Apache Corporation v. Castex Offshore, Inc. Castex Energy, Inc. Castex Energy Partners, LLC, F/K/A Castex Energy Partners, L.P. Castex Energy 2008, LP And Castex Energy Developement Fund, LP

Affirmed as Modified and Opinion filed May 11, 2021.

In The

Fourteenth Court of Appeals

NO. 14-19-00605-CV

APACHE CORPORATION, Appellant

V. CASTEX OFFSHORE, INC.; CASTEX ENERGY, INC.; CASTEX ENERGY PARTNERS, LLC, F/K/A CASTEX ENERGY PARTNERS, L.P.; CASTEX ENERGY 2008, LP; AND CASTEX ENERGY DEVELOPMENT FUND, LP, Appellees

On Appeal from the 133rd District Court Harris County, Texas Trial Court Cause No. 2015-48580

OPINION

In this breach of contract case between oil and gas companies, the main question presented is whether the non-operator produced sufficient evidence that its damages resulted from the operator’s willful misconduct. To answer that question, we must first determine the meaning of “willful misconduct.” We hold that willful misconduct means deliberate mismanagement committed without regard for the

consequences. We further conclude that there is sufficient evidence that the operator engaged in such willful misconduct with respect to one of the non-operator’s counterclaims, but not both of them.

As for the remaining questions presented, we conclude that the trial court did not abuse its discretion by awarding prejudgment interest on a recovery of attorney’s fees, and that the operator did not establish that it is entitled to a new trial on its separate claims for relief against the non-operator.

I. BACKGROUND

The disputes in this case arise out of several joint operating agreements between Apache Corporation and Castex Energy Partners, LLC, f/k/a Castex Energy Partners, L.P. One of the agreements involved the expansion of a natural gas processing facility known as Belle Isle. Another agreement involved the drilling of a well under a lease known as Potomac. And the remaining agreements, which the parties collectively dubbed the “Apache Contracts,” involved unrelated projects from the same region in southern Louisiana where Belle Isle and Potomac are both located. In all of these various agreements, Apache was the operator and Castex was the non-operator.

Apache sued Castex, alleging breaches of the Potomac agreement and the Apache Contracts. These claims were based on Castex’s alleged failure to pay its proportionate share under the agreements.

Castex countersued, alleging breaches of the Belle Isle and Potomac agreements. These claims were based on Apache’s alleged mismanagement of the two projects. More specifically, Castex alleged that Apache’s mismanagement led to gross overspending at Belle Isle, and to irreversible damage to the natural gas reservoir at Potomac.

All of these issues were submitted to a jury, which returned findings in favor of Castex. The trial court rendered a judgment based on those findings, and this appeal followed.

II. BELLE ISLE

A. Overview of Castex’s Claim

When the parties acquired their interests in Belle Isle, it was already a fully functional facility, capable of processing 18 million cubic feet of natural gas per day. But with the ongoing development in the region, Apache projected that more than ten times that amount would be produced each day. To capitalize on this greater potential, Apache entered into an agreement with Castex to upgrade the facility and increase its capacity. Under the terms of their agreement, Apache was responsible for 75% of the costs and Castex was responsible for the remaining 25%. That allocation corresponded with the parties’ respective interests in the facility.

As the operator, Apache had sole control over the project, and it consulted with an outside engineering firm, which estimated that the upgrades would cost approximately $16.9 million. That estimate formed the basis of an Authorization for Expenditure (“AFE”), which both parties executed. Well before the project was finished, Apache overspent the amount in the AFE, which prompted the parties to execute a supplemental AFE for a total completion cost of $37.7 million. But Apache overspent that amount too, leading to a second supplemental AFE for $78.5 million. Once again, Apache overspent that amount. By the time the project was completed, the spending had totaled $102 million.

Castex sought damages for the cost overruns, claiming that Apache had mismanaged the project from the very beginning. This claim implicated an exculpatory clause in the parties’ model form joint operating agreement, which

provided that Apache, as the operator, could not be held responsible for any losses sustained or liabilities incurred, except those resulting from gross negligence or willful misconduct.

Castex argued both theories of liability to the jury. In response, Apache countered that it would not knowingly engage in conduct that risked such an extreme escalation in costs, especially when it bore a disproportionate share of those costs. According to Apache, the cost overruns were the result of ordinary negligence at most.

In furtherance of these theories, the parties presented the jury with a substantial amount of evidence, and that evidence showed that many things went wrong with the project.

B. The Evidence at Trial

Planning began in late 2011. Apache appointed Bill Billman to lead the expansion, and he engaged with Nova Consulting, an outside engineering firm, to design an upgraded facility capable of processing 220 million cubic feet of natural gas per day.

Billman did not remain on the project for very long. He was terminated in February 2012 for reasons that are not clear from the record. Nova was released from the project around that same time as well, supposedly because of its affiliation with Billman.

Following Billman’s termination, Doug Orr was elevated to project manager, and there is a general consensus among the parties that this elevation proved to be costly. Orr’s background was in the completion of wells. He had no experience in engineering facilities or managing their construction. For those skills, Orr relied on two other outside firms.

For support in engineering, Orr turned to M&H Energy Services, which replaced Nova in February 2012. M&H estimated that the upgrades would cost $16.9 million. This estimate was not very detailed, but it provided the basis for the parties’ original AFE, which was executed in September 2012.

For support in project management, Orr turned to Foster Wheeler Upstream, which joined the project in September 2012. Later, in April 2013, Foster Wheeler was tapped to also take over the engineering of the facility because M&H was not moving quickly enough.

When he was in charge, Orr did not develop a detailed scope for the project, and he allowed the scope to change over time. For example, in September 2012, engineers were instructed to increase the capacity of the facility so that it would process 350 million cubic feet of natural gas per day, instead of the 220 million cubic feet that had been planned earlier. And in December 2012, engineers were instructed to design the facility so that it was equipped with one less amine unit, which is a device that treats natural gas by removing its concentration of carbon dioxide. These changes necessitated reengineering of the facility, which led to delays and increased costs.

Orr also failed to ensure that there was a detailed project schedule. He allowed for the procurement of equipment and materials in March 2012, and he allowed for construction to begin in February 2013. These actions were premature, as they both occurred before the engineering had even been completed. These actions also necessitated reengineering in the field, which led to more delays and increased costs.

By the end of February 2013, Orr was already overbudget. He had spent $9.1 million, and he had committed to spend another $8 million for labor and equipment that had not yet been invoiced. And at this point, the project was nowhere near completion.

By the end of July 2013, Orr was asked to provide an update to Jon Jeppesen, an executive vice president at Apache. Orr advised Jeppesen that the project was overbudget, and that the expected cost to complete was $26 million. This estimate was far from accurate. In a separate cost-tracking report that only he had received, Orr knew that, as of the beginning of July 2013, the total spending on the facility had already approached $25.9 million. And even then, the project was still nowhere near completion.

David Carmony, a vice president at Castex, learned of the overspending no later than August 2013, after he received a joint interest billing statement for the previous month. The billing statement reflected more than $19.6 million in spending, but the true total was much higher. In the cost-tracking report that had been sent to Orr in early August 2013, the total spending approximated $28.6 million.

Carmony mentioned the overspending during an August 2013 lunch with Jeppesen. After the lunch, Jeppesen scheduled a meeting to discuss the project with David Tirey, who had a supervisory role over Orr, and Paul McKinney, who had a supervisory role over Tirey. During the meeting, Jeppesen went “ballistic,” to borrow his own language. Jeppesen got the impression that the project was being run by the engineering companies and by major vendors, who had little incentive to control costs. He ordered an internal audit review of all invoices of the contractors involved. He also appears to have called for the preparation of a supplemental AFE, which Orr assisted in drafting in early September 2013.

Orr explained in a memorandum that the project required additional expenditures primarily because of changes in scope. He wrote that some of the changes were needed so that the existing facility could operate safely during the course of its own expansion. Other changes were needed because of findings after a government inspection. Orr also advised that previous cost estimates had been too

low, that Foster Wheeler was modifying the engineering plans from M&H, and that there had been expensive weather delays. For all of these reasons and more, Orr determined that the AFE should be supplemented for a total completion cost of $37.7 million.

After this supplemental AFE was prepared, Jeppesen terminated McKinney, and additional personnel changes soon followed. Tirey was given the option of taking a demotion or retiring early; he chose the latter. Orr was largely reassigned to another project, which made his continued involvement with Belle Isle significantly reduced. He was replaced with Brandon LePretre, a production foreman who took over as interim project manager in October 2013. LePretre had only a high school education and lacked the necessary qualifications to manage a project of this scale. He stepped aside as project manager in November 2013, when Kenneth Moore took the lead.

Moore was an engineer from another division at Apache, and he had over twenty years of experience in project management. Unlike with Orr and LePretre, there is a general consensus among the parties that Moore was competent to serve as project manager, a role that he kept until the upgrades were completed in May 2014. But despite his competence, costs doubled under Moore’s watch.

In March 2014, Moore prepared a supplemental AFE for $78.5 million. He separately explained in a memorandum that more funds were needed for reasons that predated his appointment as project manager. He also explained that he had brought in an experienced team of engineers from another division at Apache, and that changes were planned to make the facility more efficient.

By the middle of April 2014, and after the parties had executed the second supplemental AFE, Moore sent an email forecasting that the spending would exceed $95 million. He explained, “At a high level the increased cost comes from the

schedule moving out an extra month and invoices on work completed but not known about.” Mark Bauer, who replaced McKinney, responded to Moore’s email with these words: “I’m speechless. I don’t have any concept how these costs are so high.”

The project was completed roughly two weeks later at a final cost of $102 million.

C. The Verdict

After hearing this evidence and more that we discuss below, the jury was asked a series of questions.

One of the first questions was, “Did Apache fail to comply with the Belle Isle Facility Agreement?” This question included instructions that Apache had duties under the facility agreement (also known as the joint operating agreement) to conduct all operations in a good and workmanlike manner; to keep accurate books, accounts, and records of operations; and to the extent reasonably possible, to furnish pertinent information to Castex regarding the facility. The jury answered this question “Yes.” Though the question did not specify whose acts and omissions could be attributed to Apache for liability purposes, the parties are in agreement that the corporation could be bound by any single employee, not just a vice principal.

The next question asked, “Did Apache’s failure to comply result from gross negligence?” There was an instruction defining gross negligence, but the jury’s answer was “No.”

The jury was then asked, “Did Apache’s failure to comply result from willful misconduct?” The jury’s answer was “Yes.”

The next question was, “What sum of money, if any, if paid now in cash, would fairly and reasonably compensate Castex for its damages, if any, that resulted from such failure to comply?” There was a further instruction that the measure of

damages was the difference, if any, in the amount Castex agreed to pay for the facility, and the amount that Castex actually paid. The jury found that this amount was $5,566,577.

D. Sufficiency Challenge

Apache does not challenge the finding that it failed to comply with the joint operating agreement. However, Apache does challenge whether there is legally and factually sufficient evidence to support the finding of willful misconduct.

In such a challenge, we measure the sufficiency of the evidence under the charge that was given, unless there was a timely and valid charge objection. See St. Joseph Hosp. v. Wolff, 94 S.W.3d 513, 530 (Tex. 2003). Because Apache timely objected to the charge, we begin by determining whether its charge objection was valid.

1. Meaning of Willful Misconduct

Though Belle Isle is located in Louisiana, the parties’ joint operating agreement contained a Texas choice of law provision, and in keeping with the Texas Pattern Jury Charges, the trial court did not provide a definition for willful misconduct. See Texas Pattern Jury Charges (Oil & Gas) § 305.28 (2018).

Apache objected to the omission of a definition and requested that the jury be instructed that “willful misconduct requires a subjective, intentional intent to cause harm.” The trial court overruled that objection.

At least one intermediate court of appeals has recognized that willful misconduct has a meaning consistent with Apache’s proposed definition. See IP Petrol. Co. v. Wevanco Energy, L.L.C., 116 S.W.3d 888, 898 (Tex. App.—Houston [1st Dist.] 2003, pet. denied) (“A finding of willful misconduct required evidence of a specific intent by [the Defendant] to cause substantial injury to Plaintiffs.”).

However, the Texas Supreme Court has more recently opined that willful misconduct is “short of genuine intentional injury.” See Mo-Vac Serv. Co. v. Escobedo, 603 S.W.3d 119, 125–26 (Tex. 2020). Because we are bound to follow that higher court, we conclude that willful misconduct does not require a subjective intent to cause harm, and that the trial court correctly overruled Apache’s objection.

In the absence of a technical legal definition, we apply the ordinary meaning of willful misconduct. See Green Tree Acceptance, Inc. v. Combs, 745 S.W.2d 87, 90 (Tex. App.—San Antonio 1988, writ denied) (concluding that the words “willful misconduct . . . are words of ordinary meaning and are readily understood by the average person”). According to one leading dictionary, the word “willful” means “intentional; deliberate” or “having or showing a stubborn and determined intention to do as one wants, regardless of the consequences or effects.” See New Oxford American Dictionary 1978 (Angus Stevenson & Christine Lindberg eds., 3d ed. 2010). That same dictionary defines “misconduct” as “unacceptable or improper behavior, esp. by an employee or professional person” or “mismanagement, esp. culpable neglect of duties.” Id. at 1117.

Based on these definitions, a plaintiff can show that a defendant is liable for willful misconduct if the evidence establishes that the defendant intentionally or deliberately engaged in improper behavior or mismanagement, without regard for the consequences of his acts or omissions.1 We will measure the sufficiency of the evidence against this standard.

1 This standard differs from the standard for an intentional tort, which requires proof of a more culpable mental state. To recover damages on an intentional tort, the plaintiff must show that the defendant desired the consequences of his acts or omissions, or that the defendant knew that those consequences were substantially certain to occur. See Mo-Vac Serv. Co., 603 S.W.3d at 125.

2. Legal Sufficiency

When analyzing the legal sufficiency of the evidence, we review the record in the light most favorable to the challenged finding, crediting favorable evidence if a reasonable factfinder could and disregarding contrary evidence unless a reasonable factfinder could not. See City of Keller v. Wilson, 168 S.W.3d 802, 827 (Tex. 2005). The evidence is sufficient to support a finding if the evidence rises to a level that would enable reasonable and fair-minded people to differ in their conclusions. Id. at 822. The evidence is insufficient to support a finding only if (a) there is a complete absence of evidence of a vital fact, (b) the court is barred by rules of law or evidence from giving weight to the only evidence offered to prove a vital fact, (c) the evidence offered to prove a vital fact is no more than a mere scintilla, or (d) the evidence conclusively establishes the opposite of the vital fact. Id. at 810.

Castex contends that there is legally sufficient evidence of willful misconduct based on the acts and omissions of multiple individuals within the Apache organization. We need only focus on the acts and omissions of Orr because there is objective proof that Orr was aware of the cost overruns for months and did nothing about them.

The evidence showed that Orr had hired Courtney Bonura at M&H to provide cost-tracking services. As part of these services, Bonura would compile all of Apache’s expenditures on a monthly basis. Then he would tabulate those expenditures, put them in a report, and forward the report to Orr, who was his only contact at Apache.

Bonura began making these reports in June 2012, and he delivered his final report in October 2013, after which he was apparently let go. We reproduce a portion of the reports in the following table:

Date Spent Committed Total

Feb. 1, 2013 $8,422,052.27 $6,731,358.02 $15,153,410.29

Feb. 28, 2013 $9,109,270.86 $8,022,868.52 $17,132,139.38

Mar. 31, 2013 $9,386,063.34 $9,424,347.23 $18,810,410.57

Apr. 30, 2013 $13,401,039.94 $9,689,622.00 $23,090,661.94

May 31, 2013 $17,041,191.46 $6,139,992.00 $23,181,186.46

June 30, 2013 $22,191,170.98 $3,668,858.00 $25,860,028.98

July 31, 2013 $25,312,276.94 $3,315,025.00 $28,627,301.94

Aug. 31, 2013 $31,134,475.42 $2,698,025.00 $33,832,500.42

Sept. 30, 2013 $36,582,958.06 $5,957,193.00 $42,540,151.06

The first column of this table does not represent the date of Bonura’s report, but rather the accrual date for the expenditures in the corresponding row. The second column represents actual expenditures for invoices already received. The third column represents committed expenditures for invoices that were anticipated but not yet received. And the fourth column represents the total of the second and third columns.

Bonura testified that his standard practice was to transmit his reports to Orr shortly after the date indicated in the first column, which means that Orr would have known in early February 2013—when construction was just beginning—that he had already exhausted more than $15 million, or 89% of the approved budget for the entire project. By the time Bonura transmitted his next report in early March 2013, Orr would have known that the spending had topped $17 million, which exceeded the AFE.

Jeppesen wrote in an email that Orr and his superiors “had absolutely NO authority” to spend an amount in excess of the AFE. Orr likewise testified that Apache had a policy about requiring supplemental AFEs, which he admitted to having violated. That violation is plain evidence of misconduct, and it continued until September 2013, when Orr prepared the first supplemental AFE.

Orr did not admit during the trial that his misconduct had been willful. Or stated another way, he did not testify that when the cost overruns first occurred, he knew that he needed to pursue a supplemental AFE, but he deliberately did not do so. Nevertheless, a jury could have reasonably inferred that ultimate fact from the circumstantial evidence.

The circumstantial evidence included the subsequent reports from Bonura, which alerted Orr month after month for more than five months that he was overbudget. Yet Orr did nothing in response to those reports, other than allow the overspending to grow.

In addition to Bonura’s reports, the circumstantial evidence included an email from LePretre to Jeppesen, in which LePretre described how he had asked Orr on several occasions about the money being spent, and Orr responded that “money is not a problem.” The jury could have reasonably determined from this evidence that Orr was consciously indifferent to the cost overruns.

There is also ample evidence of causation. Chris Sullivan, who testified as an expert on behalf of Castex, said that Orr’s failure to timely supplement the AFE impacted the effectiveness of any mitigation efforts. Sullivan continued, “Had they realized that the costs were increasing pretty dramatically, nearly doubling or more than doubling, they could have taken a hard look at what these changes were and tried to eliminate the changes before they incurred additional cost.”

That sentiment was echoed by Tom Thweatt, whom Apache retained as an expert. He said that “the earlier you identify a problem . . . the more likely it is to be mitigateable, if that’s a word.” Thweatt’s testimony is also consistent with Moore’s memorandum for the second supplemental AFE, in which Moore indicated that additional funding was required because of the acts and omissions of his predecessors.

By the time that Orr had prepared the first supplemental AFE in early September 2013, the spending had already surpassed $33.8 million, which was nearly double the amount of the original AFE. And by the end of September 2013— before Apache had ever forwarded the supplemental AFE to Castex—the spending had surpassed $42.5 million, which was even higher than the supplemental AFE of $37.7 million.

From this evidence alone, the jury could have reasonably determined that Castex’s share of the cost overruns amounted to more than $6 million.2 The jury’s actual damages finding of $5,566,577 is well within that range of evidence. We therefore conclude that the evidence is legally sufficient to support the jury’s findings.

3. Factual Sufficiency

When a party challenges the factual sufficiency of a finding for which the party did not bear the burden of proof at trial, we review all of the evidence in a neutral light and will reverse only if the evidence supporting the finding is so contrary to the overwhelming weight of the evidence as to make the judgment clearly wrong and manifestly unjust. See Maritime Overseas Corp. v. Ellis, 971 S.W.2d 402, 406–07 (Tex. 1998). Under this standard, we may not pass upon the credibility of

2 ($42,540,151.06 – $16,950,863) × 0.25 = $6,397,322.02

witnesses or substitute our judgment for that of the factfinder, even if the evidence would clearly support a different result. Id. at 407.

Apache does not articulate a clear argument for why the evidence is factually insufficient. Instead, Apache proposes that the judgment must be reversed because “no sane company would purposefully increase its own costs.” This proposition does not correctly apply the standard for willful misconduct. The standard is whether the defendant deliberately engaged in misconduct without regard for the consequences—not, as Apache has suggested, whether the defendant sought the consequences of his own misconduct, which describes an intentional tort.

Apache also seems to believe that the evidence is insufficient based on the brief statements of three separate witnesses. Apache refers first to Orr, who testified that he did not intentionally fail to do things that he knew he ought to do. Apache refers next to the testimony of Mark Cook, a project manager at M&H, who said that he never saw anyone at Apache do anything that he would consider to be willful misconduct. And finally, Apache refers to the testimony of Jeppesen, who opined that the personnel in charge of Belle Isle, which would include Orr, did not know enough about what they were doing to appreciate the risks of their own decisions.

Weighing against this testimony are the statements from LePretre, who explained that Orr was not concerned about the spending because he believed that Apache’s developments in the region would be extremely profitable. According to LePretre, Orr said “in a nice way it’s not [LePretre’s] problem to worry about what’s being spent, that it’s going to all be worth it at the tail end.” Orr rationalized that there were “hellacious reserves” of natural gas, and that the spending on the facility would be a drop in the bucket once that natural gas was produced.

Having considered all of his evidence in a neutral light, we conclude that the jury had a substantial basis for making an implied finding that Orr knew, but did not

care, about his own overspending. Therefore, we cannot say that the jury’s express finding of willful misconduct was clearly wrong and manifestly unjust.

4. Prejudgment Interest

For its claim that Apache had breached the joint operating agreement relating to Belle Isle, Castex sought reasonable and necessary attorney’s fees, which the jury awarded in the amount of $3,152,301. The trial court rendered judgment based on that award, and it further granted Castex $238,862.67 in prejudgment interest on the attorney’s fees.

Apache objected to the prejudgment interest on two separate grounds. First, Apache argued that prejudgment interest is not available under Texas law, at least according to some courts of appeals, including the First Court of Appeals, with which we share jurisdiction. See Power Reps, Inc. v. Cates, No. 01-13-00856-CV, 2015 WL 4747215, at *22 (Tex. App.—Houston [1st Dist.] Aug. 11, 2015, no pet.) (mem. op.). Our court, on the other hand, has previously determined that an award of prejudgment interest is within the trial court’s discretion, provided that the client has actually paid the attorney’s fees before the entry of judgment. See Nova Cas. Co. v. Turner Constr. Co., 335 S.W.3d 698, 706 (Tex. App.—Houston [14th Dist.] 2011, no pet.). Apache recognized this split in authority at the time of its objection, but Apache did not know to which court its appeal would ultimately be assigned. Now that the appeal has been assigned to our court, Apache urges us to follow the First Court, but we decline that invitation because we are bound to follow our own precedent.

Apache also objected that even if prejudgment interest were allowable, it should not be awarded in this case because Castex waited too long to prove its entitlement to prejudgment interest. Apache emphasized that Castex did not produce any evidence to the jury that it had already paid its attorney’s fees. Instead, Castex

produced its evidence of payment to the trial court after the jury’s verdict had been received and after the jury had been discharged. This evidence came in the form of an attorney’s declaration, which Castex submitted in support of a motion for entry of final judgment. During post-verdict proceedings, Apache objected that the declaration was “too late,” but the trial court overruled that objection.

Apache now contends that the trial court’s ruling violated Rule 270 of the Texas Rules of Civil Procedure, which provides as follows: “When it clearly appears to be necessary to the due administration of justice, the court may permit additional evidence to be offered at any time; provided that in a jury case no evidence on a controversial matter shall be received after the verdict of the jury.”

Apache suggests that the payment of attorney’s fees was a “controversial matter” which required a factual determination by the jury, but Apache cites to no authority applying Rule 270 in that manner. We are not aware of any such authority either.

The Texas Supreme Court has stated that “the recovery of prejudgment interest does not require any evidentiary proof at trial.” See Benavides v. Isles Constr. Co., 726 S.W.2d 23, 26 (Tex. 1987). That statement suggests that Rule 270 would not preclude the trial court from considering evidence of payment that was produced for the first time during post-verdict proceedings. Indeed, trial courts routinely decide matters relating to prejudgment interest in post-verdict proceedings. See State Farm Mut. Auto. Ins. Co. v. Norris, 216 S.W.3d 819, 821 (Tex. 2006) (remanding for the trial court to determine when prejudgment interest began to accrue); Figueroa v. Davis, 318 S.W.3d 53, 66 (Tex. App.—Houston [1st Dist.] 2010, no pet.) (stating that evidence relating to prejudgment interest and tolling “may be presented at any time during the court’s plenary power”). We therefore conclude

that the trial court did not abuse its discretion by awarding Castex prejudgment interest.

III. POTOMAC

A. Overview of Castex’s Claim

While construction at Belle Isle was underway, a separate team at Apache was drilling a natural gas well under the lease known as Potomac. Apache and Castex each owned an equal 50% interest in that well, but under the terms of the parties’ joint operating agreement, Apache was the operator and retained sole control over the drilling operation.

Apache completed the well, but the drilling operation did not proceed according to plan. First there were technical difficulties in the original wellbore, which required Apache to pursue a sidetrack, or secondary wellbore. Then there were technical difficulties with the sidetrack, which required Apache to pursue additional sidetracks. Only after the drilling of a fifth sidetrack was the well able to produce any gas, and that production did not last for very long. After a certain point, the well yielded so much water as compared to gas that continued production became uneconomical.

Castex sued Apache, claiming that Apache had mismanaged the drilling of the well by failing to set certain liners, which would have protected the gas reserves from the influx of water. Castex sought damages for its share of the cost overruns in drilling the well, and for its share of the gas reserves that were lost because of Apache’s mismanagement.

As with its other claim concerning Belle Isle, Castex’s claim concerning Potomac implicated an exculpatory clause under the parties’ model form joint operating agreement. That exculpatory clause was similarly worded, and it provided

that Apache, as the operator, could not be held responsible for any losses sustained or liabilities incurred, except those resulting from gross negligence or willful misconduct.

Castex argued both theories of liability to the jury, whereas Apache largely countered that it had done nothing wrong.

B. The Evidence at Trial

To help frame the ensuing discussion, we begin with an illustration of the well, its sidetracks, and the gas formations that the parties tried to develop:

This well was situated in the shallow waters of a bay. Deep beneath that bay at more than three miles below sea level were several natural gas reservoirs, which were known in the region as Cib Op sands.3 These sands were all separated by layers

3 Cib Op is short for Cibicides carstensi var. opima, the taxonomic citation for a prehistoric marine microorganism.

of shale, a mostly impermeable type of rock. Within each sand were variable concentrations of water and natural gas.

For purposes of identification, the sands were assigned different numbers, only five of which are relevant to this case. At more than sixteen thousand feet below sea level, the shallowest of the five sands is Cib Op 1. Then at roughly eighteen thousand feet is Cib Op 4. These two sands exist under high pressure, and in this particular area of the bay, they are both heavily saturated with water. Beneath Cib Op 4 are the three remaining sands: Cib Op 5, Cib Op 6, and Cib Op 7. These are all gas-producing sands, none of which is deeper than nineteen thousand feet.

Apache’s plan was to drill down to Cib Op 7, produce the gas that was there, and then work its way back up by producing gas from Cib Op 6 and Cib Op 5. Certain precautions had to be taken with this approach because of Cib Op 5. That sand had already been drilled by other producers in the area, which meant that the pressure within Cib Op 5 was lower than the pressures within Cib Op 1 and Cib Op 4. This pressure differential likewise meant that if a pathway ever opened between Cib Op 5 and those higher sands, then water from Cib Op 1 or Cib Op 4 could travel down to Cib Op 5, in an event known as “cross-flow.” And if the cross-flow were not corrected, then the well could “water out,” an event that essentially forces the end of production.

The risks of cross-flow were known to Apache, and its plan to prevent that occurrence was to isolate the sands by using a liner set in cement. But as indicated earlier, Apache encountered many difficulties when trying to execute this plan.

Apache drilled a hole all the way down to its target depth in Cib Op 7. The diameter of this hole varied according to its depth. At the surface, the hole was at its widest point, and Apache installed casing that was correspondingly wide. The hole

narrowed as the depth increased, which meant that the casing also narrowed. When imagined in profile, the casing from top to bottom resembled an inverted spyglass.

Apache set liners of varying diameters along the way, depending on the size of the casing. When Apache tried to hang one of the final liners, a mechanical issue occurred. Apache attempted to pull the liner back out of the hole, but the liner got stuck, so Apache made the decision to cement the liner in place. Apache then tried to drill back through the cement to complete the well, but Apache encountered an obstruction. A mill was brought in to grind through the obstruction, but after several days of trying, the effort proved to be unsuccessful.

Apache determined that it would not be able to produce gas from the original wellbore. As an alternative, Apache decided to attempt a sidetrack. Even though the target was still Cib Op 7, Apache started the sidetrack above Cib Op 4.

During the drilling of the sidetrack, Apache experienced a “kick,” which is an unexpected flow of formation fluid into the wellbore. Apache tried to run a liner, but the liner got stuck. Apache abandoned the sidetrack while the drill pipe was still within the limits of Cib Op 4.

Apache attempted a second sidetrack, but a similar problem arose. A liner got stuck, and Apache abandoned the sidetrack when the drill pipe was still within the limits of Cib Op 4.

Apache then drilled a third sidetrack, which managed to pass through Cib Op 4 and penetrate the very top of Cib Op 5. But once again, Apache experienced a kick, and when Apache tried to set a liner, the liner got stuck. Apache plugged the hole with cement, and then it tried a fourth sidetrack.

Apache approached the fourth sidetrack differently. Beginning at a much shallower depth, Apache drilled through Cib Op 1 and then around Cib Op 4, which

was confined by a natural fault line. Apache was then able to drill down to Cib Op 5, but its pipe still got stuck. Pressures began to climb within the well, and Apache evacuated the drill rig out of concerns for a blowout. After a well control team stopped the blowout, Apache squeezed cement into the hole and attempted a fifth sidetrack.

As with the previous sidetrack, Apache drilled through Cib Op 1, around Cib Op 4, and into Cib Op 5. Because Apache projected that Cib Op 5 had paying quantities of natural gas that would make the venture profitable, Apache determined to complete the well in Cib Op 5, rather than risk another stuck pipe in pursuit of the deeper sands. This fifth sidetrack produced gas for about two months, and then it watered out.

C. The Jury’s Verdict

The first question that was asked of the jury was, “Did Apache fail to comply with the Potomac JOA?” The question included an instruction that Apache was required by the joint operating agreement to conduct all operations in a good and workmanlike manner. The jury’s answer was “Yes.”

The next question asked, “Did Apache’s failure to comply result from gross negligence?” The jury’s answer was “No.”

The next question asked, “Did Apache’s failure to comply result from willful misconduct?” The jury’s answer was “Yes.”

The final question asked the jury to determine two measures of damages: one relating to the cost overruns in the drilling of the well, and the other relating to the loss of gas reserves rendered unrecoverable by Apache’s failure to comply. The jury determined that Castex’s damages for the cost overruns amounted to more than $8.9

million, and its damages for the unrecoverable reserves amounted to more than $44.6 million.

D. Sufficiency Challenge

Apache contends that the evidence is legally and factually insufficient to support the jury’s finding of willful misconduct. Apache also contends that there is no evidence to support the damages finding for the unrecoverable reserves. We need only address the liability finding because it is dispositive.

1. Meaning of Willful Misconduct

As before, we begin with the threshold issue of the meaning of willful misconduct. Unlike the joint operating agreement covering the facility at Belle Isle, which had a Texas choice of law provision, the agreement covering the well at Potomac had a Louisiana choice of law provision. And during the charge conference, Apache requested that the trial court submit the following definition, which was taken from the Louisiana case of Koonce v. St. Paul Fire & Marine Ins. Co., 172 So. 3d 1101, 1105 (La. Ct. App. 2015, writ denied): “To constitute willful misconduct, there must be some voluntary, intentional breach of duty, which may be unlawful, dishonest, or improper, or perhaps all three, that is committed with bad intent or, at best, with wanton disregard for the consequences.” The trial court denied Apache’s request.

To the extent that Apache argues that a finding of willful misconduct requires proof of an intent to cause harm under Louisiana law, that argument finds no support in Koonce, which specifically noted that willfulness describes “that degree of fault which lies between intent to do wrong, and the mere reasonable risk of harm involved in ordinary negligence.” Id. at 1106. This understanding comports with the

Texas view of willful misconduct, which as we mentioned above is “short of genuine intentional injury.” See Mo-Vac Serv. Co., 603 S.W.3d at 125–26.

Because Koonce recognizes that willful misconduct should be understood according to its ordinary meaning, we conclude that the trial court did not err by denying Apache’s requested definition. See Longview Energy Co. v. Huff Energy Fund LP, 533 S.W.3d 866, 872 (Tex. 2017) (holding that the law of the forum state governs matters of procedure); Vast Constr., LLC v. CTC Contractors, LLC, 526 S.W.3d 709, 722 (Tex. App.—Houston [14th Dist.] 2017, no pet.) (“The trial court does not abuse its discretion by refusing to submit unnecessary instructions, even when the instructions represent correct statements of law.”). We therefore measure the sufficiency of the evidence under the same standard that we applied previously. In other words, to support the finding of willful misconduct, there must be some evidence that Apache intentionally or deliberately engaged in improper behavior or mismanagement, without regard for the consequences of its acts or omissions.

Castex contends that it satisfied this standard because “Apache deliberately left three open pathways for cross-flow.” Castex alleges that these pathways were located in the original hole, the third sidetrack, and the fourth sidetrack. We address each of them in turn.

2. The Original Hole

As it drilled through Cib Op 4, Apache pumped mud down through the middle of the drill pipe. The mud then exited from the bottom of the pipe and filled in the voids between the exterior of the pipe and the outer limits of the hole. One purpose of this mud was to protect the wellbore from the water in the sands. The mud was capable of deterioration though, which meant that it could not serve as a permanent barrier to cross-flow. For that permanent barrier, Apache needed to set a liner with cement.

a. Liner Depth

Apache planned to cement a liner across the entirety of Cib Op 4 before the drilling ever proceeded to one of the deeper sands. The stated reason for this plan was that Apache did not want its drill pipe to become stuck, which could happen if the high-pressure Cib Op 4 ever became exposed to the low-pressure Cib Op 5.

Pinpointing the best place for the base of the liner was a collaborative effort between Apache’s drilling engineers and Allen Peterson, who was Apache’s geologist. They determined to set the base of the liner at a depth of 17,974 feet. This liner required three attempts to set, and it was ultimately cemented at a depth of 17,972 feet, which is two feet higher than what had originally been planned.

Castex contends that the setting of this liner amounted to willful misconduct because the liner did not fully cover Cib Op 4 at its cemented depth of 17,972 feet. Furthermore, Castex contends that the liner would not have covered Cib Op 4 even at its planned depth of 17,974 feet.

To support this claim, there must have been some evidence that Apache knew, but was consciously indifferent, that its liner was not deep enough.

As for whether Apache possessed the requisite knowledge, Castex does not refer to any particular evidence identifying what liner depth would have been sufficient to cover Cib Op 4 and prevent cross-flow. In fact, Castex does not even identify the bottom depth of Cib Op 4. Instead, Castex simply refers to evidence showing that Apache knew that its liner was in sand, rather than in shale. This evidence included a drilling report, which indicated that the rock composition at 17,974 feet was 60% shale and 40% sand. The evidence also included a gamma ray curve, which is a chart that reflects the radioactivity of the rock composition. This

gamma ray curve indicated that the concentration of shale had been increasing, but according to Castex, it also indicated the presence of sand.

As for whether Apache was consciously indifferent, Castex only refers to evidence that Apache ultimately decided to list all of its interests in the bay for sale. From that evidence, Castex contends the jury could have reasonably concluded that Apache did not set the liner at the appropriate depth because Apache did not care about the long-term productivity of the well.

For the following reasons, this evidence of alleged willfulness is too slight to rise above a scintilla. See City of Keller, 168 S.W.3d at 813 (“In claims or defenses supported only by meager circumstantial evidence, the evidence does not rise above a scintilla (and thus is legally insufficient) if jurors would have to guess whether a vital fact exists.”).

Beginning in reverse order with the point about conscious indifference, Castex does not cite to any evidence revealing the date when Apache decided to offer its interests in the bay for sale. Our review of one witness’s testimony indicates that this decision was made during the “summer” of 2014, which would have been after April 2014, when Apache set the liner in the original wellbore. There is no suggestion from that earlier time that Apache knew that a sale would occur, or that Apache did not care whether the well would be successful.

As for the remaining point, there was objective evidence that Apache exercised care in setting the liner.

Email correspondence from the time of the drilling showed that Peterson had to estimate the heights and depths of the various sands. When estimating the bottom depth of Cib Op 4, Peterson considered data from the “mudlogger,” who is the contractor tracking the rock’s composition as it is brought back to the surface. The

mudlogger here reported higher concentrations of shale, which was consistent with being beneath the limits of Cib Op 4.

Peterson also considered the penetration curve, which tracked the rate of drilling. This curve showed that the drilling had slowed down, which was consistent with the drill being in shale as opposed to sand.

Peterson also considered a seismic map, and in an email to colleagues, he wrote, “Your map indicates we should see the 5 at about 17980 to 18000’.” In other words, Peterson had a substantial basis for believing that the top of Cib Op 5 was as close as six feet away from the bottom of the hole. Even if Peterson’s belief turned out to be wrong, Castex’s evidence of willfulness is so slight that the jury would have to speculate that Peterson was deliberately wrong, as opposed to simply mistaken or negligent. And under the equal-inference rule, when the circumstances are equally consistent with either of two facts, neither may be inferred. Id. at 813– 14.

b. “Cross Talk” Email

After it encountered difficulties during the drilling of the second sidetrack, which was about two months after Apache had set the previously described liner, Apache became aware of cross-flow in the original wellbore. Steve Hopkins, Apache’s head drilling engineer, devised a plan to stop the cross-flow, which he discussed in the following email:

All: We have managed to circulate 18 ppg[4] mud around and will be making an attempt to POOH[5] after this morning’s tour change and safety meeting.

4 Pounds per gallon. 5 Pull out of hole.

Once we see that the hole is stable, our plan is to slug the drill pipe and continue POOH carefully monitoring fill-ups and checking for flow as necessary. Electric line will be mobilized once we start out and our forward plan is to set a 7-3/4” cement retainer above the window and pump a cement squeeze to hopefully establishing a barrier between the CO[6] 4 and the deeper CO 5, 6 and 7. This is necessary before attempting another sidetrack. I sent e-mails yesterday and this morning to Mark Bauer giving him an update. Also called Kevin Ivy with Castex and discussed our status with him. Any questions let me know. Peterson, who apparently learned of the cross-flow around this same time, responded with the following email, using the term “cross talk” as a synonym for cross-flow:

So you think we have cross talk between the deeper pay sands and the Cib op 4? If this is the case, wouldn’t the partially depleted Cib op 5 have taken the gas? I guess it doesn’t really matter. Regardless we have gas and salt water coming from somewhere and it needs to be plugged off. Castex claims that Apache did nothing to stop the cross-flow, even though Apache had clearly recognized the need to do so. In support of this claim, Castex relies entirely on the testimony of Clay Kimbrell, its retained expert.

Kimbrell was asked a simple question about whether Apache had taken any remedial action to address the cross-flow that was occurring in the original hole. Kimbrell’s response was this: “Not that I recall. I don’t remember them doing anything. Of course, the e-mails are discussing trying to squeeze it off; but I don’t think we had confirmation of that.”

6 Cib Op.

Kimbrell’s testimony was negated by a well chronology report, which tracked the daily activities on the drill rig. On the day after Hopkins sent his email, the well chronology report indicated that Apache had rigged up lines in preparation for a squeeze job. And on the day after that, the well chronology report indicated that Apache had performed the squeeze job. This evidence demonstrates that Apache was actively trying to stop the cross-flow, which is the opposite of willful misconduct.

Kimbrell did not acknowledge the evidence in the well chronology report, and we cannot disregard it. Id. at 813 (“If an expert’s opinion is based on certain assumptions about the facts, we cannot disregard evidence showing those assumptions were unfounded.”).

3. The Third Sidetrack

Apache successfully drilled through Cib Op 4 with its third sidetrack, but Apache encountered problems soon after it penetrated Cib Op 5, and those problems eventually required Apache to abandon the sidetrack altogether. Castex contends that Apache engaged in willful misconduct during this sidetrack operation because Apache never set a liner across Cib Op 4, even though Apache had previously acknowledged the need for a liner in the original wellbore.

There is no dispute that Apache chose to not set a liner across Cib Op 4, but Apache explained that choice during the trial. Calvin Barnhill, who was retained by Apache as an expert, testified that the diameter of the third sidetrack was only six and a half inches, which is significant for at least two reasons. First, that diameter was not as wide as the section of the original wellbore that spanned Cib Op 4, which accommodated a liner that was seven and three quarters inches across. And second, the six and a half inch diameter was the size of the hole that Apache desired at the bottom of Cib Op 6 and Cib Op 7.

Barnhill explained that the sidetrack did not have a liner across Cib Op 4 because Apache wanted to preserve the size of its hole for those deeper sands, which were still being targeted at this point of the drilling operation. “We kind of jealously guard how small that hole gets,” he testified, “because the smaller the hole gets, the harder it is to work in.” He also indicated that a hole with a six and a half inch diameter is considered a “pretty good minimal size hole.”

Without a liner in the sidetrack, Apache used mud to prevent cross-flow from Cib Op 4. Apache encountered a circulation problem with its mud after drilling into the top of Cib Op 5. Apache also encountered a different problem when it took some kicks. To remedy these problems, Apache decided to place a five inch liner through Cib Op 5 so that it was protected as Apache continued down into the deeper sands. But that plan failed. The liner got stuck and a piece of pipe blocked Apache from doing other operations deeper down in the hole. Faced with that obstacle, Apache cut off the pipe and bullheaded cement into the hole.

Castex contends, with the benefit of hindsight, that Apache could have attempted a different procedure for cementing the stuck liner in place, which had been performed at another well in the area. But Castex cites to no other evidence showing that Apache’s failure to try that other procedure was deliberately improper, or that Apache was otherwise indifferent to the consequences of its own acts and omissions.

Castex seemingly claims that Apache was indifferent because, as mentioned earlier, Apache had decided to offer its assets for sale. But once again, that evidence of willfulness is no more than a scintilla. The liner in the third sidetrack got stuck in July 2014, whereas Apache sold its assets much later, in November 2014. There was no evidence that Apache had even received an offer in July 2014. In fact, Castex submitted its own offer for the assets in October 2014.

There is uncontroverted evidence that Apache tried to run a liner and to squeeze cement into the third sidetrack, both of which demonstrate that Apache was concerned about the risks of cross-flow. Even if Apache’s efforts at preventing cross-flow were ineffective, the jury could not reasonably conclude that Apache had engaged in willful misconduct by failing to set a liner across Cib Op 4 when the evidence gave rise to an equal inference that Apache was simply negligent in attempting a different approach.

4. The Fourth Sidetrack

Apache attempted to reach the deeper sands in its fourth sidetrack by starting the sidetrack at a shallower depth, then drilling through Cib Op 1, and going around Cib Op 4 altogether. This plan worked until Apache drilled through Cib Op 5, where Apache encountered numerous difficulties, including one that nearly risked a blowout. Apache then tried to set a liner, but that effort did not work, so Apache bullheaded cement twice into the hole and abandoned the sidetrack.

There is no dispute that Apache did not set a liner across Cib Op 1, much like its approach with the third sidetrack and Cib Op 4. Castex contends that this failure to set a liner amounts to willful misconduct, but the analysis is no different from the previous section involving the third sidetrack.

Barnhill explained that Apache intentionally did not set a liner across Cib Op 1 because Apache was trying to preserve the size of the hole, which was still six and a half inches in diameter. There is no evidence that Apache did not care about the long-term productivity of the well; this sidetrack operation happened in August 2014, which was still before Apache sold its assets in November 2014. Moreover, Kimbrell—Castex’s own expert—testified that by attempting to set a liner and by squeezing cement into the hole, Apache had tried to protect the well from cross- flow.

The record established that drilling operations are risky, which explains why the model form joint operating agreement shields the operator with a broad exculpatory clause. Castex could have overcome that exculpatory clause with legally sufficient evidence that Apache knew, but did not care, that it was mismanaging the drilling operation. For the reasons stated above, we conclude that Castex failed to produce such evidence of willful misconduct.

E. Apache’s Separate Claim for Relief

Apache sued Castex, claiming that Castex had failed to pay its proportionate share under Potomac’s joint operating agreement. In response, Castex conceded that it had stopped paying its bills. The parties even tendered a joint stipulation which stated that, for the costs of the well, Castex had an unpaid balance of $2,706,408. Nonetheless, Castex argued that its failure to pay was excused, which Apache disputed. To settle that dispute, several questions were submitted to the jury, which returned the following findings:

1. Castex failed to pay Apache the amounts that were due under the joint operating agreement;

2. The amount unpaid by Castex is $2,706,408—i.e., the same amount reflected in the joint stipulation;

3. Apache failed to comply with the joint operating agreement by not conducting all operations in a good and workmanlike manner;

4. Apache’s failure to comply preceded Castex’s failure to comply; and

5. Apache is estopped from claiming that Castex failed to comply because Apache made a false representation or concealed material facts to the detriment of Castex.

Apache now argues that these findings should be set aside and that a new trial should be ordered because Castex’s failure to pay was not excused and because Castex actually owed more than $10 million, despite the joint stipulation. Apache acknowledges that to obtain relief on this issue, it must successfully challenge both excuse findings.

We start with the first excuse finding, and we begin by noting that the charge did not ask the jury whether Apache’s prior breach was “material.” Instead, the charge merely asked whether Apache had committed a prior breach. Apache never objected to the omitted element of materiality.

Apache argues that the jury’s finding of prior breach cannot stand because there is legally insufficient evidence that Apache engaged in willful misconduct. This argument lacks merit because the absence of willful misconduct (or gross negligence) does not mean that a breach did not occur. A breach can occur simply when the operator fails to conduct its operations in a good and workmanlike manner.

Of course, the mere occurrence of such a breach does not mean that the operator will necessarily face liability. Under the terms of the joint operating agreement, the non-operator can only recover damages from the operator if the damages resulted from the operator’s gross negligence or willful misconduct. If the breach was the result of ordinary negligence, then the non-operator has no recourse. See Reeder v. Wood County Energy, LLC, 395 S.W.3d 789, 796 (Tex. 2012) (emphasizing that under the model form agreement the liability of an operator “requires more than a mere breach; rather it requires a breach attended by gross negligence or willful misconduct”).

Apache has not challenged the finding that it failed to conduct its operations in a good and workmanlike manner. Nor has Apache challenged whether there is

sufficient evidence that its breach occurred before Castex’s breach. Any such challenge has therefore been waived.

Apache proposes one additional reason for setting aside the finding of prior breach. In a reply brief, Apache argues that the model form exculpatory clause precludes a non-operator from defending a claim of non-payment on the basis of an operator’s prior breach, unless the operator engaged in gross negligence or willful misconduct. We need not consider this argument because it was raised for the first time in a reply brief, which means that it was waived. See Metro. Transit Auth. of Harris Cnty. v. Douglas, 544 S.W.3d 486, 498 (Tex. App.—Houston [14th Dist.] 2018, pet. denied) (“Arguments raised for the first time in a reply brief are waived.”).

Because Apache has not provided any basis for setting aside the jury’s finding of prior breach, we need not consider Apache’s remaining arguments on the excuse of estoppel.

IV. APACHE CONTRACTS

Apache and Castex entered into more than a dozen additional agreements concerning projects other than the facility at Belle Isle and the well at Potomac. These agreements were collectively identified during the trial as the “Apache Contracts,” and Apache alleged that Castex had failed to pay certain amounts that were due under them. Castex disputed the allegation, and the jury found that Castex had not failed to pay anything. Apache now argues that the jury’s finding should be set aside and that a new trial should be ordered because Apache conclusively proved that Castex has withheld payments under the Apache Contracts.

As the party attacking an adverse finding on an issue for which it bore the burden of proof at trial, Apache must demonstrate on appeal that the evidence

establishes, as a matter of law, all vital facts in support of the issue. See Dow Chem. Co. v. Francis, 46 S.W.3d 237, 241 (Tex. 2001) (per curiam).

Apache contends that it satisfied this burden, largely for two reasons. First, there was testimony from Deanna Doxakis, Apache’s manager of revenue accounting and collections, that Castex owed more than $15 million under the Apache Contracts. And second, there was testimony from Caran Crooker, Castex’s controller, that Castex withheld roughly $18 million in payments.

Apache contends that Crooker “freely admitted” Castex’s liability under the Apache Contracts, but the record is not so clear. Crooker said that Castex had withheld $18 million “at one point,” and then she touched on the subject of disputed bills. The questioning with Apache’s counsel proceeded as follows:

Q. You can dispute an expense? A. Correct. Q. And if you dispute an expense, you don’t have to pay it while the dispute is pending? A. Some people would argue that, but yes. Q. Okay. But in the case of the $18 million that Castex didn’t pay to Apache, those weren’t any expenses that were in dispute, were they? A. Well, maybe. It depends on how you define “dispute.” Q. Okay. Fair enough. I define “dispute” for this question as when you got a bill from one of these other 13 joint operations, you weren’t saying to Apache, “We’re not going to pay you because of something you did on that particular operation?” A. That’s correct. We were—yes, that’s correct.

This last response was ambiguous. We cannot determine if Crooker was indicating that Castex had a dispute but Castex did not notify Apache of the dispute in the manner described by counsel, or if she was indicating that the $18 million in

withheld payments were not actually in dispute. Crooker did not clarify her answer in her follow-up questions, and she did not otherwise testify that Castex was currently withholding payments that it knew were due and not in dispute. This testimony does not conclusively prove Castex’s liability.

Also, the jury had a substantial basis for not crediting Doxakis and her claim that Castex owed more than $15 million. Doxakis arrived at that calculation by running a computer-generated report, and that report depended on joint interest billing statements that Doxakis never personally reviewed.

Doxakis also created a summary of records which was presented to the jury as a demonstrative exhibit, but never formally admitted into evidence. The exhibit was included in our appellate record though, and it consists of a bare list of contract names and an amount allegedly owed under each contract. The largest amount under this list exceeds $10 million, and it corresponds to a single contract that is merely identified as “154011.” There is no contract bearing that same name or number under the list of contracts defined in the jury charge as one of the “Apache Contracts.” Apache did not explain that discrepancy.

Based on this record, we cannot say that Apache proved Castex’s liability as a matter of law.

V. CONCLUSION

We modify the trial court’s judgment by deleting Castex’s recovery of damages, prejudgment interest, and postjudgment interest arising out of Apache’s breach of the Potomac joint operating agreement (which appears in the judgment at paragraphs c, d, g, h, and k), and we affirm the judgment as so modified.

/s/ Tracy Christopher Chief Justice

Panel consists of Chief Justice Christopher and Justices Jewell and Zimmerer.