dissenting.
I do not think the district judge was “clearly wrong” in directing a verdict for Volkswagen on the negligence claim at the close of all the evidence. The issue the judge had to decide was whether there was “any credible evidence which under a reasonable view would support a verdict contrary to” a verdict for Volkswagen. Thompson v. Howe, 77 Wis.2d 441, 448, 253 N.W.2d 59, 62 (1977). Chaulk presented no credible evidence of negligence.
She was a passenger in the front seat of a 1977 Volkswagen Rabbit and was thrown out of the car when it was “broadsided.” Whether she went through the door or the window is unclear. It is possible, at least in theory (for there is no evidence that such an accident has ever occurred), that the crash pushed in the door handle, causing the latch to be released and the door to open. Chaulk was allowed to present to-the jury her theory that an automobile door latch which releases when pushed inward is defective, but the only evidence of this was the testimony of Martens, her expert witness, and the jury brought in a verdict for Volkswagen. The judge, believing that Martens’ testimony would not allow a rational factfinder to find that Volkswagen had been negligent, had refused to allow the jury to consider the case under Chaulk’s alternative theory, negligence, as well as strict liability. It is that refusal of which she complains on this appeal.
There is little if any practical difference between negligence and strict liability in a products liability case complaining about the design of the product. To win on a theory of strict liability the plaintiff must prove that the product was defective in its design, which means, unreasonably dangerous. Sumnicht v. Toyota Motor Sales, U.S.A., Inc., 121 Wis.2d 338, 367-70, 360 N.W.2d 2, 15-17 (1984). To win on a theory of negligence the plaintiff must prove that the defendant, “in the exercise of ordinary care, should have foreseen that his design and method of installation would be unreasonably dangerous to others.” Greiten v. LaDow, 70 Wis.2d 589, 602, 235 N.W.2d 677, 685 (1975). (This quotation is from the majority opinion, mislabeled concurrence, as explained in Howes v. Deere & Co., 71 Wis.2d 268, 274, 238 N.W.2d 76, 80 (1976).) Unreasonable danger is the key under either theory; if there are any substantive differences between the two theories they are not apparent from the facts of this case. One assumes, however, that strict liability would be easier to prove than negligence, as implied by the statement in Greiten that the plaintiff’s task in proving negligence in a products case is “onerous.” 70 Wis.2d at 602, 235 N.W.2d at 685. So the jury that brought in a verdict for Volkswagen on the strict liability count would not have brought in a verdict for Chaulk if her negligence count had been allowed to go to the jury also, and while this consideration is not conclusive on the question whether the judge should be reversed for withdrawing the issue of negligence from the jury, it is pertinent. For the question for us is not whether he was *644wrong but (as the majority acknowledges) whether he was clearly wrong. An error that is not prejudicial is still an error, but it is unlikely to be a clear error. Compare the treatment of “plain error” in the criminal law, as in United States v. Silverstein, 732 F.2d 1338, 1349 (7th Cir.1984).
But I do not think there was any error here, except an error of which no one complains: allowing the strict liability claim to go to the jury. No reasonable jury could have believed Martens’ testimony that the design of the Rabbit’s door latch was defective in a sense relevant to products liability law. Judge Reynolds was extending charity to Chaulk in allowing the jury to consider her strict liability claim. Of course it does not follow that if the judge erred in allowing the jury to consider that claim he erred in withholding the very similar negligence claim from the jury.
Let us consider the evidence on whether Volkswagen should have foreseen that the design of the door latch was unreasonably dangerous. The design is of a type used on millions of automobiles, including very expensive ones such as Cadillac. The design met or exceeded federal automobile safety standards. There is no evidence that this design has ever caused a door to open in a crash test conducted on the Rabbit, or for that matter on any other car. Indeed, with the disputed exception of the present accident (for Volkswagen presented evidence that the door latch had nothing to do with Chaulk’s being flung from the car — that she had been hurled through the window), there is no evidence of any accident — ever—in which a door opened because a broadside crash pushed in the door handle and released the latch.
All Chaulk has is the testimony of Martens, whose involvement in the design of automobile door latches had ended in 1972, thirteen years before the trial. He had gone to work that year for an insurance company, and later he became and he remains a professional expert witness against automobile companies in cases involving issues of door-latch design. On direct examination he testified that other automobile companies had safer latch designs which Volkswagen easily could have adopted and in subsequent model years did adopt. It would have been a jury question whether this evidence outweighed Volkswagen’s contrary evidence. But on cross-examination Martens made clear that his position, whether reflecting a genuine and disinterested conviction or pecuniary self-interest, is that almost all door latches, including one for which he holds a patent and Volkswagen’s subsequent latch design, are unreasonably dangerous: specifically, all door latches that can be sprung by inward rather than downward pressure; perhaps all automobile door latches, period, except that of Mercedes-Benz; at the very least, the door latches found in 30 million cars now on American roads. Martens was unwilling to concede that any automobile door latch except that of the Mercedes-Benz, including latches he himself had designed, is reasonably safe. Although he believes that some are safer than the one on the 1977 Rabbit, he presented no statistical or other evidence indicating that an accident (with the possible exception of the one in this case) had ever occurred that a safer door latch would have prevented.
Martens’ was the testimony either of a crank or, what is more likely, of a man who is making a career out of testifying for plaintiffs in automobile accident cases in which a door may have opened; at the time of trial he was involved in 10 such cases. His testimony illustrates the age-old problem of expert witnesses who are “often the mere paid advocates or partisans of those who employ and pay them, as much so as the attorneys who conduct the suit. There is hardly anything, not palpably absurd on its face, that cannot now be proved by some so-called ‘experts.’ ” Keegan v. Minneapolis & St. Louis R.R., 76 Minn. 90, 95, 78 N.W. 965, 966 (1899).
It is not the law, in Wisconsin or anywhere else, that the standard of care is set by the designers of $60,000 automobiles, so that the omission of any safety device found in such automobiles is negligent. “It is boilerplate law that, merely because *645a product or an operation is not as safe as possible, because there are better methods of manufacture or performing an operation does not lead to the conclusion that the method employed was undertaken with a lack of ordinary care or the product was defective.” Greiten v. LaDow, supra, 70 Wis.2d at 602, 235 N.W.2d at 685. The buyer of a Mercedes 560 may be willing to pay extra for minuscule, perhaps wholly theoretical, improvements in safety, but such a buyer’s willingness to buy the ultimate refinement in safety technology does not define the standard of care for the whole industry.
Ours is not a system of people’s justice, where six laymen are allowed to condemn an entire industry on the basis of absurd testimony by a professional witness. If Martens had testified that the Rabbit in which Chaulk was riding should have been equipped with radar or a force field or an ejection seat, my brethren would agree that the directed verdict was proper. Judge Reynolds was entitled to conclude that Martens’ actual testimony, which amounts to a wholesale condemnation of the automobile industry for failing to adopt safety precautions that so far as the evidence shows would not have prevented a single accident in the history of transportation, likewise did not provide a rational basis for finding Volkswagen negligent. Compliance with industry custom is not a defense in a tort suit; but something more than a professional witness’s conclusion, offered without substantiation, that a whole industry is lagging behind the standard of due care is necessary to create a jury issue.
In Greiten v. LaDow the Supreme Court of Wisconsin upheld a directed verdict for the defendant in a products liability case in which, as in the present case, the theory of liability was negligence. Again the focus was on the plaintiff’s expert and among the factors which caused the trial judge to direct a verdict for the defendant was that “the theoretical alternatives as to design suggested by the expert witness for the plaintiff ... did not ‘... appeal to reason or good common sense,’ ” and that “there had been no prior accidents — ” 70 Wis.2d at 597, 235 N.W.2d at 682 (concurring opinion — with whose evaluation of the evidence the majority agreed). Greiten shows that Judge Reynolds did not step out of bounds in directing a verdict for Volkswagen. See also Chart v. General Motors Corp., 80 Wis.2d 91, 110-12, 258 N.W.2d 680, 688-89 (1977).
One more point. The distinction my brethren make between “direct force” and “stretching” in an effort to make Martens’ testimony seem less wild-eyed does not appear in Chaulk’s briefs, was not mentioned at argument, and therefore has not been subjected to the fires of the adversary process. See Bonds v. Coca-Cola Co., 806 F.2d 1324 at 1329 (7th Cir.1986); Tom v. Heckler, 779 F.2d 1250, 1259 (7th Cir.1985) (dissenting opinion). It may provide useful guidance to Chaulk and Martens in a new trial; but it is not our proper function as federal appellate judges, and it is not fair to trial judges or defendants, to assist plaintiffs’ attorneys in products liability cases by scouring the record for evidence they overlooked. I would affirm.