Justice Lehrmann delivered the opinion of the Court.

In this case, we examine whether a judgment against an insured defendant was the product of a "fully adversarial trial" and is thus enforceable against the defendant's insurer. Homeowners sued their builder for failing to construct their home in a good and workmanlike manner, and the builder's commercial general-liability insurer wrongfully refused to defend the builder in that suit. The case went to trial, resulting in a judgment in the homeowners' favor. The builder subsequently assigned most of its claims against its insurer to the homeowners, who now seek to recover the judgment from the insurer under the applicable insurance policy. We are asked whether the judgment against the builder is binding on the builder's insurer in this suit. In the event it is not, we are also asked whether the deficiencies in the underlying trial were effectively remedied by this subsequent insurance litigation. The court of appeals answered yes to the first question and affirmed the trial court's judgment in the homeowners' favor. We answer no to the first. As to the second, we hold that this insurance litigation may serve to determine the insurer's liability, although the parties in this case understandably focused on other issues during the trial. Accordingly, we reverse the court of appeals' judgment and, in the interest of justice, remand to the trial court for a new trial.

I. Background

A. The Damage Suit

Glen and Marsha Hamel own a single-family home in Flower Mound, Texas. The Hamels hired a contractor, GSM Corporation (the Original Builder), to build the home in the mid-1990s, but the Original Builder abandoned the project before completion. The Hamels then hired Terry Mitchell Builders, Inc. (the Builder) to finish the home, which was completed in October 1995.1 Terry Mitchell is the president and sole owner of this company.

The home's exterior was finished with an Exterior Insulation and Finish System (Exterior Stucco), which is a type of synthetic stucco cladding that can cause wood rot and other problems relating to water damage if installed improperly or if defective materials are used. In August 2000, the Hamels noticed signs of water damage in the home, including stained walls and warped baseboards. They sued the Builder in April 2002 for breach of implied warranty, negligence, Deceptive Trade Practices Act violations, and Residential Construction Liability Act violations, alleging that the Builder failed to perform its services in a good and workmanlike manner.2 In their original petition, the Hamels alleged that the water damage resulted from the improper use or installation of the Exterior Stucco. They subsequently amended the petition to attribute the water damage to the home's improper construction or, alternatively, the use of Exterior Stucco on the home.

Great American Insurance Company insured the Builder under commercial general-liability insurance policies, issued on an annual basis. The first three policies, effective May 3, 1996, to May 3, 1999, did not exclude damage relating to Exterior Stucco. However, the fourth and fifth policies, effective May 3, 1999, to May 3, 2001, excluded property damage "arising out of" Exterior Stucco.

The Builder notified Great American of the Hamels' suit (the Damage Suit), but Great American declined to defend the Builder, citing the fifth policy's Exterior-Stucco exclusion (effective May 3, 2000, to May 3, 2001). Great American took the position that this was the applicable policy because the Hamels' August 2000 discovery of the damage fell within that period. See Don's Bldg. Supply, Inc. v. OneBeacon Ins. Co., 267 S.W.3d 20, 26 (Tex. 2008) (explaining that some Texas courts had chosen to follow "a `manifestation rule' that imposes a duty to defend [on insurers] only if the property damage became evident or discoverable during the policy term"). However, Great American now concedes that this position was erroneous. See id. at 25 (clarifying that Texas follows the "`actual injury' or `injury-in-fact' approach, [under which] the insurer must defend any claim of physical property damage that occurred during the policy term"). Great American also concedes that, in light of the Hamels' allegations in the Damage Suit, Great American wrongfully refused to defend the Builder in that suit. See GuideOne Elite Ins. Co. v. Fielder Rd. Baptist Church, 197 S.W.3d 305, 310 (Tex. 2006) ("A plaintiff's factual allegations that potentially support a covered claim is all that is needed to invoke the insurer's duty to defend....").

Without the benefit of insurance coverage, the Builder had limited assets to fund its defense. In fact, shortly before trial, the Builder terminated its counsel, Robert Hudnall, for financial reasons, and Hudnall prepared a motion to withdraw. However, the trial court apparently never heard that motion, and Hudnall continued to represent the Builder during and after trial.

In May 2005, a week before trial, the Hamels entered into a Rule 11 agreement with the Builder. The Hamels agreed that, in the event they obtained a judgment against the Builder, they would not attempt to pierce the corporate veil and enforce the judgment against the Builder's owner, Mitchell, individually. They essentially agreed to enforce any judgment only against assets in the company's name, excepting any "personal tools of the trade and truck," which the Hamels agreed not to pursue "even if in [the company's] name." Mitchell would later testify that, at the time the agreement was executed, the company had no assets beyond the excepted "tools of the trade and truck." For his part, Mitchell agreed to appear at the scheduled trial and not to seek a continuance, and the Hamels contend that securing Mitchell's trial appearance was the reason they entered into this agreement.

The day before trial, the Builder executed stipulations of fact in lieu of responding to the Hamels' outstanding requests for admissions. The stipulations included, inter alia:

These stipulations demonstrate a shift from the position the Builder took in discovery responses served earlier in the suit, in which the Builder had asserted that the Hamels' claims "relate to areas or matters for which [the Builder] was not paid . . . by [the Hamels] and for which [the Builder] had no responsibility or control."

The Damage Suit proceeded to a bench trial on May 26, 2005. Neither the stipulations nor the Builder's written construction contract with the Hamels was offered as an exhibit. The Hamels called Mitchell, who testified consistently with the stipulations that the Builder "agreed to make sure the house was finished in a good and workmanlike manner," which included an "obligation to inspect all of [the Original Builder's] work and make sure there weren't any problems." He also testified that, in inspecting and completing the home, he did not notice the "issues that the Hamels have had with their house," including steel nails in the roofing system, a short roof deck, lack of a drip edge, inadequate securing of a fascia board to the framing, improper framing of the roof ridge and second-floor window opening, improper roof sloping and drainage, and shower leaks. He testified that these were "honest mistake[s]" and that the failure to discover such problems would amount to a failure to complete the home in a good and workmanlike manner. He further testified that these problems had nothing to do with the Exterior Stucco.

Donald Yeandle, a contractor whom the Hamels hired in 2002 to inspect the home and evaluate the extent of the water damage, testified about the existence of the various problems laid out during Mitchell's questioning and the resulting water damage. He also testified that these problems were unrelated to Exterior Stucco or its components and opined that the Builder should have noticed the problems or at least performed a more thorough inspection. He concluded that the Builder did not complete the home in a good and workmanlike manner. Both Yeandle and Glen Hamel testified about damages. The Builder presented no witnesses.

The trial court accepted the Hamels' attorney's suggestion that the parties submit proposed findings of fact and conclusions of law in lieu of closing arguments; however, only the Hamels submitted proposed findings. The trial court rendered judgment in the Hamels' favor and adopted their proposed findings without modification, awarding them $365,089 in damages — composed of $169,089 in repair costs, $100,000 in loss of market value due to stigma, $50,000 in mental-anguish damages, $15,000 in costs to repair landscaping that would be damaged during the home repair, $24,000 in temporary housing costs, and $7,000 in moving costs — plus prejudgment interest and court costs. The Builder subsequently assigned most of its rights against Great American to the Hamels.

B. The Insurance Suit

The Hamels, as the Builder's assignees and judgment creditors, brought the current suit (Insurance Suit) against Great American for breach of contract and declaratory relief, seeking to recover the judgment from the Damage Suit under the Builder's insurance policy.3 They also initially asserted claims for Texas Insurance Code violations, but abandoned those claims before trial. The Insurance Suit was tried to the bench. The entire record from the Damage Suit was introduced into evidence, as were the stipulations, the Rule 11 agreement, and the contract between the Builder and the Hamels. Excerpts from the depositions of Mitchell, Glen Hamel, and their respective attorneys were also admitted. The trial court heard live testimony from both the Hamels' and Great American's expert witnesses.

The trial court rendered judgment for the Hamels and entered findings of fact and conclusions of law, including, in pertinent part:

The judgment in the Insurance Suit awarded the Hamels covered damages in the underlying Damage Judgment of $355,838, plus interest, court costs, and attorney's fees.

Great American appealed. It argued in pertinent part that the Damage Judgment was not binding on Great American under this Court's holding in State Farm Fire & Casualty Co. v. Gandy prohibiting enforcement of such judgments, if rendered without a fully adversarial trial, in an action by the plaintiff as the insured's assignee. 925 S.W.2d 696, 714 (Tex. 1996). The court of appeals affirmed the trial court's judgment in most respects, holding that Great American breached its duty to defend the Builder from the Hamels' suit,4 the Damage Judgment was the result of a fully adversarial trial, and the Builder's assignment of its claims against Great American to the Hamels was valid. 444 S.W.3d 780 (Tex. App.-El Paso 2014). However, the court of appeals reversed the portion of the judgment awarding mental-anguish damages, holding they were not compensable as a matter of law. Id. at 812.5 We granted Great American's petition for review.

II. Discussion

We have said that, generally, an insurer that wrongfully refuses to defend its insured is barred from collaterally attacking a judgment or settlement between the insured and the plaintiff. Evanston Ins. Co. v. ATOFINA Petrochemicals, Inc., 256 S.W.3d 660, 671 (Tex. 2008); Emp'rs Cas. Co. v. Block, 744 S.W.2d 940, 943 (Tex. 1988). However, as discussed below, in Gandy we narrowed the scope of that rule under certain circumstances in which the plaintiff seeks to enforce the judgment against the insurer as the insured's assignee. 925 S.W.2d at 714. Great American argues that the circumstances of this case preclude enforcement of the Damage Judgment against it.

A. Gandy

Because the resolution of the issues hinges on the proper reach and application of Gandy, we begin with a discussion of that case. Gandy sued her stepfather for damages relating to sexual abuse. Id. at 697. State Farm, which had previously issued the stepfather a homeowner's policy, was notified of the suit and agreed to provide a defense under a reservation of rights. Id. at 698-99. Despite State Farm's participation, the stepfather settled the case without notice to State Farm. Id. at 698. As part of the settlement, the stepfather assigned his claims against State Farm to Gandy, who agreed that she would not attempt to collect the agreed judgment from the stepfather. Id. The plaintiff then sued State Farm to collect the judgment and for breach of the duty to defend. Id.

We held that the assignment violated public policy and was void, noting two principal defining characteristics of the overall settlement: (1) it served to prolong the litigation rather than end it; and (2) it distorted the litigation, causing the parties to take "positions that appeared contrary to their natural interests for no other reason than to obtain a judgment against State Farm." Id. at 712. We outlined the following general rule:

Id. at 714. We expressly declined to address "whether an assignment is also invalid if one or more of these elements is lacking." Id. But we independently concluded: "In no event, however, is a judgment for plaintiff against defendant, rendered without a fully adversarial trial, binding on defendant's insurer or admissible as evidence of damages in an action against defendant's insurer by plaintiff as defendant's assignee." Id.

Great American contends that the Damage Judgment was not the product of a fully adversarial trial and that Gandy therefore precludes the judgment's enforcement against Great American by the Hamels as the Builder's assignees. Specifically, Great American argues that the pretrial agreement and stipulations entered into by the Hamels and the Builder — which were not presented to the trial court in the Damage Suit — ensured that the Builder had no real stake in the trial's outcome. According to Great American, this resulted in a sham trial shaped entirely by the Hamels and designed to aid in the prosecution of the subsequent insurance litigation. The Hamels respond that no evidence of fraud or collusion exists with respect to the Damage Trial and that Great American improperly seeks to extend Gandy's holding beyond its intended scope. With these arguments in mind, we examine Gandy's impact on this case.

B. Enforceability of Damage Judgment against Great American

As an initial matter, we note that the validity of the Builder's assignment of its claims against Great American to the Hamels is not at issue here.6 As Great American recognizes, the circumstances underlying our invalidation of the assignment in Gandy are lacking in this case. First, the Builder assigned its claims following, not preceding, a trial and judgment. Second, unlike the insurer in Gandy, Great American breached its duty to defend. Third, Great American neither accepted coverage nor made a good-faith effort to adjudicate coverage before the Hamels' claims against the Builder were resolved. We have stated that Gandy's holding "was explicit and narrow, applying only to a specific set of assignments with special attributes." ATOFINA, 256 S.W.3d at 673. Although in Gandy we declined to address whether an assignment that lacked one or more of Gandy's characteristics could be invalid, we see no reason to invalidate an assignment when none of those characteristics are present. Great American took a significant risk by refusing to defend, or at least litigate its duty to the Builder. See Gandy, 925 S.W.2d at 714 (noting that "[d]isputes between [the insurer] and [insured defendant] can often be expeditiously resolved in an action for declaratory judgment while [the plaintiff's] claim is pending," and that insurers often "will assume the burden of having the issues resolved" to prevent undue burden on the insured). We therefore confirm that the Builder's assignment of its claims against Great American to the Hamels was valid.

But that does not end the inquiry. Great American argues that, separate and apart from the assignment's validity and regardless of Great American's failure to defend, Gandy precludes the Hamels from enforcing the Damage Judgment against Great American in the Insurance Suit. Great American relies on our statement in Gandy that "[i]n no event . . . is a judgment for plaintiff against defendant, rendered without a fully adversarial trial, binding on defendant's insurer or admissible as evidence of damages in an action against defendant's insurer by plaintiff as defendant's assignee." Id. In so holding, we "disapprove[d] the contrary suggestion" in Block. Id. (citing Block, 744 S.W.2d at 943).

In Block, as in this case, the defendant's insurer wrongfully refused to defend in the underlying suit. 744 S.W.2d at 942. The parties settled, and the trial court rendered an agreed judgment. Id. The defendant sued its insurer for failure to defend under the policy, and the plaintiffs intervened as judgment creditors and assignees seeking recovery of the judgment. Id. We held that, in light of its failure to defend, the insurer "was barred from collaterally attacking the agreed judgment by litigating the reasonableness of the damages recited therein." Id. at 943. We revisited this issue in ATOFINA, another case in which the insurer breached its duty to defend in the underlying suit and then sought to challenge the parties' settlement as unreasonable. 256 S.W.3d at 670-74. We applied the Block rule notwithstanding Gandy's pronouncement that plaintiffs may not enforce an underlying judgment against the defendant's insurer absent a fully adversarial trial, explaining that "Gandy's key factual predicate [was] missing" because the insured defendant in ATOFINA had not assigned its claims but had sued the insurer directly. Id. at 673. We further noted that the "risk of distorting litigation or settlement motives" present in Gandy did not exist in ATOFINA because the insured defendant "settled without knowing whether or not it would be covered by the policy, leaving in place its motive to minimize the settlement amount in case it became solely responsible for payment." Id. at 674.

This case gives us the opportunity to clarify how our holdings in Block, Gandy, and ATOFINA apply in cases that do not match their exact factual circumstances. Block indicated that the insurer's breach of its duty to defend necessarily renders any covered judgment binding on the breaching insurer. 744 S.W.2d at 942-43. Since Gandy, however, an insurer's wrongful failure to defend is no longer dispositive. In holding that a plaintiff may not enforce an underlying judgment against the defendant's insurer absent a "fully adversarial trial," we shifted focus toward whether the underlying judgment accurately reflects the plaintiff's damages and thus the insured's covered loss. 925 S.W.2d at 714; see also ATOFINA, 256 S.W.3d at 673-74 (binding the insurer to a judgment arising from a settlement agreement, rather than a trial, primarily because the defendant retained a stake in the litigation even upon settlement).7

One way to ensure that a judgment accurately reflects the plaintiff's damages is to require that the loss be determined through a proceeding in which the parties "fully" — or at least actually and effectively — oppose and contest each other's positions. The difficulty with this approach, however, is in determining just how effective each party's trial performance must be. In this case, for example, the parties present diametrically opposed positions on whether the Damage Judgment resulted from a fully adversarial trial. In arguing that the underlying trial was not fully adversarial, Great American relies on the pretrial agreement allegedly removing the Builder's incentive to defend and subsequent conduct purportedly conforming to that state of affairs, including the Builder's abandonment of earlier-asserted defenses, stipulations as to liability, and minimal trial participation. Great American accuses the parties of secrecy and collusion, noting that the pretrial agreement and stipulations were not put into evidence, nor was the trial court advised of their existence. Great American also engages in a lengthy and detailed critique of the Builder's trial strategies, accusing the Builder of presenting a sham defense. The Hamels respond that the Builder "presented the best defense it could in light of the undeniable facts," noting that the Builder participated in the trial and was represented by an attorney who engaged with the court and the witnesses.

The court of appeals rejected Great American's arguments and concluded:

444 S.W.3d at 803. The court of appeals' approach necessarily requires courts to retroactively evaluate and thus second-guess trial strategies and tactics, which — as we have noted in other circumstances — often produces an inaccurate and unreliable result. C.f., e.g., Cantey Hanger LLP v. Byrd, 467 S.W.3d 477, 481 (Tex. 2015) (noting the general rule that "attorneys are immune from civil liability to non-clients for actions taken in connection with representing a client in litigation" (citations and internal quotation marks omitted)); In re JFC, 96 S.W.3d 256, 283 (Tex. 2002) (noting the difficulty of overcoming the presumption that trial counsel's acts and omissions are based on strategy in claims of ineffective assistance of counsel). Every trial presents unique challenges, requiring subjective judgment calls that may seem in hindsight to have been ill-advised. But determining whether and when those calls destroy the "adversarial" nature of the proceeding is simply not possible. Great American's criticism of the Builder's trial strategy here is particularly troubling given that it had the opportunity to control the defense in the first instance and wrongfully refused to do so.

This misplaced focus on trial details likely results from a misinterpretation of the phrase "fully adversarial" as it was used in Gandy. Fundamentally, proceedings are "adversarial" when the parties oppose each other. See Adversarial, WEBSTER'S NEW WORLD COLLEGE DICTIONARY (5th ed. 2014) (defining "adversarial" as "of or characterized by opposition, disagreement, hostility, etc., as between adversaries"). Today we clarify that the controlling factor is whether, at the time of the underlying trial or settlement, the insured bore an actual risk of liability for the damages awarded or agreed upon, or had some other meaningful incentive to ensure that the judgment or settlement accurately reflects the plaintiff's damages and thus the defendant-insured's covered liability loss.

Our holding in ATOFINA is consistent with this standard. In ATOFINA, the insurer wrongfully refused to defend the underlying claims, and the insured settled without assigning his rights against the insurer to the plaintiff. 256 S.W.3d at 673. More importantly, the insured retained the risk that he would be liable for the damages if the policy did not cover them. Id. at 674. That incentive to contest the plaintiffs' alleged damages was sufficient to ensure that the settlements accurately reflected the insured's covered loss, even without a trial. By contrast, in Gandy, the insurer agreed to provide a defense, and the insured assigned his claims against the insurer to the plaintiff before the case was resolved. 925 S.W.2d at 698. Further, the insured retained no financial risk or other incentive to oppose the plaintiff's damages. See id. In the absence of any such incentive, we explained, the insurer should not be bound by the damages determination. Id. at 714.

In this case, the parties' pretrial agreement eliminated any meaningful incentive the Builder had to contest the judgment. Specifically, before trial, the Hamels agreed not to enforce any resulting judgment against Mitchell's personal assets.8 They also agreed not to pursue Mitchell's truck or "personal tools of the trade . . . even if in the name of [the company]." While the agreement technically did not foreclose the Hamels from pursuing other assets of the Builder, Mitchell himself testified that the company had no other assets except the insurance policy.

When the parties reach an agreement before trial or settlement that deprives one of the parties of its incentive to oppose the other, the proceeding is no longer adversarial. Stated another way, proceedings lose their adversarial nature when, by agreement, one party has no stake in the outcome and thus no meaningful incentive to defend itself. When a plaintiff agrees to forgo execution of a judgment against a defendant's assets, whether in conjunction with a settlement or before trial, the defendant no longer has a financial stake in the outcome and thus likely has no interest in either avoiding liability altogether or minimizing the amount of damages. We believe adversity turns on the insured defendant's incentive to defend (or lack thereof), and an after-the-fact evaluation of the parties' trial strategies therefore has no place in the analysis.

Here, as noted and despite the Hamels' protestations, the pretrial agreement effectively removed any financial stake the Builder had in the outcome of the Damage Suit, thereby eliminating any incentive the Builder had to oppose the Hamels' claims. This turned the Damage Suit into a mere formality — a pass-through trial aimed not at obtaining a judgment reflective of the Hamels' loss, but instead at obtaining a potentially inflated judgment to enforce against Great American. Again, the Builder's only assets (other than the insurance policy) were "a pickup truck and some tools," and these were the assets the Hamels specifically agreed not to pursue. They also agreed not to attempt to pierce the corporate veil, ensuring that Mitchell's personal assets were protected. This left only the Builder's insurance policy as a potential source to satisfy any judgment obtained. Mitchell's testimony reflected the effect of this agreement on the Builder's incentive to defend itself:

In sum, the parties' pretrial agreement removed the Builder's stake in the out-come and any corresponding incentive to defend itself. After the agreement was executed, the Damage Suit no longer involved opposing parties, and the trial that followed was not fully adversarial. Accordingly, under Gandy, the Damage Judgment is not binding against Great American in the present suit brought by the Hamels as judgment creditors and assignees. See 925 S.W.2d at 714.

We do not suggest that a formal, written pretrial agreement that eliminates the insured's financial risk will always be either necessary or sufficient to disprove adversity. We hold instead that the presence of such an agreement creates a strong presumption that the judgment did not result from an adversarial proceeding, while the absence of such an agreement creates a strong presumption that it did. On the one hand, the insurer may overcome the presumption by demonstrating that, even though the plaintiff and insured defendant did not enter into any formal, written agreement, the evidence nonetheless establishes that the defendant had no meaningful stake in the outcome of the underlying litigation.9 Conversely, the plaintiff (acting as the defendant's assignee) may overcome the presumption by submitting evidence demonstrating that the defendant retained a meaningful incentive to defend the underlying suit despite an agreement that eliminated the defendant's financial risk.

In this case, as Mitchell expressly confirmed, the parties' pretrial agreement eliminated any incentive the Builder had to defend against the Hamels' claims. Without second-guessing subjective trial tactics and strategy decisions, we conclude that, in the absence of any such incentive, the Damage Trial was not fully adversarial and the resulting judgment is not binding on Great American.

C. Whether the Insurance Trial Cured the Lack of Adversity in the Damage Trial

The Hamels contend that, in the event we find a lack of adversity in the Damage Trial, any associated problems were cured by the Insurance Trial, at which Great American "had the opportunity to make its coverage arguments, examine witnesses, put on witnesses of its own and evaluate the damages." Great American responds that the problems associated with the Damage Trial are incurable, arguing, the "entire rationale of Gandy is that once adversity is destroyed and the defendant no longer has any incentive to oppose the plaintiff, it is impossible to go back and determine what might have been." Great American thus contends that it is entitled to rendition of judgment in its favor.

We acknowledged in Gandy that, in the context of attempting to assess a defendant's liability after settlement, "it is very difficult to determine what might have been" once the parties have changed positions. 925 S.W.2d at 719. We explained that "this inquiry should ordinarily be avoided, absent compelling reasons to the contrary." Id. (emphasis added). We believe an insurer's wrongful refusal to defend presents a compelling reason to engage in this endeavor despite its difficulty.

An insurer's refusal to defend or to even attempt to litigate its duties while the underlying suit is pending carries significant risks, and for good reason. See id. at 714. It places the burden on the insured to defend itself, often without adequate resources to do so. See id. It can also leave the plaintiff in an untenable position. The defendant's insurer is often the plaintiff's only real source of recovery, but without the insurer's involvement in the lawsuit the likelihood of a fully adversarial trial diminishes substantially. Those concerns were manifested in this case. To some degree, the parties' conduct is simply an attempt to make the best of a situation that Great American created by refusing to defend.

Accordingly, while we will not hold an insurer to a judgment that was not the result of an adversarial proceeding, we will not preclude the parties from properly litigating the underlying liability issues in a subsequent coverage suit.10 Although in Gandy we identified the difficulties inherent in this undertaking, we also emphasized the importance of determining an insurer's obligations before its insured incurs liability. Id. By declining to defend or litigate its duties early, an insurer plays a key role in making such a complicated endeavor necessary. Certainly, relitigation of underlying liability and damages issues is not a perfect solution, but it is necessitated by the circumstances. The insurer should not benefit from the problem that it helped create, as Great American's proposed solution — rendition of judgment in its favor — would allow. Rather, under the approach we adopt today, the insurer will have the opportunity to challenge its insured's underlying liability and the resulting damages, the abandoned insured is protected, and the burden on the plaintiff is fair. And of course, the insurer has every incentive to assert a strong defense during the Insurance Trial.

So did the parties effectively retry the Damage Suit in the Insurance Trial? To that, we must answer no. Although the trial court in the Insurance Suit made independent findings about the Builder's liability and held that the evidence supported the Damage Judgment's award of cost-of-repair damages, the fact remains that the arguments, evidence, and witnesses in the Insurance Trial were understandably geared toward other issues. With respect to coverage, the parties litigated primarily (1) when the property damage occurred, which affected whether the policies with Exterior-Stucco exclusions applied, and (2) whether the Hamels segregated covered and non-covered damages and properly allocated the damages among policy periods. The parties also presented evidence and testimony bearing on whether the Damage Trial qualified as "fully adversarial" by focusing on trial strategies and tactics rather than whether the parties were adverse. These overarching issues encompassed some, but not all, of the liability and damages issues adjudicated in the Damage Trial.

For example, in the course of litigating whether the Damage Trial was fully adversarial, the parties addressed portions of the evidence relating to the Builder's liability and the Hamels' damages. Great American introduced the original construction contract into evidence and elicited testimony about the scope of the Builder's contractual duties to inspect and remedy work performed by prior contractors. Great American was thus able to present one of the liability defenses purportedly ignored during the Damage Trial: that the Builder was not liable for the Original Builder's defective work.11 Great American also pursued its contention that the Hamels failed to mitigate their damages, cross-examining Yeandle on a report he had prepared in 2002 estimating repair costs in an amount significantly lower than he testified to at the Damage Trial. The trial court considered this evidence and found that "the 2002 report prepared by Mr. Yeandle, as he confirmed, . . . bears no relation to the full cost to repair estimated, that he testified to at the [Damage] Trial."

However, the parties did not thoroughly relitigate all aspects of the Hamels' claimed damages. Great American briefly questioned one expert witness about various deficiencies, eliciting opinion testimony that the Builder was entitled to a $25,000 settlement credit, that the damages attributed to landscape repair constituted a double recovery, and that Glen Hamel's testimony was incompetent as to the home's loss in market value and the Hamels' temporary housing and moving costs. But these issues were not pursued at length. The parties did not present formal motions or legal arguments regarding the potential settlement credit, and Glen Hamel was not questioned about damages even though he was the sole witness at the Damage Trial with respect to all damage categories except repair costs. Further, the court did not make specific findings on these categories of damages in the Insurance Suit.

In sum, although some of the Damage-Suit issues were indirectly raised in the Insurance Trial, we cannot say that it "cured" the problem. The scope of the Insurance Trial was simply not as broad as the Hamels suggest. However, we cannot fault the parties or the trial court for that. Prior to this opinion, Texas law was not clear as to Gandy's effect on the Damage Judgment's enforceability. Nor was it evident that the Insurance Suit provided a vehicle to remedy the problems associated with the lack of adversity in the Damage Suit.12 Accordingly, we believe a remand in the interest of justice is necessary. See USAA Tex. Lloyds Co. v. Menchaca, ___ S.W.3d ___, ___, 2017 WL 1311752 (Tex. 2017) (remanding in the interest of justice "[i]n light of the parties' obvious and understandable confusion over our relevant precedent and the effect of that confusion on their arguments in this case").

III. Conclusion

We hold that the Damage Judgment was not the product of a fully adversarial proceeding because the parties entered into an agreement that eliminated any meaningful incentive for the Builder to contest the Hamels' claims. As a result, the judgment that followed was not enforceable or admissible as evidence in the subsequent Insurance Suit against Great American by the Hamels as judgment creditors and as the Builder's assignees. However, we also hold that the Insurance Suit gives the parties the opportunity to litigate any disputed underlying issues with the benefit of full adversity. Although the parties did not do so here, they should be given the opportunity on remand. Accordingly, we reverse the court of appeals' judgment and remand the case to the trial court for a new trial.