Vascular Solutions, Inc. (‘VSI”) and Marine Polymer Technologies, Inc. (“MPT”) sell differently designed medical patches that are designed to stop bleeding after specific medical procedures, including catheterizations. Each accused the other of spreading false statements about its particular product. Awarding nothing to MPT, the jury found VSI was entitled to damages of $4.5 million in lost profits on its product disparagement claim, and it is from that judgment that MPT now appeals.
The background facts can be briefly summarized, recounted in light of the jury’s verdict. Mendez-Matos v. Municipality of Guaynabo, 557 F.3d 36, 40 (1st Cir.2009). Both MPT’s “SyvekPatch” and VSI’s “D-Stat Dry” are commonly used after cardiac catheterization procedures, which are performed to examine blood vessels in the heart. Applied to the skin surface like a bandage, the patches stem blood flow at the site of a physician’s puncture of a patient’s major artery in the upper thigh, which is made to start the catheterization. The patches are intended to operate more quickly and less invasively than older methods for halting post-catheterization bleeding.
VSI’s D-Stat Dry patch uses bovine thrombin — a compound derived from cow’s *58blood- — as its active ingredient. The ingredient creates the risk, which increases with multiple exposures of the patient to the ingredient, of triggering antibodies and therefore possible complications (which may include excessive bleeding or blood clots). An FDA-approved warning appears in D-Stat Dry’s written instructions for use. However, the ingredient seals wounds rapidly and, by the time of trial, VSI had sold more than one million D-Stat Dry patches, without reports of severe bleeding or blood clots resulting from proper use of the device.
VSI began selling D-Stat Dry in September 2003: it sold 19,000 devices during the remainder of 2003 and 164,000 in 2004. In early 2005, Howard Root, VSI’s chief executive officer (“CEO”), and VSI’s chief financial officer developed a three-year sales forecast, projecting that sales of VSI’s patch would grow from an estimated 282,000 in 2005 (one-third of the market) to 467,000 in 2007 (one-half of the market). VSI’s sales leveled off, however, and the company achieved about a 37 percent market share by 2007.
Several years earlier, MPT had won FDA approval for its own clotting bandage, now named the SyvekPatch, which uses a thin fiber produced by the diatom, a single-celled marine organism. MPT’s sales of the patch grew rapidly through 2002, but growth declined in subsequent years. MPT concluded that its sales had been affected by false information about its patch disseminated by VSI starting in 2003, and it counter-attacked by providing to its own sales agents negative information about VSI’s product.
After gentler critiques of D-Stat Dry failed to stop the decline in MPT’s sales, MPT’s marketing director, Peter Stevens, prepared an informational bulletin dated March 2, 2004 (the “marketing bulletin”), drawing on material in a scientific paper published in 2001 (the “Ortel article”). The bulletin, in an introductory large-type message, asserted, “YOU CAN’T USE D-STAT AND ‘DO NO HARM’ a reference to the Hippocratic Oath — and contained five statements later successfully challenged by VSI as product disparagement (the first two being the two sentences contained in the substantive paragraph that follows):
The message to communicate to all health care providers is:
After one application of bovine thrombin there is a 95% patient risk of developing antibodies to it, and a 51% risk of developing antibodies to human factor V. After just two exposures to topical thrombin there is a 36% risk of developing abnormal coagulation results leading to events such as excessive post-op bleeding, gastrointestinal bleeding, and hemorrhagic stroke.
The other statements challenged by VSI were the first three of five numbered points that appeared farther down on the first page of the bulletin:
1. Bovine thrombin [e.g. D-Stat] is extremely immunogenic; after just one exposure to Bovine thrombin 95% of patients raise antibodies to it.
2. Bovine thrombin carries a very high risk of causing serious permanent bleeding disorders in 51% of patients. Excessive post-op bleeding, gastrointestinal bleeding, and hemorrhagic stroke occur — after only 2 exposures to bovine thrombin.
3. After a second exposure to thrombin, 51% of patients develop antibodies to HUMAN Factor V and 36% have abnormal coagulation results — the effect is irreversible for life.
The bulletin was sent electronically and presented in training sessions to MPT’s sales force for its message to be communi*59cated to customers without giving them the bulletin itself. A VSI sales agent discovered the document while visiting a customer, and Root then wrote to MPT’s president asking for correction of false and misleading statements. Root asserted, for example, that the Ortel article rested on a study using an “extremely impure and since-discontinued” form of bovine thrombin rather than the purer D-Stat Dry version, and said that certain of the data concerned patients who had undergone “massively-invasive” surgeries rather than minimally invasive catheterization procedures.
MPT did not respond and VSI filed the present action in federal district court in Minnesota in May 2005; the case was later transferred to the District of Massachusetts. The only one of VSI’s claims to survive the trial was its product disparagement claim (we need not discuss the others or MPT’s counterclaims which also failed). Following a 10-day trial, the jury awarded VSI $4.5 million, a little more than half of the $8.4 million in damages VSI had sought. The jury made specific findings that these five statements were false and, in light of the jury instructions, had to find that MPT made them with actual malice. The district court enjoined MPT’s further use of the false statements.
On appeal, MPT challenges the jury’s verdict for VSI. MPT contends that VSI failed to prove malice by clear and convincing evidence and, alternatively, that VSI was not entitled to the judgment because the evidence did not show specific lost sales attributable to MPT’s conduct and so was insufficient evidence of special damages as a matter of law. The parties assume in their briefs that Massachusetts law governs VSI’s product disparagement claim; we therefore do likewise. See Nagle v. Acton-Boxborough, Reg’l Sch. Dist., 576 F.3d 1, 3 (1st Cir.2009).
Starting with malice, we note that neither the Supreme Court nor this one has decided whether the First Amendment requires in product disparagement actions the actual malice standard of New York Times Co. v. Sullivan, 376 U.S. 254, 279-80, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964); at least one circuit has rejected that proposition for a similar cause of action, see Procter & Gamble Co. v. Amway Corp., 242 F.3d 539, 552 & n. 26 (5th Cir.2001) (Lanham Act commercial disparagement), cert. denied, 534 U.S. 945, 122 S.Ct. 329, 151 L.Ed.2d 243 (2001). Whether Massachusetts courts might independently read such a requirement into its common law cause of action is also unclear.1
The parties tried the case before us on the premise that actual malice is required, and we accept that view solely for the purposes of this case. Bose Corp. v. Consumers Union of the U.S., Inc., 692 F.2d 189, 194 (1st Cir.1982) (accepting that the actual malice standard applied in a product disparagement case where the parties did not dispute it), aff'd, 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984); Suzuki Motor Corp. v. Consumers Union of U.S., Inc., 330 F.3d 1110, 1133-34 & n. 11 (9th Cir.) (same), cert. denied, 540 U.S. 983, 124 S.Ct. 468, 157 L.Ed.2d 373 (2003). Such a requirement might seem far afield from the concerns animating the New York Times case — itself an innovation not known to the common law — but this is an issue for another case.
*60To prove actual malice, VSI had to provide clear and convincing evidence that MPT’s marketing director made the challenged statements “with knowledge that [they were] false or with reckless disregard of whether [they were] false or not.” New York Times, 376 U.S. at 279-80, 287, 84 S.Ct. 710; Bose, 692 F.2d at 195. Thus Stevens must have had a “high degree of awareness of their probable falsity” to meet the actual malice standard, Garrison v. Louisiana, 379 U.S. 64, 74, 85 S.Ct. 209, 13 L.Ed.2d 125 (1964); for recklessness the evidence must permit the conclusion that he “in fact entertained serious doubts as to the truth of his publication,” St. Amant v. Thompson, 390 U.S. 727, 731, 88 S.Ct. 1323, 20 L.Ed.2d 262 (1968).
Subjecting the jury’s malice determination to our own independent review, as we must, see Bose, 466 U.S. at 510-11, 104 S.Ct. 1949, we find that the evidence supports the jury’s finding that the statements were false and provides ample proof of malice for at least the principal challenged statements; and, for reasons that will appear, this is an ample basis for satisfying the malice requirement, assuming it exists.
The most inflammatory of the five statements, and the most glaringly unsupported, are the two that associated D-Stat Dry with specific and serious outcomes in percentages that would be remarkable for a relatively straightforward medical task — to stop bleeding at a modest-size doctor-created incision. If believed, they would certainly deter use of a product for a result that could be achieved by other, long-used means or by newer products like that of MPT. Both statements associate YSI’s product “after just” or “only” two exposures with “excessive post-op bleeding, gastrointestinal bleeding, and hemorrhagic stroke.”
Phrasing the statistics differently, the two statements read:
After just two exposures to topical thrombin there is a 36% risk of developing abnormal coagulation results leading to events such as excessive post-op bleeding, gastrointestinal bleeding, and hemorrhagic stroke.
2. Bovine thrombin carries a very high risk of causing serious permanent bleeding disorders in 51% of patients. Excessive post-op bleeding, gastrointestinal bleeding, and hemorrhagic stroke occur — after only 2 exposures to bovine thrombin.
The Ortel article does refer to the increased risk of an “adverse event” upon second exposure to bovine thrombin, and it contains language and a table listing some of the possible adverse events, including “postoperative bleeding,” “gastrointestinal bleeding,” and “hemorrhagic complications.” The article, however, does not directly link a second exposure to bovine thrombin to the complications cited by MPT and certainly does not quantify any such risk in the percentages stated by MPT. If anything, the Ortel article disclaims proof of such an association.2
Language in the article that comes closest to the quoted statements was specific to exposure to bovine thrombin during surgery where much larger dosages are employed. Nor could MPT’s prediction of *61“permanent” bleeding disorders be supported by the article; as Stevens himself suggested, the time period of the study was too short to make that determination — in fact, the study followed the patients for only four to eight weeks. Stevens admitted that the Ortel article “doesn’t support even one person losing the ability to clot their blood permanently” or even one patient having a serious and permanent bleeding disorder.
In considering malice, the jury was also entitled to weigh MPT’s motive. Harte-Hanks Commc’ns, Inc. v. Connaughton, 491 U.S. 657, 668, 109 S.Ct. 2678, 105 L.Ed.2d 562 (1989); Bose, 692 F.2d at 196. MPT’s own sales began to fall after D-Stat Dry was introduced, and chronology shows that MPT developed the bulletin after MPT’s early and more temperate criticisms of D-Stat Dry seemingly went unheeded by customers. The scare phrasing of the document and the indication that it was to be used to persuade customers, but not to be left with them, could also be considered by the jury — the latter as reflecting a concern that the bulletin could otherwise be discovered by VSI and challenged as false.
The other three statements are a closer call as to malice. Two of them charge that bovine thrombin causes 95 percent of patients to raise antibodies to it after one exposure; two that it causes 51 percent of patients to raise antibodies to human factor V after at least one exposure; and one that a second exposure to bovine thrombin causes 36 percent of patients to have abnormal coagulation results — “the effect is irreversible for life.”3 The Ortel article uses these exact percentages with respect to the claimed risks of antibodies and abnormal coagulation results in speaking generically about bovine thrombin; and while at trial VSI offered expert medical data suggesting much lower antibody production in response to exposure to its form of bovine thrombin, Stevens could have been unaware of that other data and could have made these three challenged statements without actual malice.
But this possibility makes no difference to the damages verdict. The most inflammatory statements were recklessly false— or so the jury permissibly found — and it is a fair inference that the damages verdict rested primarily on them: their assertion of high percentage risks of lurid complications would have alarmed any doctor considering D-Stat Dry. As for injunctive relief directed at future repetition of the statements, MPT’s request for relief from the injunction rests on the premise that no violation was proved as to any of them and does not seriously attempt to show that any of the statements is true but only that they were all made without malice — a position the jury permissibly rejected.
The hard questions on this appeal concern damages. MPT’s main claim is that VSI failed to provide a particular form of proof of “special damages” — namely, lost sales to specific named customers as opposed to expert evidence and general data from which aggregate loss of sales might be inferred. In most fields of tort law, a plaintiff having established wrongdoing may prove damages in any reasonable way, e.g., Graves v. R.M. Packer Co., 45 Mass.App.Ct. 760, 702 N.E.2d 21, 28 (1998) *62(stating that the damages evidence need only “afford a ‘basis for a reasonable judgment’ ” (quoting Linkage Corp. v. Trs. of Boston Univ., 425 Mass. 1, 679 N.E.2d 191, 210 n. 38 (1997))); but for peculiar reasons, probably more historical than prudential, a number of jurisdictions following the Restatement (Second) of Torts require in product disparagement cases proof of specific lost sales to identifiable customers unless it is infeasible to provide such proof.4
Massachusetts has no authoritative case cited to us that is directly on point, but it has followed the Restatement on other aspects of product disparagement, e.g., Dulgarian, 652 N.E.2d at 609 (citing a different Restatement section on liability for injurious falsehood, which includes product disparagement), and we will assume arguendo that it would likely do so here. This is merely an assumption:
Under the impact of more detailed statistical and expert proof, courts have shown increasing willingness in recent years to award a loss of profit in other kinds of cases, and it may be expected that this will carry over into the injurious falsehood [which includes product disparagement] cases where the loss is shown with reasonable certainty.
Prosser and Keeton on Torts § 128, at 972 (5th ed.1984) (footnote omitted).
Furthermore, the Restatement proof requirement applies only where it is feasible to identify specific lost sales. But our case is not one of falsehoods about a product sold to a limited list of named regular purchasers. MPT had a number of representatives who dealt with hundreds of “cath labs” at hospitals. So one could postulate that it would hardly be easy to determine in most cases to whom MPT sales agents spoke,5 who or what group or committee made purchasing decisions, and the precise connection (if any) between the dissemination of falsehood and the absence of a set of purchases of VSI’s product.
Had MPT asserted this special proof rule in the district court, we would presumably have some developed information on the relevant facts and a ruling from the district judge as to whether the facts added up to the “wide dissemination” exception, under which even the Restatement does not require proof of individual lost sales. Restatement, supra, § 633 cmt. h; see also Sack on Defamation § 13.1.4.6 (3d ed.2007) (stating that generalized evidence of lost profits is “probably enough” to prove special damages, “particularly where it is impractical to establish specific lost sales”). But MPT made no such objection in the district court prior to or during trial, so the need for an instruction on *63proof of special damages was never established.
VSI has failed to point out that the special damages proof rule was not properly invoked in the district court. But because the parties tried the case below on the assumption that ordinary rules of damages proof applied, MPT’s objection is forfeit, subject to plain error, Davila v. Corporation de Puerto Rico Para La Difusion Publica, 498 F.3d 9, 14-15 (1st Cir.2007); the judicial system has an ample interest in not wasting resources with unnecessary retrials. See Harden v. United States, 688 F.2d 1025, 1032 n. 7 (5th Cir. Unit B 1982). Plain error is rarely found in civil cases in this circuit, Nat’l Ass’n of Chain Drug Stores v. New England Carpenters Health Benefits Fund, 582 F.3d 30, 43 (1st Cir.2009); and here, given the difficulty of proving specific lost sales, it is far from plain that any error occurred.
MPT spends much less time arguing that damages were not established even under ordinary standards, but this argument does require attention. For reasons already noted it might be impossible to identify most, let alone all, of the various individuals exposed to the false statements, or to identify many of those involved in purchasing the products, let alone to connect specific statements to specific failures to buy; but this does not mean that no such instances could be located — assuming that the falsehoods had effect. While VSI initially indicated to the district court that it would seek to present at least some specific lost sales, in the end it identified no such instance. No legal rule requires such exemplars but they could well have strengthened VSI’s case.
Instead, VSI’s central evidence was its three-year projection estimating expected growth of sales for its new and initially successful product. Admittedly, this was made before MPT’s false statements were disseminated and before litigation in this case commenced; and Root testified that the projections were well-founded on market research including third-party reports, public company reports, and available information on competitive developments, pricing, and VSI’s own sales experience. He also said that he knew of no reason that could explain the unrealized sales growth other than the false statements disseminated by MPT.
VSI also provided an expert damages witness, Donald Nicholson, a certified public accountant, who analyzed VSI’s estimate and concluded that based upon the projected sales figures, it had fallen short by $13.6 million in sales — $8.4 million in profits. It was this loss in profits that VSI attributed to MPT’s defamatory statements. Nicholson said that his calculations accounted for new entrants during the relevant period but not for other true information in the market place about bovine thrombin risks. MPT introduced no evidence of its own — expert or otherwise— to refute VSI’s damages showing.
Yet even as Nicholson stated that Root’s projection was reasonable, he reiterated that they were “really Mr. Root’s estimate[s]” and confessed that they would be too high if one were to accept a third-party report’s estimates of the market size. VSI also provided lower estimates of sales growth to its investors, although there are reasons why a company’s executives might provide more conservative estimates to shareholders. It is not surprising that the jury did not entirely accept the projection, awarding VSI only about half of what it sought for conduct that the jury itself deemed malicious.
Massachusetts cases have relied upon expert testimony by outside experts or company executives to uphold a jury’s determination that a defendant’s conduct *64caused a loss of profits.6 This is perhaps an easier inference where existing sales fall after the defendant’s wrongdoing; VSI’s damages require the further step of giving credence to Root’s projected growth in sales over and above the inference that the defendant’s conduct caused harm. Massachusetts courts have not foreclosed the use of growth estimates but are prepared to reject individual calculations as unsupported. See Augat, Inc. v. Aegis, Inc., 417 Mass. 484, 631 N.E.2d 995, 998 (1994) (rejecting a specific high growth rate, admittedly calculated after the litigation, as “implausible],” but not rejecting the method).
In this case, the evidence surely permitted the jury to infer that MPT’s malicious false statements — inflammatory and widely deployed — had a negative causal effect on VSI’s otherwise climbing sales and profits. As to the amount of damages, lost profits inherently involve estimation and courts are often inclined to think that the wrongdoer created the problem and must accept the jury’s judgment within a range of uncertainty. Herbert A. Sullivan, Inc., 788 N.E.2d at 543; Lowrie v. Castle, 225 Mass. 37, 113 N.E. 206, 210 (1916); Ricky Smith Pontiac, 440 N.E.2d at 48. The question is whether the amount in this case went beyond reasonable limits.
On this, the panel is divided even as between the two judges who believe liability has been established and vote to uphold an award of damages in favor of VSI. One member of the panel, recognizing that causation must be inferred and that any amount of damages is an estimate, thinks that the jury verdict is still within reasonable bounds; the other, that it is excessive — given the lack of proof of any specific lost sales and that its support is almost entirely a projection by the plaintiff as to what it hoped to earn — and that a sua sponte remittitur is essential.
The verdict most favorable to the judgment that two judges will support is an affirmance subject to a remittitur limiting VSI’s recovery to $2.7 million apart from interest, costs and other incidentals. See Alvira v. F.W. Woolworth Co., No. 92-2255, 991 F.2d 786, 1993 WL 101438, *2-4, 1993 U.S.App. LEXIS 7145, at *8-11 (1st Cir. Apr. 7, 1993); Catullo v. Metzner, 834 F.2d 1075, 1083 (1st Cir.1987); Kolb v. Goldring, Inc., 694 F.2d 869, 874-75 (1st Cir.1982). On remand, which we now order, VSI can accept this figure or insist upon a new trial as to damages. In all remaining respects the judgment is affirmed. Each side will bear its own costs.
It is so ordered.
. Massachusetts courts have not decided whether malice is an element of the tort, e.g., Dulgarian v. Stone, 420 Mass. 843, 652 N.E.2d 603, 609 & n. 9 (1995), but might require it if faced with the issue again, see id. at 609 (quoting the Restatement (Second) of Torts § 623A (1977), which provides the actual malice standard for injurious falsehood/product disparagement actions).
. The Ortel article states that "[ajdverse clinical complications in the postoperative period did not appear to be associated with the development of elevated antibody levels to bovine or human coagulation factors " (emphasis added). One of the experts at trial, Dr. Mann, explained that this passage "means ... that [the article's authors] could not identify an association between the generation of an antibody and ... any clinical event.”
. The claim that 36 percent of patients have abnormal coagulation results that are “irreversible for life” is significantly different, and better supported by the Ortel article in combination with Stevens’ probable understanding of immune memory, than the stronger statement that linked two bovine thrombin exposures to a substantial (impliedly 36% or higher) risk of particular bleeding disorders, which we have decided was unsupported by the Ortel article and made with actual malice.
. The phrase "special damages”' — common in defamation law — means nothing more than actual damages (as opposed to damages — e.g., to reputation — that are presumed for certain libelous statements), Peckham v. Holman, 28 Mass. (1 Pick.) 484, 486 (1831); the Restatement, followed in a number of jurisdictions, goes even further — in certain but not all cases — by demanding proof of specific lost sales. Restatement (Second) of Torts § 633 (1977) (special damages "may be established by (a) proof of the conduct of specific persons, or (b) proof that the loss has resulted from the conduct of a number of persons whom it is impossible to identify”); e.g., Fashion Boutique of Short Hills, Inc. v. Fendi USA, Inc., 75 F.Supp.2d 235, 240-41 (S.D.N.Y.1999) (rejecting suit because the plaintiff did not identify specific lost customers when it was possible to do so), aff'd, 314 F.3d 48 (2d Cir.2002).
. However, because MPT had only about 25 sales representatives in addition to a few executives who made sales calls, it should have been feasible to identify for at least some of those agents the customers and potential customers to whom they spoke during the relevant time period and so to follow the trail to at least some lost sales.
. See, e.g., Herbert A. Sullivan, Inc. v. Utica Mut. Ins. Co., 439 Mass. 387, 788 N.E.2d 522, 544 (2003); Brewster Wallcovering Co. v. Blue Mountain Wallcoverings, Inc., 68 Mass.App.Ct. 582, 864 N.E.2d 518, 529-30, 541-43 (2007); Ricky Smith Pontiac, Inc. v. Subaru of New England, Inc., 14 Mass.App.Ct. 396, 440 N.E.2d 29, 46-47 (1982).