delivered the opinion of the Court

Eight Virginia limited partnerships hired Arkoma Basin Exploration Company^[1] to estimate production from mineral properties in the Arkoma Basin in southeastern Oklahoma. When the properties failed to produce as predicted, they sued.

Based on Virginia law, a Texas jury found clear and convincing evidence of fraud and awarded $5.5 million in damages. The trial court signed a judgment reducing the verdict to $4.7 million, and later reduced that further by remittitur to about $2.9 million. When Arkoma appealed the judgment and the partnerships cross-appealed the remittitur, the court of appeals affirmed in all respects but one, holding part of the remittitur improper and restoring about $1.5 million of the jury’s verdict.^[2]

We granted Arkoma’s petition to consider whether there was legally sufficient evidence of fraud under Virginia law, or of damages under Texas law. Finding that only two of the eight limited partnerships met these standards, we affirm the court of appeals’ judgment as to FMF Associates 1988-B, Ltd. and FMF Lazare, Ltd., and reverse the remainder.

I. Evidence of Fraud: Statement of Fact or Opinion?

Arkoma argues that its reserve estimates are immune from any fraud claim under Virginia law. The parties agree that Virginia law governs this issue, and requires clear and convincing evidence to establish liability.^[3] As this heightened standard is more substantive than procedural, we apply it in our legal sufficiency review.^[4]

[*384] The parties agree that reserve estimates do not attempt to calculate the volume of gas underground, but the volume that can be economically produced from a reservoir in the future.^[5] As a result, reserve estimates inherently include analysis and assumptions about future events.

Virginia law draws a line between statements of opinion and of fact: “[t]he mere expression of an opinion, however strong and positive the language may be, is no fraud.”^[6] Additionally, Virginia law distinguishes between statements of existing and future facts: “fraud must relate to a present or a pre-existing fact, and cannot ordinarily be predicated on unfulfilled promises or statements as to future events.”^[7]

But the line between these categories is not as clear as one might expect from these statements. Virginia specifically eschews a bright-line test, judging each case on its facts, and considering the nature of the representation, the relative knowledge of the parties, their intentions, and all of the surrounding circumstances before deciding whether a statement could constitute fraud:

We have not, however, established a bright line test to ascertain whether false representations constitute matters of opinion or statements of fact. Rather, each case must in a large measure be adjudged upon its own facts, taking into consideration the nature of the representation and the meaning of the language used as applied to the subject matter and as interpreted by the surrounding circumstances.... It is not always an easy matter to determine whether a given statement is one of fact or opinion. The relative knowledge of the parties’ dealing, their intentions and all of the surrounding circumstances, which can only be gathered from the evidence, affect the interpretation which the courts put upon the representations in determining whether they be of fact or opinion.^[8]

Thus, in some circumstances Virginia law allows fraud claims based on what might otherwise appear to be opinions. For example, the Virginia Supreme Court has held that a consultant who reported “nothing to indicate that wetlands are present,” but also warned that this was a matter of opinion, could nevertheless be liable for fraud when 80% of a property[*385] was later designated as wetlands.^[9] The same court held a fraud claim could be based on a realtor’s opinion that there was no termite damage, when in fact he had in hand a report suggesting the opposite.^[10] And a builder’s opinion that a house was “free from structural defects” could support a fraud claim, but his opinion that “no significant work would be required” could not.^[11] In each of these cases, the nature of the statement, the relative knowledge of the parties, and all the surrounding circumstances were taken into account in deciding whether a statement that appeared to be an “opinion” was nevertheless actionable as fraud.

Similarly, statements about future events may constitute fraud under Virginia law in some circumstances. Thus, for example, a sales agent’s assurance that the woods behind a condominium would never be cleared constituted fraud when she knew about plans to build a playground there in the future.^[12] And the Fourth Circuit has held that a doctor stated a fraud claim under Virginia law by alleging his malpractice insurer falsely assured him that a settlement would not affect his future ability to obtain coverage.^[13] Here again, in each case the nature of the statement, the parties’ relative knowledge, and all the surrounding circumstances were considered before deciding whether a statement about future events was nevertheless actionable.

Applying these principles to this case, there is no question Arkoma had superior knowledge about likely reserves, as it was hired for precisely that reason. Arkoma represented that its database was “without equal” and “unique” in its ability to combine geologic, engineering, and economic data to assess the potential value of mineral properties.

But Arkoma’s reserve estimates cannot all be treated alike because their nature and the circumstances surrounding them were very different. The reserves prepared for two partnerships — 1988-B and Lazare — concerned mineral interests in the Wilburton field, a mature field where scores of wells had been producing natural gas since 1960. Here, Arkoma estimated production from anticipated “infill” wells drilled between existing wells for which there was a long track record of production. Experts testified that these wells could reasonably be expected to produce from the same strata at similar historical rates, and that reserve estimates in this area could not reasonably vary by more than 10 or 15 percent. Yet there was clear and convincing evidence that Arkoma’s estimates assumed that infill wells would last far longer, be much more productive, and yield 200 to 300 percent more than nearby existing wells had ever produced.

By contrast, Arkoma’s reserve estimates for the remaining six partnerships concerned mineral interests in the South Panola field, a new field in which there had been little drilling or production. Interests here were purchased “ahead of the play” — in areas where minerals had not yet been found, but might be. The plaintiffs’ own expert conceded there was less than a year of production history anywhere in this field, and that an entirely[*386] different method of calculating reserves had to be used because no one knew how much wells in this area might eventually produce. Further, while he characterized drilling in the Wilburton field as “low risk,” he characterized that in South Pano-la as “definitely high” risk, which can only mean that reserve estimates in the latter were much more speculative. There was also undisputed evidence that in the decade after Arkoma’s estimates, major oil companies had drilled very expensive wells (at $4 million or more each) unsuccessfully looking for the same reserves in the same places that Arkoma had predicted. Viewing all the surrounding circumstances, there was no clear and convincing evidence that Arkoma’s estimates in South Panola should be treated as statements of fact rather than of opinion.

Our conclusion that reserve estimates in these two fields must be treated differently is supported by DeJarnette v. Thomas M. Brooks Lumber Co., in which a seller’s estimate that standing timber would produce three million board feet of lumber was held to be a statement of opinion rather than fact.^[14] The opinion did not suggest that estimates of natural resources must always be one or the other; to the contrary, it turned on specific evidence in that case that the buyer knew the seller’s estimate was not based on a detailed survey, and had made his own inspection with an experienced “timber cruiser” before buying. By focusing on how rough or specific the parties thought the defendant’s estimate was, Virginia law would appear to require different treatment for estimates in mature and immature gas fields like those here.

We disagree with the dissent that Virginia law exonerates opinions only if they are “good-faith opinions.” Because Virginia law requires proof of scien-ter in every fraud case,^[15] honest statements are never actionable whether they are opinions or not. In affirming the fraud verdict as to both fields and all eight partnerships, the court of appeals pointed to evidence suggesting Arkoma’s engineer was pressured to inflate some reserve estimates, and may have manipulated some spreadsheet calculations to reach a predetermined result. This is surely some evidence of scienter, but the fact/opinion distinction focuses on reliance. If circumstances show an estimate should be treated as a mere opinion, it is unreasonable to rely on the estimate as a fact whether or not it was a lie.^[16]

Accordingly, we reject arguments that Virginia law would uniformly treat all reserve estimates — regardless of the circumstances — as mere opinions (Arkoma’s view) or as statements of fact (the partnerships’). Instead, viewing all the surrounding circumstances in a light favorable to the verdict, and keeping in mind Virginia’s clear and convincing evidence standard, we hold that Arkoma’s reserve estimates in the mature Wilburton field were actionable as statements of fact, while those in the South Panola field were nonactionable statements of opinion. We affirm the jury’s fraud verdict as to the 1988-B and[*387] Lazare partnerships, and reverse the remainder.

II. Evidence of Damages: Preservation and Legal Sufficiency

Arkoma also challenges the legal sufficiency of the damages evidence. For the reasons stated above, we restrict our review to the two partnerships who proved fraud. They argue Arkoma failed to preserve this error in the trial court because Arkoma’s no-evidence objection (1) did not specify why the evidence was legally insufficient, and (2) was not raised before or during trial. As these preservation points concern procedural matters, Texas law governs.^[17]

To assert a no-evidence complaint in this Court, a party must preserve error in both the trial court and the corut of appeals.^[18] The court of appeals held that Arkoma’s objections in the trial court were not “specific enough to call the trial court’s attention to the precise lack of sufficiency asserted on appeal.”^[19] We disagree, for two reasons.

First, Arkoma’s post-trial motion explicitly asserted that “there is no evidence ... to support the jury’s answers to each part of Question 4,” the damages question. Generally, a no-evidence objection directed to a single jury issue is sufficient to preserve error without further detail.^[20] Thus, as Justice Calvert wrote for this Court 50 years ago, while a single such objection to all 79 jury answers is too general, the same objection addressed to each individual issue is adequate.^[21] Several commentators suggest this is precisely what careful practitioners should do.^[22]

Second, the cardinal rule for preserving error is that an objection must be clear enough to give the trial court an opportunity to correct it.^[23] Here, the trial judge not only had that opportunity, he took it. The trial judge conducted a post-[*388]trial hearing on the sufficiency of the damages evidence, received letter briefs on the issue, and wrote a four-page single-spaced letter granting remittitur. Though damages were not reduced as much as Arkoma had hoped, there is no question the trial court was aware of its objection.

Of course, stock objections may not always preserve error. If a single jury question involves many issues, it is possible that a general objection may not tell the trial court where to start.^[24] But post-trial objections will rarely be as detailed as an appellate brief because time is short, the record may not be ready, and the trial court is already familiar with the case. In that context, an objection is not necessarily inadequate because it does not specify every reason the evidence was insufficient.^[25] Like all other procedural rules, those regarding the specificity of post-trial objections should be construed liberally so that the right to appeal is not lost unnecessarily.^[26] We hold the court of appeals erred in holding that Arkoma’s objection was too general.

It is a closer question whether Arkoma had to object to the damages evidence during trial. It was the partnerships’ burden to prove damages by expert testimony, as the value of mineral reserves is not a matter of common knowledge.^[27] Texas law requires an objection to expert testimony before or during trial if the objection “requires the court to evaluate the underlying methodology, technique, or foundational data,” but no objection is required if the complaint “is restricted to the face of the record,” as when the complaint is that an opinion was speculative or eon-clusory on its face,^[28] or assumed facts contrary to those on the face of the record.^[29] Of course, some objections will fall close to the line between these categories. But we[*389] need not categorize Arkoma’s objection, because even if no objection was necessary the expert testimony here was legally sufficient.

The 1988-B and Lazare partnerships called Michael Harper, a petroleum engineer, to calculate the amount by which reserves were fraudulently inflated. Arko-ma does not object to the volumes he estimated for the exaggerated reserves; its sole challenge is that when Harper multiplied those volumes by prices, he failed to discount the total.

Arkoma’s argument assumes Harper was using the income approach to value without saying so. In that approach, estimated future income is discounted by a capitalization rate (reflecting both risk and the time value of money) to reach a present value.^[30] Arkoma argues that no reasonable jury could credit testimony that a risky return of $100 in the future had a fair value of $100 now.

But Harper never purported to use the income approach. While his testimony about his calculations was cursory, a demonstrative exhibit shows that he believed the price reasonable investors would pay for mineral interests was the total dollars they expected to receive in the next eight years without any adjustment for risk, inflation, or interest rates. Although his calculations did not include an explicit discount rate, they included an implicit return for risk and interest because the wells here were expected to produce for substantially more than eight years. While failing to use an explicit discount rate might undermine expert testimony in other cases, we note that “payouts” in the oil and gas business are often calculated in precisely this manner.^[31]

Arkoma is certainly correct that Harper’s testimony could have been a lot clearer; his references to “up here” and “right there” on slides and posters used at trial often make it hard to tell what he is talking about. But we cannot say on this record that his opinions were unreliable or speculative. Nor were they conclusory as a matter of law; Harper did not simply state a conclusion without any explanation,^[32] or ask jurors to “take my word for it.”^[33] It is true that without the foundational data in the appellate record, we cannot confirm that “cash off my runs ... divided by mcf” yielded the $1.62, $1.41, $1.48, and $1.59 prices he calculated as the low range for damages. But experts are not required to introduce such foundational data at trial unless the opposing party[*390] or the court insists.^[34]

Accordingly, we reject Arkoma’s claim that evidence of the damages suffered by the 1988-B and Lazare partnerships was legally insufficient.

III. Deadlines for Appeal: From Judgment or Remittitur?

In its original judgment, the trial court reduced the jury’s damages verdict for the 1988-B partnership to $2,090,000 and for Lazare to $930,000, a matter they never appealed. But they did challenge by cross-appeal the trial court’s remittitur reducing damages further to $1,302,302 for 1988-B and $579,492.76 for Lazare. The court of appeals reversed the remittitur and reinstated the former amounts.^[35] Ar-koma asserts the court had no jurisdiction to reinstate anything because the partnerships’ cross-appeal was two days late.

The trial court’s original judgment was signed January 29, 2002. Arkoma’s motion for new trial extended the time for appeal to April 29th, 90 days after judgment.^[36] "When Arkoma filed its notice of appeal on the last possible day, the deadline for a cross-appeal was further extended an additional 14 days to May 13th.^[37] The partnerships filed two days later, May 15th.

But the partnerships argue their cross-appeal was timely because the original judgment was modified by remittitur. “If a judgment is modified in any respect,” appellate deadlines do not run from the original judgment but “from the date when the modified judgment is signed.”^[38] According to the partnerships, the trial court’s order granting remittitur on April 15, 2002 restarted the appellate timetables, rendering their cross-appeal filed 30 days later timely.^[39]

Technically, neither trial nor appellate courts can order remittitur; they can only suggest a remittitur on condition that a new trial will be granted if it is refused.^[40] We have held that if a court directly orders a reduction in damages, that order necessarily modifies the judgment even if it is incorrectly called a re-mittitur.^[41] By contrast, we have also held that if a party files a purely voluntary remittitur (without any order or suggestion from the court), the judgment has not been modified.^[42] If the latter rule were otherwise, a party could extend appellate deadlines indefinitely by remitting a dollar at a time.

We have never addressed whether an order that properly suggests remittitur modifies a judgment “in any respect,” and thus restarts appellate deadlines. But we have explained that the deadlines are re[*391]started by “any change, whether or not material or substantial.”^[43] Thus, appellate deadlines are restarted by an order that does nothing more than change the docket number or deny all relief not expressly granted.^[44]

By this standard, we think a signed order suggesting remittitur must restart the appellate deadlines as well. By its very nature, such an order allows only two options: a smaller judgment or a new trial. While it may not be clear when the order is signed which option a claimant will select, it is immediately clear that the original judgment will change. In many cases, the order itself will immediately change which party, if any, should begin preparing for an appeal. Given the relative impacts of an order adding a docket number and an order suggesting remitti-tur, it would be anomalous if the former restarted appellate deadlines but the latter did not.

It is conceivable, of course, that a judgment might be reinstated after a suggestion of remittitur because the trial judge withdraws the latter. But that does not mean the judgment was not modified “in any respect” in the interim; a trial judge who modifies a judgment and then withdraws the modification has modified the judgment twice rather than never.

Because appellate timetables restarted when the trial court signed its remittitur order, the court of appeals had jurisdiction to consider the remittitur. Because Arko-ma asserts no other objection to it, it is affirmed.

IV. Conclusion

For the reasons stated above, we hold that two of the eight limited partnerships presented legally sufficient evidence of fraud under Virginia law and of damages under Texas law. Accordingly, we affirm the court of appeals’ judgment as to FMF Associates 1988-B, Ltd. and FMF Lazare, Ltd. We reverse its judgment and render a take-nothing judgment against FMF Associates 1990-A, Ltd., FMF HMY, Ltd., FMF Kahn, Ltd., FMF Friedman, Ltd., FMF Greenwald, Ltd., and FMF Small-wood, Ltd.

Justice O’NEILL filed a concurring and dissenting opinion.