 F.3d at 695-96. Since price cuts are generally pro-competitive, some comparison to cost is necessary to determine injury to competition; without a cost-based test of predation, courts would inevitably punish firms for being the most efficient producers. See Brooke Group Ltd., 509 U.S. at 222-24, 113 S.Ct. 2578. Spirit Airlines is therefore best read to allow plaintiffs the flexibility to apply a cost-based test other than average variable cost (or variations on average variable cost). See 431 F.3d at 945 (finding an issue of material fact given an "intellectual disagreement" over the proper calculation of average variable costs); see also AMR Corp., 335 F.3d at 1116 ("[W]e do not favor [average variable cost] to the exclusion of other proxies for marginal cost. Whatever the proxy used to measure marginal cost, it must be accurate and reliable in the specific circumstances of the case at bar."). Plaintiffs can proffer any number of proxies for the marginal cost of production. See AMR Corp., 335 F.3d at 1115-20 (discussing several alternatives).

But without McFarland's "no economic sense" test, PBSI has not offered such a metric. And even if there is a reasonable inference from the expert reports that Gordon's prices were below average total cost, that inference still leaves the burden of production on PBSI. Spirit Airlines, Inc., 431 F.3d at 938. Thus, PBSI's claims, under both the Sherman Act and the Robinson-Patman Act, succeed or fail based on an analysis of average variable cost. See AMR Corp., 335 F.3d at 1120-21 (affirming summary judgment on a predatory pricing claim where plaintiff's cost metrics were unreliable and no one disputed that prices were above average variable cost). It is not helpful to the plaintiff that Gordon's prices may have a "close proximity" to average variable cost, or that any price cuts merely coincided with competitive entry. So long as prices are still above cost, those actions simply suggest pro-competitive behavior. Brooke Group Ltd., 509 U.S. at 224, 113 S.Ct. 2578.

Finally, PBSI provides only a single record citation, across both of its briefs, to support its assertion that Gordon priced below its variable costs. That citation purports to compare costs discussed in deposition testimony with a table of prices in McFarland's initial report. On our review of that deposition testimony, we simply do not see any cost figures with sufficient specificity to construct a meaningful comparison. The deposition transcript does provide one aggregate cost figure for a single truck hood in 2006. But it does not at all illuminate how costs varied throughout the year or how sales of the hoods were distributed across time — at least one of which is necessary to comparing Gordon's costs to its highly fluctuating prices. To infer from this evidence that Gordon sold below cost, we would be forced to assume that it sold the overwhelming majority of its output only during the few months with the absolutely lowest prices in McFarland's table. Even putting aside the issue of whether a predatory pricing scheme could succeed in such a short period of time, PBSI points to no evidence that Gordon's output was concentrated in this interval. We do not find this assumption reasonable. PBSI's last-gasp citation, then, is no more than a "scintilla" of evidence on an essential element of each of its antitrust claims, and summary judgment was therefore appropriate. Anderson, 477 U.S. at 252, 106 S.Ct. 2505.

VI.

The judgment of the district court is affirmed.