SUTTON, Circuit Judge.
Faurecia Interior Systems USA appeals (1) the district court’s summary judgment decision that it breached a confidentiality agreement with Multimatic and (2) the jury’s $10 million damage award. Multimatic cross-appeals the district court’s exclusion of its damages expert, who claimed Multimatic suffered an additional $28.7 million in damages. Because the district court properly granted summary judgment on liability, the jury’s verdict is supported by sufficient, and properly admitted, evidence and the district court did not abuse its discretion in excluding Multimatic’s damages expert, we affirm.
I.
In October 2003, DaimlerChrysler selected Faurecia to supply instrument panels for its JS41 vehicle program, which [646]*646covered the 2007 to 2012 Sebring and Avenger vehicle lines. In early 2004, Faurecia selected Multimatic to design the instrument panel’s cross-car beam. In connection with this agreement, Multimatic asked Faurecia to sign a confidentiality agreement, which it did in February 2004.
With the confidentiality agreement in place, Multimatic and Faurecia began collaborating weekly on the cross-car beam design. At the same time, Multimatic began working on an alternate design— dubbed the “mass saving design” — that would meet Faurecia’s requirements but weigh several pounds less than the current design. By November 2004, Multimatic was ready to focus solely on the mass saving design. But because the parties had not yet agreed on a price for the cross-car beams, Multimatic worried (correctly) that Faurecia might switch suppliers. Faurecia signed a pre-development letter of intent that month to assuage Multimatic’s concerns, and the two parties shifted their focus to refining the mass saving design.
In April 2005, Faurecia again looked to switch cross-car beam suppliers because the parties could not agree on a price for the product. It selected Brown Corporation as its new pre-production cross-car-beam supplier in May 2005 and terminated its relationship with Multimatic. Brown remained the cross-car-beam supplier once the JS41 program began production in 2006.
In May 2005, Multimatic filed this lawsuit against Faurecia, alleging that it breached the confidentiality agreement, breached the November 2004 letter of intent and breached several prototype purchase orders. The district court granted summary judgment for Multimatic on the breach-of-confidentiality-agreement claim and summary judgment for Faurecia on the breach-of-letter-of-intent claim. After a trial, a jury awarded Multimatic $9,381,306 for Faurecia’s breach of the confidentiality agreement and $600,515 for breach of the purchase orders. Both parties appeal.
II.
We first consider whether the district court properly granted summary judgment to Multimatic on liability. This aspect of the appeal boils down to two questions: (1) Does the confidentiality agreement cover information created after Multimatic and Faurecia executed it and (2) did Faurecia reveal information owned exclusively by Multimatic? In answering these questions, we must give Faurecia the benefit of all reasonable factual inferences. See Cavin v. Honda of Am. Mfg., Inc., 346 F.3d 713, 719 (6th Cir.2003).
A.
As to the first question, Multimatic and Faurecia signed the confidentiality agreement so that they could share “their respective CAD designs, drawings, analysis models and results” concerning the cross-car beam that “Multimatic will design and supply.” ROA 236. The agreement protects “sensitive information” that “each of Multimatic ... and Faurecia possesses,” defined as:
[Proprietary confidential information pertaining to its business and customers and ... technical information relating to its products, designs and services, including compositions, raw materials, formulations, additives, components, production processes, plant layout, engineering concepts and designs, analysis models and results, know-how, and other intellectual or industrial property, which is generally not available to the public.
Id. The agreement adds that Multimatic and Faurecia must disclose any sensitive information “necessary to develop and [647]*647supply” the cross-car beam during the agreement’s three-year term. Id. And it provides that neither party may share sensitive information obtained under the confidentiality agreement with third parties unless they first impose “similar” confidentiality obligations. Id.
Under Michigan law, as under the law of most (if not all) States, courts will enforce the terms of a contract if they have a “clear, unambiguous, and ... definite meaning.” Mahnick v. Bell Co., 256 Mich.App. 154, 662 N.W.2d 830, 832-33 (2003); see also Rory v. Continental Ins. Co., 473 Mich. 457, 703 N.W.2d 23, 28 (2005). Courts determine the meaning of unambiguous contract provisions, and juries generally determine the meaning of ambiguous provisions when there is competing extrinsic evidence over how to construe them. Mahnick, 662 N.W.2d at 833.
In our view, as in the view of the district court, the confidentiality agreement unambiguously covers pre-existing and future trade secrets. The preamble looks to the future, contemplating that Multimatic “will design” the cross-car beam and will share its “designs, drawings” and other information created during that process. ROA 236. The agreement’s disclosure obligation confirms this forward-looking perspective, mandating disclosure of sensitive information “necessary to develop” the cross-car beam. Id. And nothing in the agreement draws a distinction between pre-existing and future information. The definition of “sensitive information” does not include the modifier “pre-existing” or otherwise restrict itself to a discrete time period.
The agreement also confirms that the parties knew how to exclude information from the compass of protected trade secrets. Immediately after defining sensitive information, it excludes information in the public domain and information the receiving party lawfully possessed prior to disclosure. In this context, had the agreement meant to exclude information postdating its execution, one would expect to see language to that effect alongside the other carve-out clauses.
Our interpretation also gives effect to “every word” in the definition of sensitive information. See Associated Truck Lines, Inc. v. Baer, 346 Mich. 106, 77 N.W.2d 384, 386 (1956). The definition picks up several types of information created during the design process — many of which necessarily would occur in the future — • such as engineering designs, analysis models and results. If the agreement covers future information, as we think it does, the inclusion of these types of covered information in the definition imposes meaningful obligations. Were we to interpret the confidentiality agreement to cover only pre-existing information, by contrast, that would leave these provisions with no role to play.
Faurecia offers three rejoinders. First, it notes that the agreement repeatedly uses the present tense, including the verb “possesses” in the definition of sensitive information, suggesting that the agreement covers just pre-existing trade secrets. But it is not that easy. All contracts use the present tense — at least in part. And the parties’ use of the present tense here provides a baseline for creating existing and future obligations, and the context of the agreement and its other terms show that the agreement covers both types of commitments. Proving the point, legal drafters frequently use the present tense to cover the present and the future. See, e.g., 1 U.S.C. § 1; Mich. Comp. Laws §§ 141.602, 141.1162, 206.2, 257.80; Heniser v. Frankenmuth Mut. Ins. Co., 449 Mich. 155, 534 N.W.2d 502, 503, 505, 509 (1995) (interpreting contract term to cover the future even though it was drafted in the present tense); Bryan [648]*648A. Garner, A Dictionary of Modem Legal Usage 939 (2d ed.1995). Verb tense alone cannot hold the weight that Faurecia places on it.
Second, the obligation to disclose sensitive information “to the extent necessary to design and develop” the cross-car beam, Faurecia argues, has no work to do if the agreement includes information created during the design process. Not so. The phrase still protects pre-existing information from disclosure to the same extent as an interpretation covering only pre-existing information.
Third, Faurecia worries that extending the confidentiality agreement to cover future information — especially the cross-car beam design itself — -would give Multimatic the ability to “shut[] down production of an entire line of vehicles” because the agreement allows either party to request the immediate return of its sensitive information at any time. Faurecia Br. at 29. But Chrysler’s policies indicate that it will not incorporate third-party designs into its production vehicles unless it has an irrevocable right to produce and use that design. The confidentiality agreement thus was prologue to a supply or licensing agreement, which would address these types of concerns if and when Faurecia chose to use (and pay for) Multimatic’s design.
Nor does the promise to protect sensitive information become illusory if the confidentiality agreement covers future information. Faurecia complains that Multimatic could circumvent its disclosure obligation by immediately asking Faurecia to return any disclosed design materials. But the agreement’s implicit covenant of good faith and fair dealing alleviates this concern. See Lowes Home Centers, Inc. v. LL & 127, LLC, 147 Fed.Appx. 516, 523-24 (6th Cir.2005).
B.
Faurecia separately argues that Multimatic did not exclusively own the information that Faurecia disclosed to Brown. Yet it is beyond question, as an initial matter, that Faurecia shared the cross-car beam design and associated engineering documents with Brown in April or May 2005. Brown needed the information to provide an accurate quote, which it gave to Faurecia on May 18, 2005. Brown’s 2006 three-dimensional cross-car beam models also perfectly match Multimatic’s 2005 models “except for minor revisions ... resulting from a more mature vehicle package” by Chrysler. R.84 Ex. 33 at 21, 27. Faurecia offers no evidence suggesting that Brown recreated the design from scratch.
Once Faurecia transitioned the cross-car-beam project from Multimatic to Brown, it also disclosed Multimatic’s design-failure mode and effects analysis and its design verification plan and report. Both represent “core engineering documents” that disclose Multimatic’s quality assurance processes, such as how to test for and prevent cross-car beam failures. R.84 Ex. 33 at 4, 20. These documents by their nature come within the categories of “information pertaining to [Multimatic’s] business” and “analysis models and results” protected by the confidentiality agreement. ROA 236.
Multimatic exclusively owned this information. All of the record evidence — including deposition testimony by two Faurecia employees intimately involved in the cross-car beam development — indicates that Multimatic alone created the 3D models Faurecia shared with Brown in 2005. Faurecia may well dispute who came up with some of the ideas embodied in the 3D models, but that does nothing to counter the reality that Multimatic created the models.
[649]*649In addition, Multimatic solely created and solely owned the engineering documents Faurecia provided to Brown. Only Multimatic employees are identified as creating and maintaining those documents. And Faurecia forwarded the Multimatic emails containing those documents to Brown without any modification or comment.
Copyright law- — the legally cognizable form of ownership in 3D models, see 17 U.S.C. § 102(a)(5); Robert R. Jones Assocs., Inc. v. Nino Homes, 858 F.2d 274, 278 (6th Cir.1988) — confirms Multimatic’s exclusive ownership of the cross-car beam models. To transfer copyright ownership, the copyright owner must sign a written instrument acknowledging the transfer. See 17 U.S.C. § 204(a). Faurecia has not produced any writing signed by Multimatic that transfers its copyright in the 3D models to Faurecia or to Chrysler.
Faurecia disagrees, pointing to a January 2005 email from Lee James, a Multimatic engineer, who agrees to comply with several of Chrysler’s cross-car beam specification requirements, including some in Chrysler’s PS-7000 process standard. In that standard, Chrysler outlines technical specification requirements for designs created by third-party suppliers. But it also describes two potential IP ownership arrangements between outside suppliers and Chrysler: either the supplier retains IP ownership but grants Chrysler a license to use the IP or the supplier assigns IP ownership to Chrysler.
According to Faurecia, a jury could infer from James’ email that Multimatic agreed to give Chrysler ownership in the cross-car beam design. That is a stretch. A jury could not reasonably infer that a Multimatie engineer’s email to Faurecia acknowledging screw-torque testing procedures, welding specifications and “mill oil and lubricant requirements” simultaneously assigned the company’s intellectual property ownership to Chrysler — all without mentioning intellectual property ownership at any point in the email. R.85 Ex. R. The PS-7000 standard itself also acknowledges that suppliers do not transfer IP ownership until Chrysler and the supplier execute a purchase order for the supplier’s vehicle component. Multimatic and Chrysler never did so for the cross-car beam.
Faurecia persists that Multimatic assigned ownership of its 3D models to Chrysler by uploading them to Chrysler’s Virtual Product Manager (VPM) system. By logging into VPM, Multimatic “acknowledg[ed]” that VPM and its data are valuable DaimlerChrysler assets and that it could use these assets only in authorized ways. R.91 Ex. LL. A “Confidentiality Agreement” on the VPM welcome screen also states, “All information and content on the [Chrysler] network is the confidential and proprietary property of’ Chrysler, which VPM users cannot disclose to others. R.91 Ex. LL.
Yet these statements do not automatically assign ownership over information uploaded into VPM to Chrysler. To acknowledge that someone owns a book is not the same as acknowledging that they own the copyright in the book. By logging into VPM, Multimatic likewise acknowledged only that the data in VPM — not the intellectual property embodied in that data — belong to Chrysler.
Chrysler’s PS-7000 process standard adds force to this conclusion. All 3D models for Chrysler vehicle components must be uploaded into VPM, so any promise under the PS-7000 standard that a supplier could retain IP ownership over its design would be a dead letter if uploading information into VPM automatically transferred ownership to Chrysler. Two acknowledgments on a login screen did not render Chrysler’s carefully crafted con[650]*650tract terms meaningless and deceptive. In the end, Faurecia fails to offer any reason why the application of a carefully negotiated confidentiality agreement in this context would implicitly give up what it explicitly protects: intellectual property rights.
III.
Faurecia separately challenges the jury’s $10 million damage award. Multimatic’s principal evidence supporting its $10 million lost-profits claim came through the testimony of Lawrence Simon, one of its damages experts. Simon, a CPA for twenty-four years, testified that Multimatic lost roughly $9.4 million in profits when Brown, not Multimatic, supplied the cross-car beam for the JS-41 program. Simon calculated this figure by multiplying the number of cars Chrysler would produce under the JS41 program (according to publicly available automotive forecasting information) by Multimatic’s estimated profits per cross-car beam. See id. at 28. He then estimated the profits per beam based on a February 2005 Multimatic price quote, which included a pricing structure similar to the November 2004 letter of intent.
Simon also testified that Multimatic incurred an additional $600,515 in damages from unpaid purchase orders for cross-car beam prototypes. He arrived at this figure by adding up the invoice amount on each unpaid prototype invoice that he could tie to a Faurecia purchase order. This included purchase orders for prototypes that Multimatic produced but did not ship to Faurecia in May 2005.
Faurecia claims the jury verdict cannot stand as a matter of law. We give fresh review to this claim, apply the law of the forum state (Michigan) in assessing the sufficiency of the evidence, see K & T Enters., Inc. v. Zurich Ins. Co., 97 F.3d 171, 175-76 (6th Cir.1996), and view the evidence in the light most favorable to Multimatic, see In re Brown, 342 F.3d 620, 626-27 (6th Cir.2003).
According to Faurecia, Multimatic cannot show that the parties contemplated lost-production profits for breaching the confidentiality agreement. A plaintiff can recover only for damages “that arise naturally from the breach or those that were in the contemplation of the parties at the time the contract was made.” Kewin v. Mass. Mut. Life Ins. Co., 409 Mich. 401, 295 N.W.2d 50, 52-53 (1980). mile we agree that lost-production profits are not an inevitable consequence of breaching a confidentiality agreement, the jury had ample evidence to conclude that the parties contemplated such damages at the time of contracting.
Mark Sullivan, a Multimatic manager who oversaw the cross-car beam project, testified about the company’s intention in signing the confidentiality agreement: Faurecia remained free to use any supplier it wished for the JS41 program, but if it used Multimatic’s design, Faurecia had to permit Multimatic to supply the beam. He pointed to language in the agreement that reflected this intention. And he testified that he communicated this understanding to Faurecia on several occasions. Based on this evidence, a jury could reasonably infer that Faurecia understood that, if it used the Multimatic design protected by the confidentiality agreement, it would have to use Multimatic to supply the product.
Faurecia fares no better in arguing that the lost-profit award is unduly speculative. Under Michigan law, sure enough, a plaintiff cannot recover damages “based on mere conjecture or speculation.” Sullivan Indus., Inc. v. Double Seal Glass Co., Inc., 192 MichApp. 333, 480 N.W.2d 623, 632 (1991). Yet damages calculations with a “reasonable basis of computation” clear [651]*651this hurdle even though the results are “only approximate,” Waskin Dev. Co. v. Weyn, 369 Mich. 121, 119 N.W.2d 662, 665 (1963), and “speculative to some degree,” Lorenz Supply Co. v. Am. Standard, Inc., 100 Mich.App. 600, 300 N.W.2d 335, 340 (1980).
The damages evidence satisfies this standard. Simon had a reasoned, non-speculative basis for projecting profits through 2012. His yearly vehicle production forecasts for the JS41 program came from a reliable industry forecaster upon whom the automakers themselves rely. Several witnesses testified that the automakers rarely, if ever, change production suppliers once production begins, so the jury had a reasoned basis for concluding that Multimatic would have been the cross-car beam supplier throughout the JS41 program’s production run.
The jury also had a sound basis for concluding that Faurecia would have selected Multimatic as its supplier absent a breach of the confidentiality agreement. Sullivan, to repeat, testified that the confidentiality agreement obligated Faurecia to use Multimatic as its supplier if it used Multimatic’s design. The jury also heard testimony that, “within a few months” of November 2004, Faurecia would have no choice but to use Multimatic’s design. ROA Tr. Vol. V at 100. It takes years, according to Sullivan, to develop a production-ready cross-car beam, and Chrysler was nearing its design-freeze deadline for the JS41 program, scheduled to hit showroom floors in July 2006. The jury had a reasonable basis for finding that, by May 2005, when Faurecia breached the confidentiality agreement and used Brown as its cross-car beam supplier, Faurecia had to use Multimatic’s design.
Taking a different tack, Faurecia argues that a Multimatic quote, which Faurecia never accepted, cannot establish Multimatic’s lost profits. No doubt, unaccepted offers and preliminary negotiations generally do not establish the basis for a damages award. See Wisconsin Bridge & Iron Co. v. City of Alpena, 238 Mich. 164, 213 N.W. 93, 95 (1927); ATACS Corp. v. Trans World Commc’ns, Inc., 155 F.3d 659, 670 (3d Cir.1998). But cf. Pelton v. Pelton, 167 Mich.App. 22, 421 N.W.2d 560, 562 (1988) (using an unaccepted offer to value a business during divorce proceedings). And no doubt, in a mine-run case, this rule makes sense because it approaches conjecture, if not enters that forbidden territory, to assume that both parties would agree to one party’s quote or even that both parties would reach a final agreement. See ATACS Corp., 155 F.3d at 670.
But in this case, the jury had more to go on. The evidence showed that Faurecia had little choice by May 2005 but to agree to Multimatic’s terms. It took no speculation for the jury to find that Faurecia and Multimatic would have reached agreement. For if they had not, Faurecia would have forgone the benefits of — and borne the consequences of breaching — its instrument-panel supply agreement with Chrysler. In this setting, where there was a reasonable likelihood that the parties would have consummated a supply agreement but for Faurecia’s breach, Multimatic’s February 2005 quote gives a contemporaneous, non-speculative indication of the price term Multimatic would have imposed from its superior bargaining position. That is particularly so where Faurecia did not even offer its own expert on damages.
Faurecia, lastly, challenges the sufficiency of the evidence to support the jury’s award of $600,515 for unpaid prototype parts. Faurecia waived this argument, however, by not raising it in its Rule 50(b) motion. It makes no difference that Faurecia mentioned that it wished to pre[652]*652serve the issue for appeal in a footnote in its brief. The point of a Rule 50(b) motion is not to tell the district court what the losing party might challenge on appeal but to give the district court a first shot at correcting the mistake — and to give the district court an explanation for doing so. See Bldg. Serv. Local 47 Cleaning Contractors Pension Plan v. Grandview Raceway, 46 F.3d 1392, 1398-99 (6th Cir.1995).
Faurecia argues in the alternative that the district court abused its discretion in making four evidentiary rulings. See Barnes v. City of Cincinnati, 401 F.3d 729, 741 (6th Cir.2005). First, the district court did not exceed its authority in denying Faurecia’s discovery motion to compel disclosure of Multimatic’s costs and profit margins on all cross-car beam programs between 2000 and 2005. Faurecia claims that it could have presented “a substantially reduced lost profits computation” to the jury had Multimatic provided this information. Faurecia Br. at 59. A magistrate judge denied this motion for two reasons: The request was overly broad in time and scope given that Multimatic’s damages experts had not relied on any of this financial information in preparing their reports, and it conflicted with a previous stipulated order withdrawing all of Faurecia’s requests for historical pricing information. The district court agreed. A district court has broad discretion to deny unduly broad discovery requests, see Info-Hold, Inc. v. Sound Merchandising, Inc., 538 F.3d 448, 457 (6th Cir.2008), and this trial judge did not abuse that discretion by denying Faurecia access to five years of tangential, sensitive financial information.
Second, the district court did not abuse its discretion by admitting Faurecia’s November 2004 pre-production letter of intent. According to Faurecia, the letter ceased having any relevance once the district court deemed it unenforceable at the summary judgment stage. Rule 401 sets a low threshold for relevance, however, which is to say, evidence “having any tendency to make the existence of any” relevant fact “more [or less] probable.” Fed.R.Evid. 401. The November 2004 letter clears this threshold because it shows both parties contemplated using the pricing structure Simon relied upon when calculating lost-profits damages, thereby making the lost-profits calculation more plausible. Likewise unavailing is Faurecia’s claim that the letter of intent would confuse the jury, which might treat it as an enforceable agreement. See Fed.R.Evid. 403. The district court instructed the jury that they could not treat the letter of intent as an enforceable agreement and indicated it would give that instruction when it admitted the letter.
Third, Faurecia objects under Rule 702 to the admission of Simon’s expert testimony. To be admissible, expert testimony must be relevant and “have a reliable basis in the knowledge and experience of’ the expert’s “discipline.” Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 592, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). Simon’s testimony is relevant because it went to the only issue before the jury: the amount of damages suffered by Multimatic. No evidence shows that Simon’s methodology, or the way he applied that methodology here, falls outside the range of accepted accounting principles. For the same reasons that Multimatic’s unilateral price quote is not too speculative to support a jury verdict, neither is Simon’s testimony based on that quote too speculative to be admissible under Rule 702. Faurecia separately attempts to incorporate by reference arguments it made in its motion in limine below seeking to exclude Simon’s testimony. Because Faurecia does not develop these arguments on appeal, however, it waived them. See Bickel v. [653]*653Korean Air Lines Co., Ltd., 96 F.3d 151, 153 (6th Cir.1996); Bldg. Serv. Local 47, 46 F.3d at 1398-99.
Fourth, Faurecia likewise failed to preserve its objection to Sullivan’s testimony about his understanding of the confidentiality agreement by not objecting in the district court. See Bowman v. Corrections Corp. of America, 350 F.3d 537, 548 (6th Cir.2003).
C.
Faurecia also claims that the district court abused its discretion by rejecting its Rule 60(b) motion to reduce the lost-profits damage award by roughly $2.2 million. Less than a month after the district court entered its final judgment in October 2007, Chrysler announced it would eliminate a shift at the plant producing the JS41 program vehicles, causing industry forecasters to reduce JS41 production estimates through 2012 by thirty-one percent. Faurecia moved for Rule 60(b)(2) and (b)(6) relief from the judgment based on this information, which the district court denied.
Faurecia does not qualify for Rule 60(b)(2) relief. Under that rule, the district court “may relieve a party ... from a judgment” if it uncovers “newly discovered evidence that, with reasonable diligence, could not have been discovered in time to move for a new trial.” Fed.R.Civ.P. 60(b). The problem is, the newly discovered evidence must relate to facts existing at the time of trial. See Davis v. Jellico Cmty. Hosp. Inc., 912 F.2d 129, 136 (6th Cir.1990). Faurecia offers no evidence that this was so.
Faurecia’s request for Rule 60(b)(6) relief fails as well. District courts may grant relief under this provision only in “extraordinary circumstances.” See Agostini v. Felton, 521 U.S. 203, 239, 117 S.Ct. 1997, 138 L.Ed.2d 391 (1997). Events disproving a damages expert’s forward-looking estimates are not such an extraordinary circumstance, as it is hardly unusual for future events to overtake the reasonable premises of an estimate. See Davis, 912 F.2d at 136 (“Death of a judgment plaintiff following so shortly after a jury award of damages based on an expected life span not realized ... is not the sort of ‘extraordinary circumstances’ contemplated by Rule 60(b)(6).”).
IV.
Multimatic cross-appeals the district court’s decision to exclude testimony by its second damages expert, Dr. Michael Hartzmark. Hartzmark, a PhD economist, would have testified that Multimatic suffered roughly $28.7 million in additional damages beyond the $10 million Sullivan calculated. Hartzmark formed the opinion that Faurecia’s breach of the confidentiality agreement caused three other sources of lost profits: (1) Lost profits from a revamped JS41 program extending from 2013 through 2018; (2) the loss of an “incumbency benefit[ ]” by not supplying the cross-ear beam, which reduced Multimatic’s likelihood of being a Chrysler supplier in the future by eighteen percent, R.144 Ex. A at 4, 21-22; and (3) “price erosion” — requiring decreased profit margins — on future GM and Ford contracts through 2016 because competitors had access to Multimatic’s designs, ROA Tr. Vol. VII at 117,121.
The district court excluded this evidence under Fed.R.Evid. 702 and 403, finding that Hartzmark’s expert opinion rested on speculation and his methodology had not been scientifically tested, had not been subject to peer review and was not generally accepted in the economic community. We review this decision for an abuse of discretion. Gen. Elec. Co. v. Joiner, 522 U.S. 136, 138-39, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997). In ensuring that ex[654]*654pert testimony is “relevant and reliable” under Rule 702, Kumho Tire Co. Ltd. v. Carmichael, 526 U.S. 137, 141, 152, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999), district courts must ensure that the testimony has a “basis in the knowledge and experience of’ the expert’s discipline and that the expert exhibits “the same level of intellectual rigor” expected of an expert outside the courtroom. Id. at 149, 152.
Consistent with this gatekeeping requirement, the district court properly excluded Hartzmark’s testimony about lost JS41-related profits from 2013 through 2017. See McLean v. 988011 Ontario, Ltd., 224 F.3d 797, 800-01 (6th Cir.2000). Hartzmark assumed that Chrysler would follow standard automotive practice by revamping the JS41 program after five years and then extend its production run through 2017. He further assumed, without any supporting industry research, that Chrysler would continue producing JS41 vehicles at the same rate in 2017 as it did in 2007. Simon, in contrast, relied on industry research for his 2007-12 lost-profits calculations. The district court hit the nail on the head when it characterized this ten-year prediction about the fortunes of the American automotive industry as utterly speculative.
Hartzmark’s opinion regarding Multimatic’s loss of its “incumbency benefit[ ]” fares no better. R.144 Ex. A at 21. His opinion rests on speculation about Multimatic’s profit margins on yet-to-be-signed contracts years into the future, using Simon’s per-beam profit estimates for the JS41 program without analyzing Multimatic’s historical profit margins or typical industry profit margins. As mentioned, however, Simon’s calculations rested on an unaccepted price quote. That quote did not rest on unacceptable speculation only because Multimatic had Faurecia in a position where it had no choice but to accept Multimatic’s terms. Multimatic will not have similar bargaining strength for future Chrysler contracts, and in the absence of that reality we cannot fault the district court for excluding Hartzmark’s testimony.
Hartzmark builds his “price erosion” opinion on an even shakier foundation. According to Hartzmark, Faurecia’s breach of the confidentiality agreement eroded Multimatic’s future profits on Ford and GM contracts by eight percent. He derived this number from the same Multimatic quote as his incumbency-benefit calculation as well as May 2005 quotes from Brown and other suppliers. His opinion also assumed, relying on Multimatic’s say-so, that Multimatic provides 2.15 million cross-car beams annually to Ford and GM and that Multimatic would continue to do so through 2016. Experts may not assume facts without some support for those assumptions in their expert report or elsewhere in the record. See McLean, 224 F.3d at 801.
Beyond being speculative, Hartzmark’s lost profit calculations fall short of the level of rigor professional economists normally exercise. See Kumho Tire Co., 526 U.S. at 152, 119 S.Ct. 1167. We doubt— and Multimatic introduced no evidence to the contrary — that an economic projection ten years into the future would be well-received by professional economists if no historical data backed up the projection and the economist made no effort to test his model against historical data to confirm its predictive power. Hartzmark’s price erosion calculation also never attempts to identify or control for factors beyond the disclosure of Multimatic’s designs that might account for the possibility that competitors might submit lower bids. (Some of the bidders, for example, were able to obtain steel at lower prices.)
Multimatic bears some of the blame for Hartzmark’s problematic analysis. He asked for historical data, but Multimatic [655]*655refused to provide it, likely because the company did not want to disclose it during discovery. District courts do not abuse their discretion, however, when they exclude expert testimony whose flaws stem in part from the expert’s efforts to do the best job he could with the limited data his client would provide.
Multimatic responds that price erosion is a well-accepted economic principle and that weaknesses in Hartzmark’s opinion go to weight, not admissibility. But the question at hand is the reliability of Hartzmark’s application of price erosion principles, not the reliability of price erosion as an economic principle. See Kumho Tire Co., 526 U.S. at 152, 119 S.Ct. 1167. Perceived flaws in an expert’s opinion go to weight only if they fall within the accepted norms of the discipline and have a non-speculative basis in fact. See McLean, 224 F.3d at 800-01. Not so here.
V.
For these reasons, we affirm.