OPINION of the court
Simons, J.Plaintiff Michael A. Opera fractured his right leg while skiing at Bluemont, a ski area owned and operated by Defendant Hyva, Inc. He and his wife sued defendant, claiming that the injury was caused by a defective binding on skis rented from it. Defendant thereafter impleaded third-party defendant Moog, Inc., the manufacturer of the binding. The jury awarded plaintiffs compensatory and derivative damages on their strict products liability cause of action and it apportioned liability 25% against defendant Hyva, Inc., and 75% against third-party defendant Moog, Inc.
The accident happened on February 12, 1977 when plaintiff, who had only been on skis twice before, was skiing downhill on the beginners’ slope. While doing so, he fell forward and slightly to the right. His left ski binding released but the right one did not and he sustained angulated fractures of the distal third of his tibia and fibula, an *375injury known as a “boot top” fracture because of its location. It was plaintiff’s claim that the bindings should have released and that defendant’s employees had wrongfully adjusted them too tightly at the time of rental, using a setting which required the application of too much force before releasing. He claims that defendant relied on a defective instruction manual supplied by the manufacturer which recommended binding settings which were “not reasonably safe”. Plaintiff’s evidence on the point included instruction manuals published by Moog after the accident and which contained modifications which plaintiff claimed should have been used in prior manuals.
There should be a reversal. The evidence of postaccident changes in Moog’s manual was admissible to prove the feasibility of using a different system of adjusting the bindings, but the court erred in instructing the jury that these postaccident modifications were evidence that the binding was not reasonably safe when marketed (see Rainbow v Elia Bldg. Co., 79 AD2d 287, affd 56 NY2d 550; Bolm v Triumph Corp., 71 AD2d 429, 436-437; cf. Caprara v Chrysler Corp., 52 NY2d 114).
To state the problem in its simplest terms, bindings for downhill skis should keep the skis fastened on the skier’s foot when he needs them but release so that ski and skier will separate before excessive force, encountered because of a fall or unexpected terrain, is applied to the skier’s body. Excessive force is force that will fracture the skier’s leg bones. And that is the variable factor in adjusting ski bindings; how much force — measured in foot-pounds to account for the leverage of the ski — will the skier’s bones withstand. Obviously, it differs for skiers of different sizes and skiing styles. All binding manufacturers address this problem in the same way. They use various adjustment tables which state the safe force (as discovered by empirical data from laboratory tests) that individual skiers can tolerate. By extrapolation, these readings are transferred from the adjustment table to the manufacturer’s conversion chart which relates it to the proper setting to be used for the binding.
At the time of this accident there were three adjustment tables available for measuring the safe force which could *376be tolerated by individual skiers. Moog’s manual for the 1976-1977 season used two of them. The first was the I.A.S. system which required measuring the diameter of the skier’s tibia. This was the most precise test of bone strength when done in the laboratory on a denuded bone, but its use presented some obvious difficulties when measurement was made on a human leg in a commercial setting. The second table was called the BfU system which calibrated the release force by considering the skier’s sex, weight and ability (measured by whether the skier was an average or fast skier). The BfU Tables were used by defendant to determine that a D setting should be used on the binding rented to plaintiff.
The third set of adjustment tables, relied on by plaintiff, were known as the Universal Binding Adjustment Tables. They used a “weight and ability” or Lipe* method of measuring the foot-pounds that a given skier’s legs could safely withstand by taking into consideration the skier’s weight and ability (as measured by whether the skier was a beginner, advanced novice, intermediate, advanced intermediate or expert). Using the Lipe method, plaintiff’s binding should have been adjusted to release more easily by adjusting it to an E or F setting. As noted, Moog’s 1976-1977 instruction manual did not contain the Lipe Tables, but in the manuals published for the 1977-1978 and 1978-1979 ski seasons, the Lipe Weight and Ability Tables were substituted for the BfU Tables. These post accident manuals were received, in evidence.
It is now familiar law that a manufacturer is liable for producing a defective product which causes injury in a cause of action for strict products liability. The product may be defective because of the way it was made (see, e.g., Caprara v Chrysler Corp., 52 NY2d 114, supra; Codling v Paglia, 32 NY2d 330), the way it was designed (see, e.g., Robinson v Reed-Prentice Div. of Package Mach. Co., 49 NY2d 471; Micallef v Miehle Co., Div. of Miehle-Goss Dexter, 39 NY2d 376; Bolm v Triumph Corp., 33 NY2d 151; Rainbow v Elia Bldg. Co., 79 AD2d 287, supra), or because the manufacturer failed to give adequate warning or instructions on its use (see, e.g., Torrogrossa v Towmotor Co., *37744 NY2d 709; Wolfgruber v Upjohn Co., 72 AD2d 59, affd 52 NY2d 768; Lancaster Silo & Block Co. v Northern Propane Gas Co., 75 AD2d 55). In the first of these, a defect in construction through mistake in manufacturing or assembly, the plaintiff proves that the product which caused his injury was defective in the sense that it did not perform as intended. Fault plays no part in the liability equation; a defectively manufactured product is one that is flawed because, without regard to fault, it was misconstructed. A defectively designed product, on the other hand, is one that presents an unreasonable risk of harm, notwithstanding that it was meticulously made according to plans and specifications (see, e.g., Robinson v Reed-Prentice Div. of Package Mach. Co., 49 NY2d 471, 478-479, supra; Rainbow v Elia Bldg. Co., 79 AD2d 287, 294, supra). The decision that a product was unreasonably dangerous because of its design involves a balancing process to assess whether the benefits of a particular design outweighed the risks of using it and also what alternative designs were available and feasible at the time the product was marketed, all subjective judgments. Defective design cases are thus similar to negligence cases and the standards for imposing liability for design defects are general negligence principles (Bolm v Triumph Corp., 33 NY2d 151, 157-158, supra; Rainbow v Elia Bldg. Co., supra; see, generally, Hoenig, Resolution of “Crashworthiness” Design Claims, 55 St. John’s L Rev 633, 659 et seq.). So, too, in the case of failure to give adequate warning or instruction, the test is a subjective one of what is reasonable and feasible under the circumstances (Torrogrossa v Towmotor Co., supra). Where the theory of liability is failure to warn or adequately instruct, negligence and strict products liability are equivalent causes of action (Wolfgruber v Upjohn Co., 72 AD2d 59, 62, supra; Rainbow v Elia Bldg. Co., 49 AD2d 250, 253; Lancaster Silo & Block Co. v Northern Propane Gas Co., 75 AD2d 55, 63-64, supra). Since negligence principles apply, evidence of postaccident modifications are not admissible in design defect or failure to warn cases to establish fault, and the court’s erroneous instruction requires a new trial (see Rainbow v Elia Bldg. Co., supra; cf. Caprara v Chrysler Corp., supra).
*378Plaintiff’s counsel contends that the evidence was properly received without objection or limitation, but the record does not support that claim. He attempted to use the postaccident manuals to show far more than feasibility. Over objection he was permitted to use them to corroborate plaintiff’s expert’s claim that F was the proper setting for plaintiff’s binding, to cross-examine other witnesses on liability and to argue that the change in the manual after the accident was proof of fault (see Bolm v Triumph Corp., 71 AD2d 429, 436, supra; and see, generally, Weinberger, Caprara over the Rainbow — New York Grapples with Post-Accident Design Changes in Products Liability Actions, 46 Albany L Rev 132, 144-145; Madden, Admissibility of Post-Incident Remedial Measures: A Pattern Emerges, 5 Journal of Products Liability 1).
One other evidentiary point requires comment. Moog attempted to introduce a chart showing that none of its competitors used the Lipe method in their instructions in 1976. The court correctly ruled that the evidence was not relevant. A manufacturer may defend a case claiming improper instructions by showing that a suggested alternative was not within the state of the art at the time of manufacture, but the custom in the industry does not establish the state of the art (Bolm v Triumph Corp., 71 AD2d 429, 438, n 2, supra).
Defendant’s third-party complaint asserted separate causes of action in strict products liability, breach of warranty and negligence and Moog raises points with respect to each of them.
First, it contends that the court erred in charging strict products liability principles similarly in the main action by plaintiff as lessee of the equipment against defendant as lessor and also in the third-party action between defendant as purchaser and third-party defendant as manufacturer. Essentially, its point is that it should not be strictly liable because defendant failed in its duty to it, third-party defendant, to inspect and use the product as intended when renting it to others (see Codling v Paglia, 32 NY2d 330, 343, supra). Since defendant was in the business of renting equipment, it was liable to plaintiff in strict liability if it rented him defective equipment (see *379Stang v Hertz Corp., 83 NM 730, 733-734; and cf. Nastasi v Hochman, 58 AD2d 564). As for the liability between the parties in the claim over, their relationship also would support a cause of action in strict liability between defendant as purchaser and third-party defendant as manufacturer. It was for the jury to decide, in apportioning fault, which of the two, lessor or manufacturer, was ultimately liable or if jointly liable, what portion of the loss they should bear inter sese.
Next, Moog contends that the warranty cause of action was erroneously submitted to the jury. We agree with the trial court, however, that whether Moog’s advertising statements were warranties and a part of the bargain of sale to defendant or mere puffing was a question of fact (see Uniform Commercial Code, § 2-313, subd [1], pars [a], [b]; Spiegel v Saks 34th St., 43 Misc 2d 1065, 1070-1071, affd on opn below 26 AD2d 660).
Finally, Moog contends that the court should have dismissed defendant’s negligence cause of action against it for failure of proof. On all the evidence in the case, however, the jury could have found that Moog in preparing its instruction manuals was negligent in using the IAS Table, which was impractical, or the BfU Table, which resulted in too strong a binding setting.
We have considered third-party defendant’s other points and found them without merit.
Insofar as defendant’s appeal is concerned, plaintiff contends that it is untimely because third-party defendant appealed 29 days after entry and service of judgment and defendant did not appeal until 10 days later. We consider defendant’s appeal timely even though more than 30 days had elapsed before it appealed plaintiff’s judgment (see CPLR 1008; Matter of Gebbie, 33 AD2d 1093, 1094). We would reverse the judgment against it in the interest of justice, in any event, since we are reversing its judgment against the third-party defendant (see Rome Cable Corp. v Tanney, 21 AD2d 342, 345; McLaughlin, Practice Commentaries, McKinney’s Cons Laws of NY, Book 7B, CPLR C1008:1, pp 119, 120; Siegel, Supplementary Practice Commentaries, McKinney’s Cons Laws of NY, Book 7B, CPLR C5522:2, 1981-1982 Supp, pp 29-30).
*380The judgment should be reversed and a new trial granted.
Dillon, P. J., Moule and Schnepp, JJ., concur with Simons, J.; Doerr, J., not participating.
Judgment reversed, on the law and facts, without costs, and new trial granted.
The developer of the system was an American named Gordon Lipe.