Affirmed by published opinion. Judge DIAZ wrote the opinion, in which Senior Judge DAVIS joined. Judge SHEDD wrote a dissenting opinion.
DIAZ, Circuit Judge:In this Sherman Act case, we review the district court’s grant of summary judgment in favor of Defendant-Appellee E.I. Du Pont de Nemours and Company (“DuPont”). We also consider challenges by Plaintiff-Appellant Kolon Industries Incorporated (“Kolon”) to certain of the district court’s discovery rulings and its denial of Kolon’s recusal motion. Finding no reversible error, we affirm.
I.
A.
The merits of this case concern Kolon’s claim that DuPont attempted to wield, or did wield, monopoly power over the U.S. para-aramid fiber market in violation of Section 2 of the Sherman Act, 15 U.S.C. § 2.1 Para-aramid is a strong, complex *163synthetic fiber used in body armor, tires, fiber optic cables, and a variety of other industrial products. Three para-aramid producers — DuPont, Teijin Aramid (formerly a division of the Dutch company Akzo N.V.), and Kolon — sell their para-aramid fibers to U.S. consumers.
DuPont invented para-aramid fiber in 1965, and for a period controlled the entire U.S. para-aramid market with its Kevlar© fiber. Teijin introduced its competing Twaron© fiber to the U.S. market in 1987 and has chipped away at DuPont’s share of that market every year since 1990. According to one of Kolon’s expert witnesses, during 2006-2009 (the relevant time period), DuPont’s share of the U.S. para-aram-id market (the relevant geographic and product markets) fell from a high of 59% in 2006 to 55% in 2009, with most of this loss going to Teijin.
The U.S. para-aramid market is highly concentrated between DuPont and Teijin, which together account for 99% of U.S. sales. This extreme market concentration owes at least in part to the industry’s high entry barriers. As Kolon showed, para-aramid production is time-intensive and expensive, and potential customers test and “qualify” each para-aramid product to ensure it meets their particular needs, a process that typically takes six months to three years. In addition to this evidence of market concentration and high barriers to entry, Kolon adduced evidence that DuPont, despite Teijin’s encroachment, earned profit margins of as high as 75% between 1997 and 2005 and had the ability to price discriminate among its customers.
Kolon made its foray into the U.S. para-aramid market in 2005 with its Heracron© fiber. Kolon’s evidence showed that DuPont considered Kolon’s market entry to be a threat. Anticipating Kolon’s potential encroachment, DuPont began identifying segments of the market that it viewed as hospitable to entry, including auto short fibers (pulp for brakes and gaskets), tires, manufactured rubber goods, and fiber optic cables. According to Kolon, DuPont then undertook a strategy of executing multi-year supply agreements with high-volume customers in each identified segment, requiring these customers to purchase most or all of their para-aramid requirements from DuPont during the relevant time period.
These supply contracts contained restrictions, such as volume purchase commitments and “meet and release” clauses, that required any competing para-aramid seller to propose a bid at a designated lower amount than DuPont’s existing price, prohibited the customer from informing the competing seller of DuPont’s price, and gave DuPont a right to match any competing offer. As a result of these allegedly anticompetitive practices, Kolon insists, it never achieved more than a de minimis market share during the relevant time period. By contrast, Kolon was able to penetrate other comparable para-aram-id markets, such as Europe.
DuPont, for its part, attributes Kolon’s failure to penetrate the U.S. market to Kolon’s own shortcomings. According to DuPont, Kolon undertook only a “feeble effort” to establish a U.S. foothold, using only seven sales agents, inadequately investing in product offerings and supply capacity, and contacting only a small percentage of potential customers as of October 2009. Meanwhile, DuPont defends its supply agreements as a competitive response to Teijin’s use of such practices, and as driven by consumer demands.
*164DuPont also attempts to diminish the reach of its supply agreements, noting that it entered into only twenty-five agreements with twenty-one U.S. customers, collectively accounting for only a small percentage of its U.S. revenue. Of those agreements, DuPont says, only a portion obligated the customer to purchase some amount of Kevlar, and these had typical durations of two years or shorter. Meanwhile, none of DuPont’s supply agreements precluded competitors from qualifying their products with the customer while the DuPont agreement was in effect. And of the group of “key” customers Kolon identified as necessary to establish a foothold for effective competition, DuPont notes that the majority had no supply agreement with DuPont during the relevant period. Appellee’s Br. at 38. In short, DuPont submits that myriad self-inflicted failures — not DuPont’s supply agreements— frustrated Kolon’s U.S. market penetration.
B.
DuPont brought suit against Kolon alleging the theft and misappropriation of its Kevlar trade secrets (the “trade secrets case”). Kolon’s answer included the instant counterclaim (the “antitrust case”), alleging that DuPont had illegally monopolized and attempted to monopolize the U.S. para-aramid market through its supply agreements with high-volume para-aramid customers. DuPont moved, under Federal Rule of Civil Procedure 12(b)(6), to dismiss Kolon’s counterclaim. The district court granted that motion, with leave to amend. Kolon filed an amended counterclaim, followed by a second amended counterclaim, which was also dismissed for failure to state a claim, again with leave to amend. Kolon declined to further amend the counterclaim, opting instead to appeal the dismissal. We reversed, holding that Kolon had adequately pleaded both its actual and attempted monopoly claims, and remanded the matter to the district court for further proceedings. E.I. du Pont de Nemours & Co. v. Kolon Indus., Inc. (DuPont I), 637 F.3d 435 (4th Cir.2011).
On remand, the district court tried the trade secrets claim separately, culminating in a $919.9 million jury verdict for DuPont on September 14, 2011. The court then formally severed the two claims. The court also issued several rulings adverse to Kolon in the antitrust case. First, the district court denied Kolon’s motions to compel DuPont’s production of transaction-level sales and cost data, concluding that this discovery would significantly burden DuPont and would not be any more useful than the aggregate sales and cost data DuPont had already produced. The district court also denied Kolon’s motions — filed on grounds described below— for recusal and disqualification in both the antitrust and trade secrets cases. The district court further denied Kolon’s request to depose a DuPont corporate representative concerning its strategic use of supply agreements. Finally, the district court granted summary judgment to DuPont on both Sherman Act claims, dismissing them with prejudice. Kolon timely noted this appeal.
II.
Before turning to the merits of the antitrust claims, we first consider Kolon’s argument that the district court judge was required to recuse himself in both the instant antitrust case and the trade secrets case, which is also now before us on appeal. See E.I. du Pont de Nemours & Co. v. Kolon Indus., Inc., No. 12-1260 (argued May 17, 2013).
A.
Kolon’s recusal motion is based on the district court judge’s involvement, while in *165private practice, in litigation that, according to Kolon, presents a matter in controversy in the instant dispute. In the 1980s, DuPont and Akzo N.V., Teijin’s predecessor, became embroiled in several patent lawsuits relating to the manufacture and sale of para-aramid fibers. In one such dispute, Akzo sued DuPont in the United States District Court for the Eastern District of Virginia for infringement of an Akzo para-aramid patent.
In the Akzo case, DuPont was defended by the law firms Fitzpatrick, Celia, Harper & Scinto (“Fitzpatrick Celia”) and McGuire Woods & Battle (now “McGuire-Woods”). In the 1980s and at the time of the Akzo case, the district court judge was a partner at McGuireWoods. As a result, he was a limited partner in an affiliated entity, and continued to receive small payments from McGuireWoods of rent for furnishings. Because McGuireWoods also served as counsel to DuPont in the present litigation, in May 2009 the clerk of court issued a notice informing the parties of the judge’s related financial interest. The judge noted then that he did not believe grounds for disqualification existed, but he instructed the parties to file a motion within 20 days if they believed otherwise. Neither party filed a motion or otherwise objected to the judge’s continued participation in the case.
Kolon believed the Akzo matter was central to both its defense of the trade secrets case and its maintenance of the antitrust case. In the antitrust case, Kolon began seeking discovery of the Akzo case files in August 2009. It contended that the Akzo litigation — which also included a counterclaim filed by DuPont, and which was ultimately resolved pursuant to a settlement that restricted Akzo’s exports to the United States — was relevant evidence of DuPont’s anticompetitive practices and its intent to monopolize the U.S. para-aramid market.
In the trade secrets case, beginning in April 2010, Kolon conducted extensive discovery into the Akzo litigation on the theory that DuPont’s asserted trade secrets had been revealed in the course of that litigation and therefore were no longer secret. In August 2010, the district court ordered McGuireWoods, DuPont, and Fitzpatrick Celia to review the Akzo case files and to produce responsive documents. Later that month, DuPont produced approximately thirty boxes of documents, along with a privilege log, from Fitzpatrick Celia’s files.
One of the entries on the privilege log showed that in May 1985, Mr. Fitzpatrick of Fitzpatrick Celia had sent the district court judge, then a partner at McGuire-Woods, a letter confirming a telephone conversation in which Fitzpatrick had asked the judge to send him a facsimile of the complaint filed by Akzo. The privilege log also indicated that, per Mr. Fitzpatrick’s request, the judge had faxed him a copy of the complaint.
Nearly a year later, on July 20, 2011, Kolon filed a memorandum opposing one of DuPont’s proposed jury instructions in the trade secrets case. In that memorandum, Kolon stated that it was “compelled to point out that there is some question whether Your Honor should be adjudicating these matters [due to DuPont’s production of documents] indicating a role by Your Honor in the earliest stage of the Akzo litigation.” Kolon Indus., Inc. v. E.I. du Pont de Nemours & Co., 846 F.Supp.2d 515, 519 (E.D.Va.2012) (quoting 3:09-cv-58, Docket No. 1247) (emphasis omitted). In a subsequent telephone conference with the court, Kolon’s counsel explained that the source of its concern was the May 1985 letter from Mr. Fitzpatrick to the district court judge.
*166The district court judge then ordered DuPont to produce any documents concerning his involvement in the Akzo litigation, which ultimately comprised only the two documents described above: the May 1985 letter from Mr. Fitzpatrick to the judge requesting a copy of the Akzo Complaint, and the judge’s responsive facsimile cover sheet, with the Complaint attached. Following further inquiry into the matter, Kolon’s counsel represented that they had reviewed all the non-privileged documents from the Fitzpatrick Celia files and that none of those documents contained the judge’s name; DuPont’s counsel made similar representations with respect to the files’ privileged documents.
After reviewing the two relevant documents, the district court judge determined that he had no recollection of the communication with Fitzpatrick or of any involvement in the Akzo litigation. Kolon said nothing more then about the issue.
During these inquiries, the trade secrets trial began as scheduled, and the jury returned a $920 million dollar verdict for DuPont after a seven-week trial. Meanwhile, discovery in the antitrust case was underway following our March 2011 reversal of the district court’s initial dismissal. As in the trade secrets case, the parties came to a head over Kolon’s proposed discovery of the Akzo litigation files, and in September and October of 2011 filed reciprocal motions respectively seeking to compel and protect that information.
Then, on November 30, 2011, two months after the jury verdict in the trade secrets case, and two days before motions for summary judgment were due in the antitrust case, Kolon filed its recusal motion, which the district court denied.2
B.
28 U.S.C. § 455(b)(2), the provision on which Kolon relies, provides that any judge of the United States shall disqualify himself
[wjhere in private practice he served as a lawyer in the matter in controversy, or a lawyer with whom he previously practiced law served during such association as a lawyer concerning the matter, or the judge or such lawyer has been a material witness concerning it.
28 U.S.C. § 455(a), which also has some relevance to our inquiry, provides that “[any judge] of the United States shall disqualify himself in any proceeding in *167which his impartiality might reasonably be questioned.”3
The district court reasoned that although § 455 is itself silent on whether a party seeking recusal must timely file a motion with the court, and despite the mandatory text of § 455(b) (“[Any judge] ... shall disqualify himself....”), the majority of circuits, including this one, have found that § 455 includes a timeliness requirement. Kolon Indus., 846 F.Supp.2d at 522 (citing, inter alia, United States v. Owens, 902 F.2d 1154, 1155 (4th Cir.1990)). Accordingly, because Kolon had delayed in filing its recusal motion for almost a year after it learned of the alleged conflict, the district court denied Kolon’s motion as untimely.
Ruling in the alternative on the merits, the district court concluded that even ignoring the untimeliness of Kolon’s motion, recusal was unnecessary under § 455(b)(2) since the Akzo litigation was not “sufficiently related” to the instant action to “constitute parts of the same matter in controversy.”4 Id. at 528 (quoting United States v. DeTemple, 162 F.3d 279, 286 (4th Cir.1998)).
C.
We review a judge’s recusal decision for abuse of discretion. United States v. Mitchell, 886 F.2d 667, 671 (4th Cir. 1989).
We first consider Kolon’s challenge to the district court’s holding that § 455(b)(2) includes a timely-filing requirement that Kolon failed to satisfy. Kolon maintains that the district court erred by relying on Owens, in which we said that “[t]imeliness is an essential element of a recusal motion,” and that notwithstanding the absence of an explicit timely-filing requirement in § 455, such a requirement is “judicially implied.” 902 F.2d at 1155.
This language, Kolon submits, does not control here because it speaks only “broadly about section 455 and did not specify whether this requirement should apply to both subsections (a) and (b) or solely to [subjsection 455(a).” Appellant’s Br. at 58. In Kolon’s view, Owens “more likely” involved a situation under § 455(a) in which the judge’s “impartiality might [have] reasonably be[en] questioned,” not a § 455(b)(1) scenario implicating “personal bias or prejudice.”5 Id. at 59. Accordingly, Kolon believes Owens merely resolved that a timely motion is required when re-cusal is implicated under § 455(a), leaving open that question with respect to § 455(b).
In Kolon’s view, the § 455(a) and (b) provisions are different enough to explain the presence of a timeliness requirement in the former despite the absence of such in the latter. Unlike § 455(a), § 455(b) may not be waived by the parties. See 28 U.S.C. § 455(e) (“No [judge] shall accept from the parties to the proceeding a waiver of any ground for disqualification enumerated in subsection (b). Where the *168ground for disqualification arises only under subsection (a), waiver may be accepted-”). From Kolon’s perspective, subsections 455(b) and (e) create a “jurisdictional limitation on the authority of a judge to participate in a given case,” leaving the judge with a sua sponte obligation to recuse himself or herself when he or she knows the predicate facts implicating § 455(b). Appellant’s Br. at 60 (quoting United States v. Gipson, 835 F.2d 1323, 1325 (10th Cir.1988) (internal quotation marks omitted)). Thus, Kolon continues, “requiring a timely party motion as a condition precedent to enforcing section 455(b) runs contrary to statutory design, effectively relieving the judge of his personal statutory duty.” Id.
D.
While Kolon’s arguments do not entirely lack merit, we conclude that § 455(b), like § 455(a), includes a timely-filing requirement under Owens and that Kolon failed to comply with it.
As the parties and the district court have acknowledged, the party seeking re-cusal in Owens did not specify which provision of § 455 required it, and we did not cabin our holding to any specific provision of that section. In Owens, the defendant, after publicly accusing the then-Governor of West Virginia of bribery, filed a motion for recusal based on the presiding judge’s “long association” with the Governor, who was responsible for the judge’s appointment to various offices. 902 F.2d at 1155.
In our view, these facts could plausibly fit under either of two subsections, 455(a) or 455(b)(1). On the one hand, the scenario in Owens could certainly speak to § 455(a)’s concern with situations where a judge’s impartiality might reasonably be questioned. But on the other, as the district court here determined, the judge’s perceived allegiance to the West Virginia Governor could reasonably have concerned “a personal bias or prejudice concerning” the defendant — the Governor’s accuser— thus implicating § 455(b)(1). Given this ambiguity, we are left only with Owens’s unqualified announcement that “[timeliness is an essential element of a recusal motion” which is “judicially implied in § 455.” 902 F.2d at 1155. Given that blanket prescription, we decline to read Owens’s timeliness requirement so narrowly as to exclude § 455(b).
Our dissenting colleague correctly observes that Owens cites a case discussing § 455(a) alone, and that our later cases in that line have yet to address § 455(b).6 The limitations of past cases, however, are not controlling, particularly because the policy rationale underlying Owens’s timeliness requirement applies just as forcefully to § 455(b) as to any other recusal scenario.
Here — just as with a § 455(a) recusal, for example — the requirement of timeliness “prohibits knowing concealment of an ethical issue for strategic purposes,” United States v. York, 888 F.2d 1050, 1055 (5th Cir.1989), and “is vital ... to prevent waste and delay,” Owens, 902 F.2d at 1156. Meanwhile, the non-waivability of a *169§ 455(b) recusal does not excuse a party’s delay in filing. As the Fifth Circuit explained, “waiver and timeliness are distinct issues.” York, 888 F.2d at 1055. Whereas “section 455(e) prohibits the judge and the parties from agreeing among themselves to abrogate section 455(b),” a “timeliness requirement forces the parties to raise the disqualification issue at a reasonable time in the litigation.” Id. And even if the mandatory text of § 455 does imply a quasi-jurisdictional limitation on a judge’s authority to hear a case, in our view, that limitation must be balanced against the interests of fairness and efficiency served by the timeliness requirement we announced in Owens.
The dissent criticizes our reading of Owens as finding no support in the statutory text of § 455(b). We note that textual support for a § 455(a) timeliness requirement is similarly lacking, and that the language under that provision is similarly mandatory, yet under our precedent that requirement is beyond dispute. In any event, while we certainly agree that our analysis must begin with the statute’s plain language, the absence of a timeliness provision does not foreclose further inquiry here. Even the plain meaning of a statute is not conclusive “in the rare cases [in which] the literal application of a statute will produce a result demonstrably at odds with the intentions of its drafters.” United States v. Ron Pair Enters., Inc., 489 U.S. 235, 242, 109 S.Ct. 1026, 103 L.Ed.2d 290 (1989) (alteration in original) (internal quotation marks omitted). Section 455 “serves to ‘promote public confidence in the integrity of the judicial process.’ ” Dissent at 181 (quoting Liljeberg v. Health Servs. Acquisition Corp., 486 U.S. 847, 858 n. 7, 108 S.Ct. 2194, 100 L.Ed.2d 855 (1988)). In our view, failing to insist on a timeliness requirement for seeking recusal under § 455(b) directly undermines that legislative goal.7
In keeping with that end, our sister circuits have overwhelmingly found a timely filing requirement to be implied despite the text’s silence. See, e.g., Am. Prairie Constr. Co. v. Hoich, 560 F.3d 780, 789-91 (8th Cir.2009) (§ 455(a) and (b)); Omega Eng’g, Inc. v. Omega, S.A., 432 F.3d 437, 447-48 (2d Cir.2005) (§ 455(b)); Stone Hedge Props, v. Phoenix Capital Corp., 71 Fed.Appx. 138, 141 (3d Cir.2003) (unpublished) (§ 455(b)); United States v. Rogers, 119 F.3d 1377, 1380-83 (9th Cir.1997) (§ 455(a) and (b)); Summers v. Singletary, 119 F.3d 917, 920-21 (11th Cir.1997) (§ 455(b)); York, 888 F.2d at 1053-55 (5th Cir.1989) (§ 455(a) and (b)).
Meanwhile, only two circuits have refused to read in a timeliness requirement. The Seventh Circuit did so first, in SCA Services v. Morgan, 557 F.2d 110, 117 (7th Cir.1977), but has since called that decision into question on more than one occasion, see Schurz Commc’ns, Inc. v. FCC, 982 F.2d 1057, 1060 (7th Cir.1992) (“SCA Services is a weak precedent.]”) (Posner, J., in chambers); United States v. Murphy, 768 F.2d 1518, 1539 (7th Cir.1985) (observing that “our decision [in SCA Services] *170stands alone”). The Federal Circuit has also declined to impose a formal § 455 filing requirement, but in doing so created what amounts to a de facto filing obligation under principles of equity. See Polaroid Corp. v. Eastman Kodak Co., 867 F.2d 1415, 1421 (Fed.Cir.1989) (finding no strict timeliness requirement but denying Kodak’s requested relief due to its unreasonably tardy § 455 objection).
We recognize the countervailing interest in removing any judge who bears even the slightest appearance of partiality. But we should not ignore the harm that would ensue if litigants were permitted to treat motions for recusal as little more than a stratagem. As the Fifth Circuit observed, “it might legitimately be asked whether the spectacle of an attorney dragging his opponent through a long and costly proceeding, only to conclude by moving for disqualification of the judge, is not equally detrimental to public impressions of the judicial system” as is a potentially biased judge. Delesdernier, 666 F.2d at 121 (internal quotation marks omitted). “Congress did not enact § 455(a) to allow counsel to make a game of the federal judiciary’s ethical obligations.... ” Id. We should not subvert that legislative intent merely because a party instead seeks re-cusal under § 455(b).
Nor are we moved by the fact that parties may not waive recusal under § 455(b). In that context, everyone (the judge and the parties) has acknowledged a conflict, but seeks nonetheless to ignore it. Thus, waiver cannot be said to prejudice one party in particular, and will not produce the gamesmanship we condemn here.
The same must be said of § 455(f). That provision permits a judge with a financial conflict of interest under § 455(b)(4) to remain on the case if he has devoted substantial time to the matter and divests himself of the interest8. The dissent reads the limitation of this provision to financial conflicts alone as a legislative determination that “timeliness and efficiency are less important than ensuring that the impartiality of the judiciary is upheld.” Dissent at 181. As with the waiver provision, however, § 455(f) presents a clean trade-off between efficiency and impartiality. It does not address the concerns about tactical sandbagging present in this case.
The dissent also contends that our decision today “pivots responsibility [for recu-sal] from the judges to the litigants.” Dissent at 182. That is not entirely correct. We agree with our friend that when a judge “is aware of grounds for recusal under section 455, that judge has a duty to recuse himself or herself.” Dissent at 182 (internal quotation marks omitted). The scenario we address here arises only when a judge independently determines, even if wrongly, that he need not recuse and a party does not affirmatively seek recusal— that is, until an adverse decision has been handed down. Both efficiency and integrity require that we not reward a party’s tactics in these circumstances.
E.
Having held that Owens’s timely-filing requirement applies to recusal motions under § 455(a) and (b) alike, we next consider whether Kolon complied with that requirement by “raising] the disqualification ... [of the judge] at the earliest moment after [its] knowledge of the facts.” Owens, 902 F.2d at 1156 (quoting Satterfield v. *171Edenton-Chowan Bd. of Educ., 530 F.2d 567, 574-75 (4th Cir.1975)). It did not.
The fact that the district court judge had been a partner at McGuireWoods at the time of the Akzo litigation was public knowledge when Kolon first sought discovery of the Akzo case materials in the antitrust case in August 2009. Given Kolon’s scouring of the Akzo litigation court records, it further seems clear that Kolon had long known that McGuireWoods represented DuPont in the Akzo case. Kolon was also formally alerted to the potential conflict in May 2009, when the clerk of court issued its notice to the parties informing them of the judge’s financial interest in an entity affiliated with McGuireWoods. Finally, Kolon became aware of the judge’s direct (if negligible) involvement in the Akzo litigation in August 2010, when DuPont produced the Akzo files and privilege log.
In sum, Kolon knew every fact that eventually predicated its recusal motion almost a year before it first suggested recusal might be appropriate, in July 2011, and over a year before it finally filed its first recusal motion, in November 2011. On this record, Kolon quite clearly failed to “raise the disqualification ... [of the judge] at the earliest moment after [its] knowledge of the facts.” Id.9
Nor, in our view, is Kolon’s untimeliness excused by the fact that DuPont knew first of the district court judge’s involvement in the Akzo case and failed to alert the court of that fact until it eventually produced its privilege log in August 2010. For one, the judge’s direct involvement in the Akzo case was not the only (or even necessarily the strongest) basis for Kolon’s eventual § 455(b)(2) recusal motion: if, as Kolon believed, the Akzo litigation was actually a matter in controversy, the mere involvement of the judge’s former law partners— of which Kolon was clearly aware — would have required his recusal. For another, DuPont’s initial withholding of the relevant communications does not explain why, after its eventual disclosure, Kolon failed to raise the disqualification issue for nearly a year.
Our dissenting colleague warns that our decision as to recusal will only diminish public respect for our profession. “At the end of the day,” he writes, our “determination that Kolon’s recusal requests were untimely means that a district judge who ... is no longer permitted to conduct further proceedings involving the trade secrets claims[] presided over a trial that ended in a one billion dollar verdict and a twenty-year worldwide production shutdown injunction.” Dissent at 184 (emphasis omitted). The district judge did preside over such a trial, and our decision here cannot rewrite the past. We have concluded separately, however, that the trade secrets verdict must be vacated based on the judge’s evidentiary rulings. In our view, a single verdict — however large — that no longer exists can hardly impair public confidence more than would a rule transforming recusal under § 455 *172into an “additional arrow in the quiver of advocates in the face of [anticipated] adverse rulings.” In re Kan. Pub. Emps. Ret. Sys., 85 F.3d 1353, 1360 (8th Cir.1996) (alteration in original) (internal quotation marks omitted). We therefore hold that the district court acted within its discretion in denying Kolon’s recusal motion on timeliness grounds.10
III.
We next consider Kolon’s challenges to certain of the district court’s discovery rulings. We review such rulings for abuse of discretion, which may be found where “denial of discovery has caused substantial prejudice.” Nicholas v. Wyndham Int’l, Inc., 373 F.3d 537, 543 (4th Cir.2004).
A.
Throughout discovery, to enable its experts to perform their analysis, Kolon sought access to DuPont’s transaction-level and market-segment sales, pricing, and margin data. The district court denied Kolon’s initial requests for this information as overly broad and unduly burdensome, but gave Kolon leave to reformulate its request. Conceding that its initial requests had been overbroad, Kolon eventually requested production of a spreadsheet with data fields relevant to certain contested issues. The request indicated that the responsive document should be in native format from “any existing database.” Ap-pellee’s Br. at 43 (emphasis omitted).
The district court again denied Kolon’s request, concluding that (1) DuPont had already produced extensive documentation on the pertinent topics, such that the requested data would not be any more relevant than the information DuPont had already provided; (2) the request remained “sweeping and extensive”; (3) DuPont had shown that production of the requested documents would be “significantly burdensome”; and (4) the request had been filed very late in the discovery period, without adequate explanation for the delay by Ko-lon. J.A. 1020-22.
While we do not necessarily share the district court’s view that Kolon’s requested transaction-level data would have been no more relevant than the aggregate data DuPont had theretofore provided, we nevertheless find that the discovery denial was sufficiently justified by the court’s determination that the production would have been unduly burdensome. Kolon insists that thfe burden to DuPont was minimal because it requested only a “single spreadsheet” which it said could be “readily compiled from any existing database,” Appellant’s Br. at 43 (emphasis omitted), and because an affidavit from DuPont’s Global Financing Director indicated that DuPont already had an existing spreadsheet containing some of the requested transaction-level data. But this ignores the sweeping nature of the information requested, which included all transaction-level details regarding customers, geographic location, dates, products, amounts, price, cost, margins, and profits. And even if DuPont did have this information in “existing database[s],” that does not mean it would not have been very burdensome to compile the information into a “single spreadsheet.” Id.
Particularly considering the district court’s “wide latitude in controlling discovery,” Rowland v. Am. Gen. Fin., Inc., 340 *173F.3d 187, 195 (4th Cir.2003), we decline to disturb its ruling.
B.
Kolon also appeals the district court’s grant of a protective order barring a Rule 30(b)(6) deposition of DuPont on its strategic use of supply agreements. Justifying that order, the district court explained that Kolon had violated Federal Rule of Civil Procedure 30(b)(1) and Rule 30(H) of the Local Rules for the United States District Court for the Eastern District of Virginia by “failing to give reasonable written notice ... for [a] replacement deposition notice that it served on October 21, 2011,” and had wasted the time the court had extended it for completion of its depositions. J.A. 2715.
Again, we see no cause to disturb the district court’s discretionary ruling. While Kolon attempts to pin blame for the discovery delays on DuPont, it concedes that it gave only five days’ notice for the replacement deposition notice it served on October 21, 2011. As the district court determined, this violated Local Civil Rule 30(H), which generally requires eleven days’ advance notice of a deposition, and Federal Rule Civil Procedure 30(b)(1), which requires “reasonable” notice. Although Kolon maintains that the five-days’ notice was reasonable under the circumstances, the district court acted within its discretion in concluding otherwise.
IV.
Finally, we consider Kolon’s challenge to the district court’s grant of summary judgment on its two antitrust claims—a ruling we review de novo, viewing all facts and reasonable inferences therefrom in the light most favorable to Kolon, the nonmov-ing party. See Pueschel v. Peters, 577 F.3d 558, 563 (4th Cir.2009).
In general, summary judgment is appropriate where there is no genuine issue as to any material fact. Fed.R.Civ.P. 56(a). A genuine issue of material fact exists when there is sufficient evidence on which a reasonable jury could return a verdict in favor of the nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). We have explained that summary judgment is “an important tool for dealing with antitrust cases,” Oksanen v. Page Mem’l Hosp., 945 F.2d 696, 708 (4th Cir.1991) (en banc), and that antitrust cases are “particularly well-suited for Rule 56 utilization” due to the “unusual entanglement of legal and factual issues” they often present, Thompson Everett, Inc. v. Nat’l Cable Adven, L.P., 57 F.3d 1317, 1322 (4th Cir. 1995).
A.
We first review the district court’s grant of summary judgment on Kolon’s monopolization claim.
Under § 2 of the Sherman Act, a defendant is liable for a monopolization claim when that defendant (1) possesses monopoly power and (2) willfully acquires or maintains that power. DuPont I, 637 F.3d at 441. In granting summary judgment to DuPont, the district court held that Kolon failed to create a genuine issue of material fact on either prong, concluding that DuPont neither possessed monopoly power nor engaged in willful maintenance of such power. We address each element in turn.
1.
“Monopoly power is the power to control prices or exclude competition.” United States v. E.I. du Pont de Nemours & Co., 351 U.S. 377, 391, 76 S.Ct. 994, 100 L.Ed. 1264 (1956). A defendant possesses *174monopoly power in the relevant market if it is “truly predominant in the market.” White Bag Co. v. Int’l Paper Co., 579 F.2d 1384, 1387 (4th Cir.1974). Although there is no fixed percentage market share that conclusively resolves whether monopoly power exists, the Supreme Court has never found a party with less than 75% market share to have monopoly power. Antitrust Laws & Trade Regulation: Desk Ed. § 3.02[2][c][ii]. And we have observed that “when monopolization has been found the defendant controlled seventy to one hundred percent of the relevant market.” DuPont I, 637 F.3d at 450 (quoting White Bag, 579 F.2d at 1387).
Beyond percentage market share, “some courts have also focused on the durability of the defendant’s market power, particularly with an eye toward other firms’ (in)ability to enter the market.” Id. at 451 (citing cases).
Applying these standards, the district court held that DuPont lacked monopoly power. Whereas (according to our ruling in DuPont I) Kolon had adequately pleaded the monopoly power element by alleging that DuPont had controlled over 70% of the relevant market, at the summary judgment stage the district court found that DuPont actually possessed significantly less than the alleged 70% of that market. Kolon Indus., Inc., v. E.I. du Pont De Nemours & Co., No. 3:11-ev-622, 2012 WL 1155218, at *12 (E.D.Va. Apr. 5, 2012). “In fact,” the court observed, “Kolon’s own expert takes the view that DuPont had a maximum market share of 59 percent during the relevant time period, and that DuPont’s market share decreased to 55 percent during that three year period rather than increased.” Id.
This decline in DuPont’s market share, combined with Teijin’s corresponding ascendance and the fact that DuPont was charging lower prices in the United States than in Europe (which Kolon identified as a comparable market), led the court to its conclusion. “[T]he fact that there are significant entry barriers,” the district court continued, “is insufficient to fill the factual gaps in Kolon’s monopolization claim.” Id. “DuPont clearly lacks the power to control prices and exclude competition,” the court summarized, “otherwise, it would have been able to prevent the decrease in its market share and the rise of one of its major competitors.” Id.
Even viewing all evidence in the light most favorable to Kolon, we agree with the district court that DuPont did not possess monopoly power in the U.S. para-aramid market during the relevant period between 2006 and 2009. First, although Kolon is correct that DuPont’s market share of less than 60% during the relevant period does not necessarily foreclose a finding of monopoly power, it does weigh heavily against such a finding. Quite simply, this percentage falls significantly short of where we have previously drawn the line for monopoly power. See DuPont I, 637 F.3d at 450 (identifying 70% market share as the bottom of the range for a finding of monopoly power).
Meanwhile, although Kolon is also correct that certain other factors do demonstrate DuPont’s strength in the market (e.g., high barriers to entry, ability to price discriminate, high profit margins), a showing of DuPont’s “market power” is not itself sufficient to prove that DuPont possesses “monopoly power.” See Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 481, 112 S.Ct. 2072, 119 L.Ed.2d 265 (1992) (“Monopoly power under § 2 requires, of course, something greater than market power under § 1.”). Furthermore, this evidence falls short of showing DuPont’s durability in the market. As the district court observed, uncontested facts demonstrate that DuPont has experi*175enced a steady, decades-long loss in significant market share to Teijin.
Ultimately, in light of DuPont’s reduced market share and lack of durable market power, the evidence cannot sustain a jury finding that DuPont had the “power to control prices or exclude competition,” United States v. du Pont, 351 U.S. at 391, 76 S.Ct. 994, or was “truly predominant in the market” during the relevant period, White Bag, 579 F.2d at 1387.
2.
Even if Kolon had presented a triable issue on the monopoly-power element, Ko-lon also needed to show that DuPont willfully maintained that power. To violate this prong, a defendant must engage in conduct “to foreclose competition, to gain a competitive advantage, or to destroy a competitor.” Eastman Kodak, 504 U.S. at 482-83, 112 S.Ct. 2072. On this element, Kolon’s theory was — and is — that DuPont maintained its alleged monopoly power through the use of long-term, multi-year, exclusive supply agreements with certain U.S. para-aramid customers.
Although exclusive dealing agreements are not per se illegal, they “may be an improper means of acquiring or maintaining a monopoly.” DuPont I, 637 F.3d at 451 (citing United States v. Grinnell Corp., 384 U.S. 563, 576, 86 S.Ct. 1698, 16 L.Ed.2d 778 (1966)). The Supreme Court has held that an exclusive dealing arrangement does not violate antitrust laws unless its probable effect is to “foreclose competition in a substantial share of the line of commerce affected.” Tampa Elec. Co. v. Nashville Coal Co., 365 U.S. 320, 327, 81 S.Ct. 623, 5 L.Ed.2d 580 (1961). The Court explained:
To determine substantiality in a given case, it is necessary to weigh the probable effect of the contract on the relevant area of effective competition, taking into account the relative strength of the parties, the proportionate volume of commerce involved in relation to the total volume of commerce in the relevant market area, and the probable immediate and future effects which pre-emption of that share of the market might have on effective competition therein.
Id. at 329, 81 S.Ct. 623.
Along these lines, we have observed that “[t]he market share foreclosed is important because, for the contract to adversely affect competition, ‘the opportunities for other traders to enter into or remain in that market must be significantly limited[.]’ ” DuPont I, 637 F.3d at 451 (quoting Tampa Elec., 365 U.S. at 328, 81 S.Ct. 623). Once a plaintiff has demonstrated substantial foreclosure, it must then also demonstrate that the conduct had “a negative impact on competition in the market as a whole.” Chuck’s Feed & Seed Co. v. Ralston Purina Co., 810 F.2d 1289, 1295 (4th Cir.1987).
The district court held there was no genuine issue that DuPont’s supply agreements had not foreclosed a substantial portion of the market. In its view, Kolon had not sufficiently attempted to quantify foreclosure of the entire relevant market, and instead had focused only on DuPont’s alleged foreclosure of particular market segments. Kolon’s evidence of the degree of foreclosure in those segments, which the court characterized as “scant at best,” did “nothing to reveal the amount of foreclosure in the [para-aramid] market as a whole.” Kolon, 2012 WL 1155218, at *14. The court concluded that since DuPont had supply agreements — many of which were non-exclusive — with only twenty-one of approximately 1,000 potential commercial U.S. para-aramid customers, the percentage of foreclosure could not, “as a matter of law, constitute sufficient grounds *176for a finding of substantial foreclosure.” Id. at *15.
The court also concluded that Kolon had “put forth no evidence” that DuPont’s supply agreements- had a negative effect on overall competition, noting that Teijin’s “relentless ascendance” fatally undercut that claim. Id. Finally, the court rejected Kolon’s argument that DuPont’s twenty-one supply arrangements substantially foreclosed the entire relevant market by blocking Kolon from crossing a “critical bridge” to “high volume” customers. Id. at *16-18.
On appeal, Kolon again stresses its “critical bridge” theory. While it does not deny that DuPont had supply agreements with only twenty-one of the roughly 1,000 potential U.S. commercial para-aramid customers, Kolon contends that the district court’s emphasis on those figures — and its disregard of the “probable effect of the contracts] on the relevant area of effective competition,” Tampa Elec., 365 U.S. at 329, 81 S.Ct. 623 — -was shortsighted. Pointing to evidence that DuPont perceived Kolon’s market entry as a threat, Kolon argues that DuPont “strategically entered into supply agreements with high-volume customers in the key commercially sustainable entry segments ... that Kolon sought to enter.” Appellant’s Br. at 7, 27-28. Kolon submits that despite the relatively low number and short duration of DuPont’s supply agreements, these agreements “choked off the ‘critical bridge’ to Kolon’s entry into the U.S. market” because they foreclosed Kolon’s access to the most important high-volume customers. Id. at 32.
Kolon points to two cases from the Third Circuit which, in its view, embrace this “critical bridge” approach. See United States v. Dentsply Int'l, Inc., 399 F.3d 181 (3d Cir.2005) (reversing summary judgment, holding that Dentsply’s exclusivity agreements with key product distributors could deny efficient scale to competitors); LePage’s Inc. v. SM, 324 F.3d 141, 160 (3d Cir.2003) (en banc) (reversing summary judgment, holding that 3M’s bundled rebate agreements with superstores like K-Mart and Wal-Mart could have cut Le-Page’s off from “key retail pipelines necessary to permit it to compete profitably”).
While we acknowledge that a singular emphasis on the percentage of customers foreclosed cannot resolve the inquiry (as foreclosure of a few important customers could substantially foreclose access to a market), we agree with the district court that Kolon failed to show what “proportionate volume of commerce” in the entire relevant market was foreclosed by DuPont’s supply agreements. Tampa Elec., 365 U.S. at 329, 81 S.Ct. 623; see also DuPont I, 637 F.3d at 451 (discussing the importance of market share foreclosed). Likewise, although Kolon’s “critical bridge” theory is certainly plausible, the evidence does not support its application here.
.Unlike the plaintiffs in Dentsply and LePage’s, Kolon offered no evidence that access to the foreclosed customers (or even to the identified market segments) was necessary to achieve scale in the broader U.S. para-aramid market. And even if we assume the significance of those customers and market segments, Kolon does not dispute that DuPont had supply agreements with fewer than half of its identified “key” customers within, those segments.
Meanwhile, DuPont persuasively distinguishes Dentsply and LePage’s based on the fact that the defendants in those “critical bridge” cases foreclosed the plaintiffs’ access to distribution networks rather than end-customers. We are not convinced that, as Kolon contends, this is “a distinction without a difference.” Reply Br. at *17715. As the district court observed, unlike with Dentsply’s and 3M’s agreements that foreclosed access to distribution networks shown to be necessary to reach many end-customers, “the record presents no reason to think that Kolon could not sell to other customers occupying the same segment of the para-aramid market ... as customers that have supply agreements with DuPont.” Kolon, 2012 WL 1155218, at *18.
In sum, we conclude that neither the probable nor the actual effect of DuPont’s supply agreements was to “foreclose competition in a substantial share of the line of commerce affected.” Tampa Elec., 365 U.S. at 327, 81 S.Ct. 623. Accordingly, those agreements do not violate the willful maintenance prong of our § 2 monopolization inquiry. Because Kolon failed to raise a genuine issue of material fact as to either prong, summary judgment was appropriate on its monopolization claim.
B.
We next review the district court’s grant of summary judgment on Kolon’s attempted monopolization claim.
“Attempted monopolization employs ‘methods, means and practices which would, if successful, accomplish monopolization, and which, though falling short, nevertheless approach so close as to create a dangerous probability of it.’ ” DuPont I, 637 F.3d at 453 (quoting M & M Med. Supplies & Serv., Inc. v. Pleasant Valley Hosp., Inc., 981 F.2d 160, 166 (4th Cir.1992)). To prevail on an attempted monopolization claim under § 2, a claimant must show (1) a specific intent to monopolize a relevant market, (2) predatory or anticompetitive acts, and (3) a dangerous probability of successful monopolization. Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 456, 113 S.Ct. 884, 122 L.Ed.2d 247 (1993).
Focusing only on the final two prongs, the district court found neither satisfied since (1) Kolon had failed to demonstrate substantial foreclosure of the relevant market resulting from DuPont’s supply agreements, meaning there was no anti-competitive conduct; and (2) DuPont had lost market share during the relevant period and had failed to prevent Teijin’s ascendance, meaning there was no dangerous probability of successful monopolization by DuPont.
Kolon first contends that DuPont’s supply agreements were anticompetitive, arguing that DuPont entered these agreements against its own interest in order to block Kolon’s market entry. Appellant’s Br. at 31-32 (citing M & M Med. Supplies, 981 F.2d at 166 (noting that where “a firm has been attempting to exclude rivals on some basis other than efficiency, it is fair to characterize its behavior as predatory”))- On the “dangerous probability of success” prong, Kolon maintains that even if DuPont’s share of the U.S. para-aramid market did not constitute actual monopoly power, it was at least consistent with a “dangerous probability” of achieving such power. Id. at 23-24 (citing M & M Med. Supplies, 981 F.2d at 168 (4th Cir.1992) (“[Cjlaims involving greater than 50% share should be treated as attempts at monopolization when the other elements for attempted monopolization are also satisfied.”)). And Kolon notes that even though “DuPont’s market share declined slightly over the three-year period, that does not, as a matter of law, preclude a finding of monopoly power, much less a dangerous probability of achieving it.” Id. at 24 (citing cases finding monopoly power despite a declining market share).
But again, even viewing the evidence in the light most favorable to Kolon, the claim fails.
*178First, as discussed above, DuPont’s alleged anticompetitive conduct — its customer supply agreements — did not have the probable effect of “foreclosing] competition in a substantial share of the line of commerce affected.” Tampa Elec., 365 .U.S. at 327, 81 S.Ct. 623; see also IIIB Areeda & Hovenkamp, Antitrust Law ¶ 806a, at 412 (3d ed. 2008) (“[T]he same basic definition of exclusionary conduct should apply to both monopolization and attempt claims.”). Nor, contrary to its suggestion, did Kolon show that the agreements were anticompetitive as without business justification or against DuPont’s own interest. Rather, DuPont introduced unrebutted evidence that it entered the supply agreements as a competitive response to Teijin’s use of that same practice, and because customers requested them.
Second, Kolon has not raised a genuine issue that DuPont had a “dangerous probability” of successfully achieving monopoly power during the relevant period. As the district court observed, DuPont’s market share had been in steady decline for seventeen years, and DuPont has proven unable to control U.S. prices or exclude Teijin from entering the market. And even if declining market share does not preclude a finding of monopoly power, Kolon pointed to no affirmative evidence indicating a “dangerous probability” that DuPont would sooner or later regain its former market dominance.
Accordingly, we affirm the district court’s grant of summary judgment to DuPont on Kolon’s attempted monopolization claim.
V.
In sum, we conclude that following Owens, recusals under 28 U.S.C. § 455(b) include a judicially implied timely-filing requirement, and that the district court acted within its discretion when it denied Kolon’s recusal motion on timeliness grounds.
We defer to the district court’s considerable discretion in overseeing discovery and will not disturb its discovery rulings. On the merits of Kolon’s antitrust suit, we agree with the district court that Kolon failed to raise a triable issue of material fact sufficient to sustain either its attempted or actual monopolization claims.
The judgment of the district court is hereby
AFFIRMED.
. We recite the relevant facts in the light most favorable to Kolon. In so doing, we have done our best to honor the parties’ oft-overzealous desires to keep certain purportedly *163sensitive information under seal. However, to the extent the district court has already disclosed (without objection) such information, we treat it as public knowledge.
. In its motion and supporting memorandum, Kolon also suggested that the district court "revisit its refusal to recuse” from the trade secrets case, apparently referring to the discussion of recusal in July. Docket No. 247, at 2; Docket No. 248, at 31. Kolon filed no separate motion for recusal in the trade secrets case at that time, but did briefly reference recusal in several subsequent filings. On December 9, 2011, Kolon filed a reply in support of its motion for a judgment as a matter of law; in a footnote, it reminded the court of its position that the judge should not have ruled regarding the adverse jury instructions. See Docket No. 1738, at 13. On December 23, Kolon asked the court to consider recusal in its memorandum supporting its motion to stay the injunction proceedings. See Docket No. 1813, at 19. And in its reply regarding that motion, on January 11, Kolon insisted that a formal motion for recusal was unnecessary. See Docket No. 1843, at 19 (responding to Docket No. 1830, at 26-27). During a January 2012 hearing on Kolon’s motion for a new trial and judgment as a matter of law in the trade secrets case, counsel for Kolon again requested that the district court judge recuse himself. The judge refused to do so, explaining that he did not have a recusal motion before him in that case. Three days later, on January 27, 2012, Kolon filed its motion for recusal and disqualification in the trade secrets case. The district court denied the recusal motions in both cases on Februaiy 21.
.A separate recusal statute, 28 U.S.C. § 144, provides parties with one opportunity per case to file an affidavit that the presiding judge has a personal bias or prejudice regarding a party. If the affidavit is sufficient, accompanied by a certificate of good faith, and timely filed, another judge will be assigned to the proceeding. Kolon did not seek recusal on this ground in the district court.
. Additionally, the district court held that re-cusal was unnecessary under § 455(a). Since Kolon does not appeal that ruling, we do not address it.
. 28 U.S.C. § 455(b)(1) requires a judge to recuse himself “[w]here he has a personal bias or prejudice concerning a party, or personal knowledge of disputed evidentiary facts concerning [a] proceeding.”
. The dissent says that our decision in United States v. Lindsey, 556 F.3d 238 (4th Cir.2009), cuts against imposing a timeliness requirement under § 455(b). With all respect, we do not share that view. In Lindsey, the presiding district court judge had participated in defendant Lonnie Robinson’s case twelve years earlier as an Assistant United States Attorney. Though the judge did not recall his participation in the earlier case, nor was he made aware of it, we vacated his order. But in that case, Robinson did not learn of the judge's prior involvement until after filing his appeal. See id. at 246-47. Thus, no timeliness issue ever arose: the case is wholly inapposite here.
. Section 455’s legislative history is murky at best. See Delesdernier v. Porterie, 666 F.2d 116, 119-121 (5th Cir. 1982). When Congress revised the statute in 1974, the Justice Department did suggest adding an explicit timeliness requirement like that found in § 144. Id. at 120. Congress declined to do so, and our friend in dissent believes that decision "end[s] our inquiry.” Dissent at 180. But the Justice Department’s suggestion is hardly the only piece of relevant legislative history. Rather, "prior to the 1974 amendment,] courts had generally held that a timely objection under the old § 455 was necessary.” Delesdernier, 666 F.2d at 121. ”Thus[,] Congress' [s] failure to act could as easily have been the result of a belief that the judicial gloss on old section 455 would survive.” Id. (internal quotation marks omitted).
. The district judge's May 2009 disclosure of a financial interest does not present an issue under § 455(f), as § 455(b)(4) requires recu-sal only where a financial interest not immediately in controversy "could be substantially affected by the outcome of the proceeding.”
. We recognize that Kolon filed its motion to recuse in this case before the district court's issuance of an adverse ruling on summary judgment. But its actions here should not be viewed in a vacuum. Recall that Kolon’s antitrust claims arose as a counterclaim to DuPont’s trade secrets action, which proceeded more quickly than the antitrust case, before the same district judge. In that case, the judge issued a series of rulings universally adverse to Kolon, and a jury rendered a $920 million verdict for DuPont. This had all transpired by the time Kolon filed its recusal motion in the antitrust case. So while Ko-lon’s sandbagging may not be obvious in the isolated context of the antitrust case, a global view of the relevant events makes clear that Kolon held its fire on recusal until after suffering a defeat in the trade secrets case.
. In light of our holding, we do not address the district court’s alternative ruling that, on the merits, recusal was not required.