TAMPA BAY WATER, a regional water supply authority, Plaintiff-Appellant, v. HDR ENGINEERING, INC., a Nebraska corporation, Defendant-Cross Defendant-Appellee, 731 F.3d 1171

MARCUS, Circuit Judge: Appellant Tampa Bay Water (“TBW”) challenges three pretrial rulings that preceded its defeat at trial in this diversity case. The subject matter of the lawsuit was a reservoir, owned by TBW, which developed large cracks in its earthen embankments shortly after construction. TBW sued several defendants, including HDR Engineering (“HDR”) and Barnard Construction Company (“Barnard”), alleging that HDR defectively designed the reservoir and that Barnard defectively constructed it.

The amount in controversy was very large; TBW’s initial settlement demands exceeded $200 million, and it ultimately sought more than $100 million in damages at trial. Prior to trial, TBW settled with Barnard and entered into a series of factual stipulations that essentially absolved Barnard of liability. On the basis of those stipulations, the district court granted Barnard

summary judgment, leaving only TBW and HDR to go to trial. TBW argued that HDR’s defective design caused the reservoir damage. HDR presented an alternative theory to the jury, alleging that Barnard’s construction techniques caused the damage. The jury found in HDR’s favor, leading to this appeal. The primary issue on appeal concerns the district court’s decision to allow HDR to present evidence that Barnard caused the reservoir damage. According to TBW, both Florida’s comparative negligence statute and estoppel principles should have barred the introduction of this evidence, since Barnard had been absolved of liability to TBW by the district court’s grant of

summary judgment in Barnard’s favor. After thorough review, we conclude that the district court properly denied TBW’s motion to exclude HDR’s evidence. TBW advanced one theory of causation against Barnard, which it expressly disavowed as part of its settlement agreement with Barnard. When the district court granted

summary judgment to Barnard, therefore, the court addressed and rejected only that theory. The district court had no occasion to adjudicate the merits of HDR’s alternative theory of causation. Thus, the grant of

summary judgment did not estop HDR from presenting that theory and the surrounding facts to the jury. TBW also claims that the district court improperly admitted HDR’s expert’s testimony despite the fact that his methodology was unreliable and never peer-reviewed, and that the district court wrongly denied TBW leave to amend its complaint for a second time after the close of discovery. The district court did not abuse its discretion in either respect and, therefore, we affirm its judgment. I. A. Tampa Bay Water is a regional water supply authority that provides drinking water for the Tampa Bay area.

In November 1998, TBW and HDR entered into a contract for the design of a large reservoir project. TBW also contracted with Barnard to build the reservoir based on HDR’s design. Because water in an earthen reservoir may seep away, HDR’s design included the use of plastic sheeting called a “geomembrane” within the embankment. A protective layer of two to three feet of soil was placed over the geo-membrane. On top of that layer, the inner slope of the reservoir consisted of “soil cement,” a mixture of soil and cement designed to prevent erosion of the reservoir’s inner walls.

The reservoir began operating in June 2005. By late 2006, two areas of the reservoir’s embankment — one in the northeast quadrant, one in the southwest — exhibited large cracks in the soil cement. TBW contacted HDR and asked it to investigate the cracks. According to TBW, HDR used its investigation to figure out a way to escape its own liability and to produce evidence of other parties’ fault. Eventually, TBW fired HDR and tasked its system engineer, Black & Veatch, with completing the investigation. In December 2008, TBW sued HDR and Barnard in the United States District Court for the Middle District of Florida.

The complaint alleged that HDR’s design was defective because of the failure to account for excess pore pressure. The complaint also alleged that Barnard had negligently constructed the reservoir with improperly blended soil, and that HDR had negligently performed its field inspection and quality-control duties during construction. In November 2009, TBW filed an amended complaint. The parties advanced several different theories regarding the cause of the cracking.

One of TBW’s theories was that the cracking was caused by excess pore pressure (“the excess pore pressure theory”), which meant that the layer of soil between the soil cement and the geomembrane was trapping excess water. The excess water exerted pressure on the soil cement and pushed it up, leading to the cracks. This theory put the blame squarely on HDR, since a different design could have avoided this problem. Initially, TBW also asserted that Barnard had used improperly blended soil in the intermediate layer.

The soil then created “lenses, pockets, streaks, and layers,” which permitted the excess pore pressure to develop (“the lenses and pockets theory”). HDR’s theory was that the cracking was not caused by the soil cement being pushed up but rather by the cement collapsing. According to HDR, the protective soil layer on top of the geomembrane was too thick, too loose, and too dry in the two areas where the soil cement cracked. When the soil became saturated with water, it became denser and lost volume, causing the soil cement to collapse and crack (“the collapse upon wetting theory”).

This theory placed the blame on Barnard. Notably, TBW never asserted the collapse upon wetting theory as an alternative theory of causation against Barnard. B. Three of the district court’s pretrial rulings are at issue in this appeal. In December 2010, after discovery closed, TBW filed a timely motion requesting leave to amend its complaint for a second time. Among other things, TBW wished to add claims for professional negligence and breach of an implied covenant of good faith and fair dealing based on HDR’s post-cracking investigation.

The district court granted the motion in part, allowing TBW to bolster its allegations in Count One (concerning HDR’s defective design), but denied leave to amend to add the two new claims pertaining to HDR’s post-cracking investigation. The district court found that TBW knew of the factual basis for both of those claims well before December 2010 and had unduly delayed bringing the motion. The second contested ruling concerned HDR’s expert, Dr. Bromwell.

TBW moved to exclude Bromwell because his testimony relied on using results from a test called “ASTM D 4767” to determine whether the reservoir’s protective soil layer was susceptible to collapse upon wetting. According to TBW, ASTM D 4767 was not a reliable or widely accepted method for measuring collapse potential. The district court denied TBW’s motion and permitted Bromwell to testify for two reasons. First, HDR had produced several peer-reviewed articles that suggested that triaxial tests like ASTM D 4767 could be used to determine the collapse potential of a soil sample.

Second, Bromwell had explained in his deposition how he had derived his assessment of the collapse potential from the ASTM D 4767 data. The final contested ruling requires some more background. In March 2011, TBW and Barnard entered into a high-low settlement agreement, which obligated Barnard to pay a minimum of $750,000 and capped its maximum exposure at $5,000,000. The agreement also contained an extensive set of stipulated facts that essentially absolved Barnard of liability.

The agreement did not dispose of TBW’s claims against Barnard, however, and in fact required Barnard to stand trial and call certain witnesses and prevented Barnard from calling others. The ultimate thrust of the agreement was that Barnard would remain a nominal defendant at trial but would (like TBW) blame HDR for the damage. HDR moved to void the settlement agreement as collusive or, in the alternative, suggested that the district court enter

summary judgment in Barnard’s favor so that Barnard could not participate at trial. The district court ordered TBW to show cause why

summary judgment should not be entered in favor of Barnard. TBW argued that, although the stipulations disavowed Barnard’s responsibility under the excess pore pressure theory, TBW could still pursue Barnard under its lenses and pockets theory. The district court granted Barnard

summary judgment and found it not liable to TBW. To start with, the court noted that TBW had abandoned any claim against Barnard not based upon the lenses and pockets theory. However, TBW and Barnard expressly stipulated that the lenses and pockets in the soil were not the cause of the damage. Moreover, they further stipulated that HDR, not Barnard, was responsible for the existence of the lenses and pockets. The district court concluded, therefore, that TBWs sole theory of Barnard’s liability was “foreclosed by its own admissions.” Since the district court granted

summary judgment to Barnard, it did not decide on the validity of the settlement agreement. After the district court granted Barnard

summary judgment, HDR moved to include Barnard as a so-called Fabre defendant, which would allow the jury to apportion responsibility for the damages between HDR and Barnard under Florida’s comparative negligence statute. The district court denied HDR’s motion based on Florida cases holding that a former defendant cannot be treated as a Fabre defendant after

summary judgment has been entered in its favor. According to the district court,

summary judgment had resolved the question of Barnard’s liability once and for all. These twists and turns brought about the final contested ruling. TBW moved in limine to exclude all evidence of HDR’s collapse upon wetting theory, based on the fact that the district court had granted Barnard

summary judgment and then denied HDR’s request to include Barnard as a Fabre defendant. TBW raised three grounds in its motion: Florida’s comparative negligence law, federal collateral es-toppel, and state judicial estoppel. The district court denied TBW’s motion. The district court reasoned that, at bottom, the issue was one of collateral estop-pel: could HDR introduce evidence that Barnard had caused the damage under the collapse upon wetting theory, when the court had determined that Barnard was not liable?

In analyzing the requirements for applying estoppel, the district court determined that the issues were not identical, and that HDR’s collapse upon wetting theory had never been actually litigated. The

summary judgment proceeding disposed only of TBW’s lenses and pockets theory and never had occasion to consider or determine the merits of HDR’s collapse upon wetting theory. The court also drew a distinction between liability and causation and concluded that its

summary judgment order had adjudicated the former but not the latter. In other words, while Barnard could no longer be found liable to TBW, this was not equivalent to a determination that Barnard’s construction techniques had not caused the damage to the reservoir. The district court therefore held that HDR was not precluded from presenting its theory “to the jury from the perspective of causation as opposed to fault.” The case proceeded to trial. TBW argued that HDR was liable based on its excess pore pressure theory, while HDR argued that the cause of the damage was collapse upon wetting.

To this end, HDR put on Dr. Bromwell as its main expert. The jury returned a verdict of no liability. II. We review the

interpretation of state law de novo, FTC v. Leshin, 618 F.3d 1221, 1231 (11th Cir.2010), mindful that, “[a]s a federal court sitting in diversity, we are required to apply the law as declared by the state’s highest court,” CSX Transp., Inc. v. Trism Specialized Carriers, Inc., 182 F.3d 788, 790 (11th Cir.1999) (per curiam). In the absence of authority on point, “we follow relevant decisions of Florida’s intermediate appellate courts” and attempt “to determine the issues of state law as we believe the Florida Supreme Court would.” State Farm Fire & Cas. Co. v. Steinberg, 393 F.3d 1226, 1231 (11th Cir.2004).

We review the applicability of direct or collateral estoppel principles de novo. United States v. Weiss, 467 F.3d 1300, 1308 (11th Cir.2006). However, we review the district court’s application of judicial estoppel for abuse of discretion. Robinson v. Tyson Foods, Inc., 595 F.3d 1269, 1273 (11th Cir.2010). We review the district court’s decision to admit or exclude expert testimony for abuse of discretion. Gen. Elec. Co. v. Joiner, 522 U.S. 136, 142-43, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997). Finally, we also review the district court’s denial of leave to amend for abuse of discretion. Burger King Corp. v.

Weaver, 169 F.3d 1310, 1315 (11th Cir.1999). A. We first address TBW’s attempt to exclude HDR’s collapse upon wetting theory, since this issue warrants the most extended discussion. TBW’s position is that HDR’s collapse upon wetting theory blamed Barnard for causing the damage, but the district court’s grant of

summary judgment in Barnard’s favor had already adjudicated Barnard’s fault. Thus, in the first place, Florida’s comparative negligence law would have barred the presentation of this evidence. Moreover, collateral estoppel barred HDR from introducing evidence against Barnard when the district court’s grant of

summary judgment already had exonerated Barnard. Finally, judicial estoppel barred HDR from asserting that Barnard was the cause of the damage when HDR had taken a contradictory position earlier in this lawsuit by not opposing the grant of

summary judgment. HDR initially responds that TBW waived this entire claim by failing to object to the introduction of its causation evidence at trial. According to HDR, the district court’s preliminary ruling on a motion in limine could not preserve this claim for appeal. However, under the Federal Rules of Evidence, it is no longer necessary for a party to renew an objection to evidence when the district court has definitively ruled on the party’s motion in limine.

In relevant part, Rule 103 provides, “Once the court rules definitively on the record — either before or at trial — a party need not renew an objection or offer of proof to preserve a claim of error for appeal.” Fed.R.Evid. 103(b). As the commentary explains, “The amendment applies to all rulings on evidence whether they occur at or before trial, including so-called ‘in limine’ rulings.” Fed.R.Evid. 103 advisory committee’s note. Thus, we review claims of error raised and decided in an in limine motion as long as the district court’s ruling was definitive. See, e.g., Proctor v.

Fluor Enters., Inc., 494 F.3d 1337, 1349-50 (11th Cir.2007). Because the district court’s ruling in this case was sufficiently definitive, we will consider the merits of TBWs position. TBW begins by arguing that fundamental principles of Florida’s comparative negligence law should have precluded HDR from introducing its evidence against Barnard. Florida’s comparative negligence law renders a party liable only for the share of total damages proportional to its fault. Fla. Stat. § 768.81(3). Under Florida law, once a court has entered

summary judgment in one defendant’s favor, that former defendant can no longer be included on the verdict form as a so-called Fabre defendant for purposes of apportioning fault to that defendant. See, e.g., Crowell v. Kaufmann, 845 So.2d 325, 327 (Fla. 2d DCA 2003). TBW argues for an extension of this Crowell principle: if a former defendant cannot be included on the verdict form, then no evidence attributing the cause of the damages to that defendant should be admitted either.

Although TBW asserts that the Crowell principle is a rule of Florida’s substantive comparative negligence law (and, impliedly, that an extension of the Crowell principle barring the introduction of evidence would also be a substantive rule), it is not. As Florida’s Third District Court of Appeals made clear in Southern Bell Telephone & Telegraph Co. v. Florida Department of Transportation, the very case cited by Crowell, [I]f the Fabre defendant is exonerated because there is no evidence of fault, that defendant does not go on the verdict form.

Here, because the trial judge determined as a matter of law that DOT was not at fault, ... Southern Bell would not be entitled to have DOT placed on the verdict form. Southern Bell and DOT are bound by the lower court’s judgment.... Southern Bell would be collaterally estopped from asserting DOT’s fault, either for the purpose of seeking contribution or for the purpose of having DOT appear on the verdict form in the underlying action so as to offset its liability. 668 So.2d 1039, 1041 (Fla. 3d DCA 1996) (original emphasis omitted and emphasis added).

In short, this rule does not stem from Florida’s comparative negligence statute but rather is merely an application of Florida’s state collateral estoppel doctrine. In asking us to apply and extend Crowell and Southern Bell, and by treating this issue as one of state substantive law, TBW implicitly is asserting that we should apply Florida’s collateral estoppel law. Yet, as this Court has unambiguously held, “federal preclusion principles apply to pri- or federal decisions, whether previously decided in diversity or federal question jurisdiction.” CSX Transp., Inc. v.

Bhd. of Maint. of Way Emps., 327 F.3d 1309, 1316 (11th Cir.2003); see also Precision Air Parts, Inc. v. Avco Corp., 736 F.2d 1499, 1503 (11th Cir.1984) (“When a federal court sitting in diversity examines the collateral estoppel or res judicata effect of a prior federal judgment, based either on diversity or a federal question, it must apply federal common law.”). Citibank, N.A. v. Data Lease Financial Corp., 904 F.2d 1498 (11th Cir.1990), provides an instructive example.

In that diversity case, Citibank sued Data Lease, which then filed several counterclaims against Citibank and third-party defendants who had been directors of a Citibank subsidiary. Id. at 1499. Data Lease and the directors entered into a settlement agreement that specifically stated that Data Lease would dismiss its claims against the directors while reserving its claims against Citibank, and the district court dismissed the counterclaims against the directors based on that agreement. Id. at 1500.

The district court then also dismissed Data Lease’s counterclaims against Citibank, notwithstanding the fact that the settlement agreement did not purport to dispose of those claims, because Florida law established that a principal cannot be held liable when its agent is exonerated. Id. A panel of this Court affirmed but did not rely on Florida’s preclusion rules.

Rather, the panel began with the proposition that, “even in a case which rests its subject matter jurisdiction solely upon diversity of citizenship, a federal court must apply federal law to determine the preclusive effect of a prior federal court decision.” Id. at 1501. The panel then applied the federal common law rule that “a judgment exonerating a servant or agent from liability bars a subsequent suit on the same cause of action against the master or principal based solely on respon-deat superior,” and affirmed the district court’s dismissal of Data Lease’s counterclaims. Id. at 1502.

Just as in Citibank, we must apply federal collateral estoppel law to determine the preclusive effect of the district court’s grant of

summary judgment to Barnard. For this reason, TBW’s first argument lacks merit, and there is no need to certify any question of law to the Florida Supreme Court. As for whether HDR should have been estopped under federal law, the rule for applying either direct or collateral estoppel is well-established.

A party asking the court to apply estoppel must establish that “(1) the issue at stake is identical to the one involved” in the earlier proceeding; “(2) the issue was actually litigated” in the earlier proceeding; “(3) the determination of the issue ... must have been a critical and necessary part” of the earlier judgment; and “(4) the ■ party against whom collateral estoppel is asserted must have had a full and fair opportunity to litigate the issue.” Dana v. E.S. Originals, Inc., 342 F.3d 1320, 1323 (Fed.Cir.2003) (internal quotation marks omitted).

If established, then “estoppel operates to bar the introduction or argumentation of certain facts necessarily established in a prior proceeding.” United States v. Lee, 622 F.2d 787, 790 (5th Cir.1980). The problem with TBW’s argument is that the district court never had occasion to consider HDR’s collapse upon wetting theory of causation when adjudicating the

summary judgment proceeding between TBW and Barnard. TBW had abandoned all theories of causation against Barnard except the lenses and pockets theory, and the district court found that theory and that theory alone foreclosed by TBW and Barnard’s factual stipulations. Estoppel did not bar HDR from presenting its collapse upon wetting theory when the

summary judgment did not consider, address, or in any way depend upon deciding the validity of that theory. Indeed, as the district court accurately explained, none of the first three elements for estoppel were established. In the first place, whether collapse upon wetting caused the reservoir damage was not at issue in the

summary judgment proceedings. The question there was not whether Barnard’s construction techniques had caused the damage to the reservoir under any possible theory of causation. As the district court explained, its

summary judgment order addressed only the much narrower question of whether Barnard could be liable to TBW under TBW’s own distinct theory of the case — the lenses and pockets theory — and in light of the settlement agreement’s stipulated facts. Holding TBW and Barnard to their stipulations, the district court found that there were no disputed facts between those two parties, and that Barnard could not possibly be liable to TBW under TBW’s lenses and pockets theory. The

summary judgment was necessarily limited because causation, as a factual question, and liability to a particular party, as a legal question, are distinct. The district court properly distinguished between the two concepts while explaining its ruling. Under Florida law, liability necessarily encompasses four elements: (1) a duty of care; (2) a breach of that duty; (3) causation; and (4) actual loss or damages. Curd v. Mosaic Fertilizer, LLC, 39 So.3d 1216, 1227 (Fla.2010). The district court’s

summary judgment meant that Barnard was “found not liable as a matter of law,” which implies only that TBW “could not establish an essential element” of its particular claim against Barnard. Price v. Beker, 629 So.2d 911, 912 (Fla. 4th DCA 1993). Since TBW advanced only the lenses and pockets theory of causation, and those parties’ stipulated facts stated that the presence of lenses and pockets had not caused the reservoir damage, TBW could never establish the causation element of the limited claim it chose to pursue.

HDR’s collapse upon wetting theory sought to prove that Barnard caused the reservoir damage in an entirely different way. If Barnard had used thick, dry, and loose soil, the embankment would have collapsed, causing the reservoir damage. This theory attributed the reservoir damage to a different flaw in Barnard’s construction technique than the one initially alleged by TBW and necessitated proof of a wholly different set of facts. Thus, TBW could not even meet the first requirement for applying estoppel.

See In re McWhorter, 887 F.2d 1564, 1568 (11th Cir.1989) (per curiam) (issues in two proceedings not identical for estoppel purposes if there is a different or “further inference of fact ... required”). The district court also properly concluded, for much the same reasons, that HDR’s collapse upon wetting theory was never actually litigated, nor was it a critical and necessary part of the

summary judgment. As both this Court and its predecessor Court have long recognized, “issue preclusion bars the relitigation of issues actually adjudicated, and essential to the judgment, in a prior litigation.” Kaspar Wire Works, Inc. v. Leco Eng’g & Mach, Inc., 575 F.2d 530, 535-36 (5th Cir.1978). However, it is “insufficient for the invocation of issue preclusion that some question of fact or law in a later suit was relevant to a prior adjudication between the parties; the contested issue must have been litigated and necessary to the judgment earlier rendered.” Id. at 536.

TBW never alleged that collapse upon wetting had caused the reservoir damage. Thus, TBW and Barnard never joined issue on whether collapse upon wetting had occurred and, if so, who was responsible. Moreover, the judgment in Barnard’s favor was not based on a factfinder’s determination of causation after an adversarial proceeding but rather on a set of joint stipulations.

Plainly, a set of stipulations does not “reflect the considered judgment of a judicial officer” or a jury; “it has been forged by [the parties] alone as an adjustment of conflicting claims and is not a tempered determination of fact and law after the annealment of an adversary trial.” Id. at 538. For this reason, we cannot treat the factual issue of causation as actually litigated or, in this case, necessary to the judgment. We therefore hold that, in the procedural posture of this case, the

summary judgment granted to Barnard did not directly estop HDR from introducing evidence at trial that Barnard caused the damage. Finally, TBW argues that judicial estoppel should have barred HDR’s collapse upon wetting theory. We remain unpersuaded. Judicial estoppel is “designed to prevent parties from making a mockery of justice by inconsistent pleadings.” McKinnon v. Blue Cross & Blue Shield of Ala., 935 F.2d 1187, 1192 (11th Cir.1991) (internal quotation mark omitted). While judicial estoppel “cannot be reduced to a precise formula or test,” Zedner v.

United States, 547 U.S. 489, 504, 126 S.Ct. 1976, 164 L.Ed.2d 749 (2006), three factors typically inform the inquiry: (1) whether there is a clear inconsistency between the earlier position and the later position; (2) a party’s success in convincing a court of the earlier position, so that judicial acceptance of the inconsistent later position would create the perception that either the earlier or later court was misled; and (3) whether the inconsistent later position would unfairly prejudice the opposing party. See Robinson, 595 F.3d at 1273.

The district court did not abuse its discretion when it concluded that this doctrine did not bar HDR’s collapse upon wetting theory. To begin with, there is no inconsistency between HDR’s positions during the

summary judgment proceeding and at trial. HDR did not oppose the court granting

summary judgment to Barnard. However, this was because TBW had stipulated away Barnard’s fault, at least with regard to TBW’s theory of the case. At no point in the proceedings did HDR disavow its collapse upon wetting theory or assert that Barnard was not the cause of the damage to the reservoir. In addition, the district court was never misled by HDR’s position; it did not grant the

summary judgment based on any factual finding that Barnard did not cause collapse upon wetting. Finally, HDR’s collapse upon wetting theory did not unfairly prejudice TBW. TBW’s only assertion to that effect is that, if Barnard had remained in the case, the jury could have assessed damages against Barnard. But it was TBW’s stipulations, not anything that HDR did, that foreclosed the possibility of Barnard remaining in the case as a defendant. B. TBW also claims that the district court abused its considerable discretion by admitting Dr. Bromwell’s expert testimony.

Fed.R.Evid. 702 provides that: A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if: (

a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (

b) the testimony is based on sufficient facts or data; (

c) the testimony is the product of reliable principles and methods; and (

d) the expert has reliably applied the principles and methods to the facts of the case. As the Supreme Court recognized in Daubert v. Merrell Dow Pharmaceuticals, Inc., Rule 702 contemplates that the district court will serve as a gatekeeper to the admission of scientific testimony. 509 U.S. 579, 589, 118 S.Ct. 2786, 125 L.Ed.2d 469 (1993); see also McCorvey v. Baxter Healthcare Corp., 298 F.3d 1253, 1256 (11th Cir.2002).

Under Daubert and its progeny, we conduct a three-part inquiry to determine the admissibility of expert testimony, weighing whether: (1) [T]he expert is qualified to testify competently regarding the matters he intends to address; (2) the methodology by which the expert reaches his conclusions is sufficiently reliable as determined by the sort of inquiry mandated in Daubert', and (3) the testimony assists the trier of fact, through the application of scientific, technical, or specialized expertise, to understand the evidence or to determine a fact in issue. City of Tuscaloosa v.

Harcros Chems., Inc., 158 F.3d 548, 562 (11th Cir.1998) (footnote omitted) (citing Daubert, 509 U.S. at 589, 113 S.Ct. 2786). Since we review the district court’s determinations on every prong of the Dau-bert inquiry only for abuse of discretion, we defer to the district court’s judgments and will not reverse unless the district court’s ruling was “manifestly erroneous.” Joiner, 522 U.S. at 142, 118 S.Ct. 512. “[I]t is by now axiomatic that a district court enjoys considerable leeway in making these determinations.” United States v.

Frazier, 387 F.3d 1244, 1258 (11th Cir.2004) (en banc) (internal quotation marks omitted). Indeed, under this standard of review, “there will be occasions in which we affirm the district court even though we would have gone the other way had it been our call.” In re Rasbury, 24 F.3d 159, 168 (11th Cir.1994). TBW’s attack on Dr. Bromwell addresses only Daubert’s second prong — whether his methodology was sufficiently reliable under Daubert. Bromwell used a preexisting set of data, produced by ASTM D 4767 testing of soil samples from the embankment prior to the completion of construction.

ASTM D 4767 provided only a single set of data quantifying the volume and density change in a soil sample that has been subjected to both water saturation and loading, but, according to TBW, the proper test would have measured the volume and density change due to water saturation alone. Another test, ASTM D 5333, is a test that applies only water saturation and evaluates the collapse potential of soil. Since Bromwell failed to use ASTM D 5333 data, the argument goes, his conclusion that the soil collapse was caused by wetting was not produced by a reliable methodology.

Under Daubert, the reliability of the expert’s methodology is a context-specific inquiry. The Supreme Court has instructed district courts to weigh four factors in making their decision: (1) whether the expert’s methodology has been tested or is capable of being tested; (2) whether the theory or technique used by the expert has been subjected to peer review and publication; (3) whether there is a known or potential error rate of the methodology; and (4) whether the technique has been generally accepted in the relevant scientific community. United Fire & Cas. Co. v.

Whirlpool Corp., 704 F.3d 1338, 1341 (11th Cir.2013) (per curiam) (citing Daubert, 509 U.S. at 593-94, 113 S.Ct. 2786). “At the same time, the Court has emphasized that these factors are not exhaustive and are intended to be applied in a ‘flexible’ manner.” Id. (quoting Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999)). Thus, for example, the absence of publications on a given methodology does not mean that an expert’s testimony based on that methodology is inadmissible.

See Daubert, 509 U.S. at 593, 113 S.Ct. 2786 (“Publication (which is but one element of peer review) is not a sine qua non of admissibility; it does not necessarily correlate with reliability....”); United Fire & Cas. Co., 704 F.3d at 1342-43 (finding that expert testimony should have been admitted despite the lack of publications supporting the expert’s findings). We cannot find that the district court’s admission of Bromwell’s testimony was manifestly erroneous or an abuse of discretion.

TBW’s primary objection is that only one testing method' — -ASTM D 5333 — is the accepted method for evaluating collapse potential, and Bromwell failed to use that method. Yet the district court had some basis for evaluating the relative reliability of Bromwell’s method. An

article which TBW submitted to the district court states that, in measuring collapse potential, “the most critical conditions are the degree of saturation and dry density of the soil and the total overburden pressure,” and that “[t]o accommodate these three variables, collapse tests are typically conducted utilizing standard oedometric or triaxial equipment.” The

article also notes that, while different triaxial tests produce different results, they “generally indicate only small differences in the predicted magnitudes of collapse.” TBW does not dispute that ASTM D 4767 is a triaxial test. Thus, although perhaps not ideal, Bromwell’s data was produced by a recognized triaxial testing method whose protocol is widely accepted in the scientific community. At the very least, TBW does not dispute that the ASTM D 4767 protocol accurately measured the volume and density change of the soil sample.

As for the main flaw in the use of ASTM D 4767 — the fact that it measured the volume and density change that occurred under the influence of two variables, load and water saturation — Bromwell had multiple measurements from the same soil samples under different loads but the same water saturation. Using those points of comparison, he could extrapolate to what extent the volume and density change was attributable to each variable. Since the changes in load resulted in only small differences, Bromwell concluded that the ASTM D 4767 test effectively measured the change in volume and density due to water saturation.

Bromwell’s extrapolation relies on mathematics, not unscientific supposition. The district court’s acceptance of this explanation fell well within the range of choice it possessed as the gatekeeper for expert testimony. TBW levels two additional objections. First, TBW’s own expert, Dr. Brumund, testified by affidavit that Bromwell’s methods were unreliable. But this kind of disagreement between experts ordinarily goes to the credibility of expert testimony, not its admissibility, and is the province of the jury. See Rosenfeld v.

Oceania Cruises, Inc., 654 F.3d 1190, 1193 (11th Cir.2011) (“[I]n most cases, objections to the inadequacies of a study are more appropriately considered an objection going to the weight of the evidence rather than its admissibility.” (internal quotation marks omitted)). In Quiet Technology DC-8, Inc. v. HurelDubois UK Ltd., for example, one objection to the expert’s testimony was that his modeling failed to include all variables and therefore could not produce a “meaningful correlation of flight test results with computer results.” 326 F.3d 1333, 1346 (11th Cir.2003).

Quoting the Supreme Court, we stated, “ ‘[normally, failure to include variables will affect the analysis’ probativeness, not its admissibility.’” Id. (alteration in original) (quoting Bazemore v. Friday, 478 U.S. 385, 400, 106 S.Ct. 3000, 92 L.Ed.2d 315 (1986)). The same applies here: Brumund’s and Brom-well’s disagreements could be aired out in front of the jury and tested by the crucible of cross-examination. See id. at 1345 (“The identification of such flaws in generally reliable scientific evidence is precisely the role of cross-examination.”).

TBW also asserts that HDR’s sole argument for reliability was an invocation of Bromwell’s impressive credentials. Yet Bromwell’s explanation of his method was a logical one, not a mere assertion of his credentials. Moreover, it was not improper for the district court to consider Brom-well’s credentials in evaluating whether his methodology was reliable. Although an expert’s qualifications go primarily to the first prong of Daubert’s inquiry, “an expert’s overwhelming qualifications may bear on the reliability of his proffered testimony” even if “they are by no means a guarantor of reliability.” Id. at 1341.

In this ease, Bromwell received both his bachelor’s degree, in Chemical Engineering, and his doctorate, in Civil Engineering, from MIT. He spent his academic career authoring or coauthoring dozens of articles in the area of soil science, and his nonacademic career was devoted to the design, analysis, and evaluation of earthen embankments. In fact, Bromwell had designed thirty to forty earthen embankments, most of which were reservoirs or similar projects.

These qualifications bolstered HDR’s showing of reliability, and the district court committed no manifest error by allowing the jury to hear Brom-well’s testimony. C. Finally, the district court did not abuse its discretion in denying TBW’s motion for leave to amend its complaint for a second time.

TBW claims that the district court applied the incorrect legal standard: a “distinct disfavor” standard found in the Middle District of Florida’s Local Rules rather than the permissive standard found in Fed.R.Civ.P. 15(a)(2). “A district court abuses its discretion if it applies an incorrect legal standard [or] follows improper procedures in making a determination .... ” Citizens for Police Accountability Political Comm. v. Browning, 572 F.3d 1213, 1216-17 (11th Cir.2009).

In this case, the district court correctly identified the relevant legal standard drawn from Rule 15, which states, “[A] party may amend its pleading [if later than twenty-one days after serving it] only with the opposing party’s written consent or the court’s leave. The court should freely give leave when justice so requires.” Fed. R.Civ.P. 15(a)(2). The relevant local rule, which the district court also cited, states, “It is the goal of the court that a trial will be conducted in all Track Two case within two years after the filing of the complaint. ...

A motion to amend any pleading ... is distinctly disfavored after entry of the Case Management and Scheduling Order.” M.D. Fla. Local Rule 3.05(c)(2)(E) (emphasis added). We invalidate local rules when their application is incompatible with the Federal Rules of Civil Procedure. See Fed.R.Civ.P. 83(a) (“A local rule must be consistent with — but not duplicate — -federal statutes and rales adopted under 28 U.S.C. §§ 2072 and 2075....”); Jackson v. Crosby, 375 F.3d 1291, 1296 (11th Cir.2004).

However, when the district court cites both rules but applies the controlling federal standard, we need not consider the local rule’s theoretical or possible inconsistency. See Dunlap v. Transamerica Occidental Life Ins. Co., 858 F.2d 629, 632 (11th Cir.1988) (per curiam). In this case, although Rule 15 and the local rule exhibit some tension with one another, the district court applied the federal standard.

We know this because the district court cited several Rule 15 cases in determining that TBW had unduly delayed moving for leave to amend-a reason both Supreme Court and this Court’s case law have expressly identified as a valid reason to deny leave. See, e.g., Foman v. Davis, 371 U.S. 178, 182, 83 S.Ct. 227, 9 L.Ed.2d 222 (1962); Nat'l Serv. Indus., Inc. v. Vafla Corp., 694 F.2d 246, 249 (11th Cir.1982). TBW next disputes the district court’s finding of undue delay, which was the ground for denying TBW leave to add two new counts to its complaint.

A district court may find undue delay when the movant knew of facts supporting the new claim long before the movant requested leave to amend, and amendment would further delay the proceedings. See, e.g., Campbell v. Emory Clinic, 166 F.3d 1157, 1162 (11th Cir.1999) (“The facts upon which the claims ... were based were available at the time the complaints were filed.”); Nat'l Serv. Indus., Inc., 694 F.2d at 249.

While “[t]he lengthy nature of litigation, without any other evidence of prejudice to the defendants or bad faith on the part of the plaintiffs, does not justify denying the plaintiffs the opportunity to amend their complaint,” Bryant v. Dupree, 252 F.3d 1161, 1164 (11th Cir.2001) (per curiam), prejudice “is especially likely to exist if the amendment involves new theories of recovery or would require additional discovery,” 3 James Wm. Moore et al., Moore’s Federal Practice ¶ 15.15[2]. TBW has failed to meaningfully rebut the district court’s factual findings, which TBW must demonstrate were clearly erroneous.

See Doe v. Cassel, 403 F.3d 986, 990-91 (8th Cir.2005) (per curiam). The district court concluded that TBW knew many of the facts supporting its two new claims — which both concerned HDR’s post-cracking investigation — before TBW filed either its initial complaint, in December 2008, or its first amended complaint, in November 2009. In particular, the district court found that TBW had fired HDR in October 2008 and retained a different firm to complete the investigation, and, prior to that point, TBW had repeatedly expressed frustration with HDR’s inadequate work and lack of results.

The head of TBW’s investigation team, Kennedy, was so frustrated with HDR’s head engineer, Meyer, that Kennedy didn’t even want to speak with Meyer on the phone. In addition, the two new claims alleged that HDR wasted time performing an ineffectual trench drain test, but a January 2010 report in TBW’s possession had criticized the trench drain test as poorly designed, meaning that TBW has known the basis for that allegation for almost a year prior to requesting leave to amend.

Finally, the district court found that amendment would prejudice HDR, since it would introduce an entirely new theory of recovery into the case after the close of discovery. The inevitable result of amendment would be that “HDR would need to investigate the facts and prepare to defend against the new claim[s],” which would require additional discovery and further delay the trial. TBW resists these findings in three ways, but we remain unpersuaded.

TBW asserts that, although it was dissatisfied with HDR’s conduct during the investigation, it only learned other facts pertinent to these claims toward the end of discovery. Yet it is apparent that TBW had had enough discovery to state both claims by January 2010, well before the end of the discovery period and TBW’s December 2010 request to amend for a second time. In addition to the facts that the district court specifically identified in its order, TBW was even aware of factual allegations that HDR had hired its own expert, Dr. Link, to assess HDR’s liability.

Indeed, TBW spends several pages of its brief discussing Dr. Link’s involvement as crucial evidence of HDR’s duplicity, yet admits that it had learned of Link’s “true role” as early as July 2009. As is by now well-established, a complaint need only “contain a short and plain statement of the claim” accompanied by “enough facts to make a claim for relief plausible on its face.” Resnick v. AvMed, Inc., 693 F.3d 1317, 1324-25 (11th Cir.2012) (internal quotation marks omitted). TBW knew enough to make a short, plain, and plausible statement of its claims concerning HDR’s investigation.

Next, TBW claims that there would have been little to no prejudice to HDR because TBW asked for leave to amend only shortly after the close of discovery. Thus, any additional discovery could have occurred within the seven months prior to trial. Again we are unpersuaded. After discovery closed in December 2010, HDR was entitled to, and did file a motion for

summary judgment in February 2011. The district court would necessarily have had to allow HDR to complete discovery on all of the new claims prior to the deadlines for dispositive motions, which would necessitate appending discovery to the front-end of the

schedule and pushing all of the other deadlines back. Indeed, since TBW’s new claims related to a distinct set of facts, there also could have been and likely would have been motions in limine contesting each party’s experts or other evidence. Thus, TBW’s bare claim that there would be no delay is unconvincing. At all events, we can discern no abuse of discretion in the district court’s determination. Lastly, TBW faults the district court for failing to consider the prejudice to TBW.

While it is true that the district court’s order does not explicitly discuss prejudice to TBW, the district court’s finding that TBW had the ability to plead this claim earlier contains the implicit finding that TBW was the cause of its own prejudice. Moreover, as courts of appeals have long recognized, a district court is “not obligated to list, and reject, factors that might have supported a contrary decision” in the exercise of its discretion. J.P. Stevens Co. v. Lex Tex Ltd., 822 F.2d 1047, 1053 (Fed.Cir.1987); cf. United States v.

Kuhlman, 711 F.3d 1321, 1326 (11th Cir.2013) (in sentencing, a “district court should set forth enough information to satisfy the reviewing court of the fact that it has considered the parties’ arguments and has a reasoned basis for making its decision, but nothing requires the district court to state on the record that it has explicitly considered each of the § 3553(

a) factors” (internal quotation marks and citations omitted)). The district court’s thorough, seventeen-page order gave ample and detailed consideration to both parties’ arguments and counterarguments. AFFIRMED; MOTION TO CERTIFY DENIED. . Barnard could not move for

summary judgment itself under the terms of the settlement, which barred it from filing any dispositive motion. . Fabre v. Marin, 623 So.2d 1182 (Fla.1993). . The propriety of denying HDR the opportunity to include Barnard as a Fabre defendant is not before us. Therefore, we need not and do not express any opinion regarding whether the denial of HDR's motion to add Barnard as a Fabre defendant was proper under the rather unusual procedural circumstances presented here. .

In the event we cannot decide this state-law claim since there are no Florida Supreme Court or appellate court precedents on the issue, TBW also moves us to certify the question of whether Florida's comparative negligence law would bar the presentation of evidence that a former co-defendant caused the damages. For the reasons we explain below, certification is unnecessary. .

The district court stated that the “motion [to exclude] is not going to be granted”; that “to the extent TBW thinks that by virtue of the ruling on the Fabre issue [HDR is] not going to be able to put on [its] collapse upon wetting [theory], that’s not going to happen”; and that “that is my ruling.” . HDR argues that collateral estoppel is not the proper doctrine to apply to this question because collateral estoppel typically addresses the effect of a final judgment in one proceeding on a separate proceeding. At most, this point has some semantic relevance.

While it is true that "the term ‘direct estoppel’ more appropriately characterizes the application of estoppel principles” within the context of a single case or cause of action, the "analysis ... remains the same, whether we refer to the application of estoppel principles as ‘direct’ or ‘collateral.’ ” United States v. Shenberg, 89 F.3d 1461, 1478 (11th Cir.1996); see also DuChateau v.

Camp, Dresser & McKee, Inc., 713 F.3d 1298, 1301 (11th Cir.2013) (‘"Direct estoppel' is a form of issue preclusion that arises ‘within the confines of a single claim or cause of action.' ”) (quoting 18 Charles Alan Wright, Arthur R. Miller & Edward H. Cooper, Federal Practice and Procedure § 4418 (2d ed.2002)). Thus, panels of this Court in the past have used the two terms interchangeably. See Shenberg, 89 F.3d at 1478. . In Bonner v.

City of Prichard, 661 F.2d 1206, 1209 (11th Cir.1981) (en banc), this Court adopted as binding precedent all decisions of the former Fifth Circuit handed down prior to October 1, 1981. . As we see it, TBW’s arguments to the contrary rely on two flawed assumptions. First, TBW reiterates several times that the

summary judgment proceeding definitively resolved the issue of Barnard’s fault. Again, the legal issue of Barnard's fault or liability is not equivalent to the factual issue of whether Barnard caused the damage under HDR’s collapse upon wetting theory. TBW also argues that the collapse upon wetting theory "actually was litigated at

summary judgment” because the stipulated facts “included references to HDR’s theory,” but HDR did not contradict those facts or argue that those facts precluded

summary judgment. TBW insists, in other words, that the stipulations between TBW and Barnard are binding on HDR because it did not oppose the grant of

summary judgment to Barnard. But an issue is not "actually litigated if it is the subject of a stipulation between the parties,” and at most “may ... be binding in a subsequent action between the parties if the parties have manifested. an intention to that effect.” Restatement (Second) of Judgments, § 27 cmt. e (1982). Since these stipulations were not actually litigated, and the parties to the stipulations were TBW and Barnard, they cannot fulfill the requirements for issue preclusion or be applied against another party (e.g., HDR). . In so holding, we do not mean to imply that a

summary judgment absolving one defendant of liability can never have issue-preclusive effect as to a subsidiary factual question. For example, in a case where the only disputed issue was causation, and there were no factual stipulations between the plaintiff and the defendant, then a district court's grant of

summary judgment to that defendant may well have issue-preclusive effect on co-defendants' attempts to later introduce evidence of the exonerated defendant’s causation. . Bromwell apparently used the preexisting ASTM D 4767 data because those tests were taken during the development of the embankment, and thus were the best measurement of the soil's collapse potential when the reservoir was being built. .

Because undue delay was an adequate ground for the district court to deny TBW leave to add both the express contractual claim and the claim for breach of an implied covenant of good faith and fair dealing, we need not and do not address the district court’s alternative holding that the implied-covenant claim was futile.