OPINION
CONNOR, Justice.On June 24, 1973, Deraid Allen Beck was killed when the Caterpillar 944 front-end loader which he was operating rolled over an embankment. Decedent’s widow, Paula Beck, brought this action against Caterpillar Tractor Company (hereinafter Caterpillar) for the wrongful death of her husband. Beck contended that Caterpillar’s failure to equip the 944 loader with a roll-over protective shield (hereinafter ROPS) constituted a design defect in the loader, and because of that defect, Deraid Beck suffered fatal injuries. Following a jury verdict in favor of Beck, judgment was entered against Caterpillar in the amount of $408,594.50. Caterpillar appeals from this judgment, specifying several errors by the trial court. Beck cross-appeals, alleging that it was error for the trial court to instruct the jury on Der-ald Beck’s comparative negligence.
I. FACTS
The model 944 front-end loader was designed by Caterpillar in the late 1950’s and was first sold in 1964. It was the first model of Caterpillar’s wheel loaders to be produced, and subsequent models have now replaced it. The particular loader involved in the accident was originally delivered by Caterpillar in 1964 to a dealer, Northern Commercial Company, as a standard model machine without a cab.
At the time of the accident, Deraid Beck was a partner in R. W. Beck & Sons, d/b/a Nenana Excavators. Nenana Excavators bought the 944 loader as used equipment from the Northern Commercial Company in 1967. The ten ton loader was equipped with a fiberglass all-weather shell, without supporting members, which was attached by bolts to the body of the machine.1
Although there were no witnesses to the accident, an investigating state trooper testified that Deraid Beck was apparently “roading” the loader in reverse,2 with the bucket in a raised position. The road was dirt and gravel with soft shoulders and wound around the side of a hill. On one side of the road there was an embankment of approximately seven feet. The trooper estimated that Beck was traveling at ap*875proximately 10 to 15 miles per hour as he proceeded up the road, but the loader then sank into the road shoulder and came almost to a complete stop. The loader apparently rolled over on its side and then flipped over 180 degrees and fell down the embankment.3 Beck was crushed beneath the loader, pinned between the steering wheel and part of the seat. The fiberglass canopy was flattened out and parts of it were broken off.
Paula Beck, decedent’s wife and personal representative, filed suit against Nenana Excavators for negligence in failing to provide safe equipment and against Caterpillar in strict liability. She alleged that Caterpillar’s failure to manufacture the loader with a ROPS rendered the loader defective. Caterpillar denied that the loader was defectively designed because roll bar protective devices were not readily available at the time of design and because the loader was intended to be marketed as a basic structure to which the user, through his dealer, would add auxiliary parts as necessary for his particular use. Nenana Excavators was subsequently dismissed from the suit on its motion for a directed verdict because Deraid Beck was a partner in the company and was, therefore, precluded from suing himself.
A ROPS is an overhead protective canopy which is constructed to withstand a rollover and, thus, protect the operator from being crushed. In addition, a ROPS can decrease the risk of a roll-over. There was no dispute at trial that a ROPS, as developed at the time of the accident, would have saved Deraid Beck’s life. Beck would not have been crushed and a ROPS may have even prevented the loader from overturning. There was also no dispute that it is best from a cost and technological standpoint to have a ROPS installed by the loader manufacturer at the time of initial production. However, there was considerable dispute about the availability of a ROPS at the time Caterpillar produced this loader in 1964.
Various types of protective canopies were used in the heavy equipment industry for many years prior to the development of a ROPS, particularly in the logging industry. These canopies provided protection primarily from inclement weather and falling objects. Such devices were known in the industry by the name FOPS, falling object protective shield or structure. These shields were manufactured by several auxiliary equipment manufacturers at the time the 944 loader was produced. A FOPS did not have the capability to withstand a rollover.
A consulting safety engineer, Ovid Holmes, appearing as an expert for Beck, testified that the heavy equipment industry began testing ROPS devices as early as 1961. Holmes had not heard of a FOPS, but testified that he believed that some of the canopies developed in the 1950’s did have roll-over protection capability. Although other front-end loader manufacturers did not supply a ROPS as part of the standard equipment at that time, several auxiliary equipment manufacturers did offer them.
According to Holmes, two Caterpillar dealers in California were selling an auxiliary ROPS by 1961 and the Illinois Division of Highways, in the home state of Caterpillar, had installed such structures on some of its equipment in 1962. Holmes believed that there was nothing technologically difficult about manufacturing a ROPS in 1964, but that Caterpillar did not do so because it would have been too costly. An employee of an auxiliary equipment manufacturer also testified that pre-1965 canopies were available with roll-over protection capability. Further, he stated that some of the canopies available in 1956-1960 for overhead protection were mounted to the frame of the basic machine in such a manner that the risk of roll-over was minimized.
*876On the other hand, Caterpillar stated in its answer to an interrogatory of Beck, that “[p]rior to 1966, Caterpillar Tractor was not aware and is not aware of the availability of [ROPS] from anyone in the country.” Although one Caterpillar senior staff engineer confirmed the Illinois Division of Highway utilization of ROPS, a manager of sales development, Jack Hasten, testified that he was unsure whether the Illinois canopy actually was a ROPS. Hasten also testified that the structural integrity of the type of canopies available in 1964 was such that it would have been irresponsible for Caterpillar to install them and represent them as having roll-over protection capability. In addition, he stated that there had been no need expressed for a ROPS. He claimed that users disliked attached canopies because the machines had a myriad of uses, some of which did not require a ROPS, and that a ROPS could be dangerous because it obstructed the operator’s view and impeded his ability to jump free if the loader fell into water. Other Caterpillar staff engineers testified that there was no ROPS capability available at the design stages of the 944 loader.
Caterpillar, at the time of design and production of the 944 loader, made a “deliberate decision” not to install any kind of protective canopy whatsoever on the 944 as part of its basic design. It did begin testing the ROPS concept in 1966. By that time, several auxiliary equipment manufacturers were regularly producing ROPS. In response to various state and federal regulations promulgated shortly thereafter,4 Caterpillar, in 1969, began installing ROPS as part of the basic vehicle model. No ROPS was ever made part of the 944 model manufactured by Caterpillar and the 944 loader went out of production in 1968 before Caterpillar began adding ROPS to its loaders.
By special verdict, the jury found that the absence of a ROPS was a design defect in the 944 loader and that such defect proximately caused the death of Deraid Beck. Damages were set at $817,189.00. However, the jury also found that Deraid Beck knew of the design defect and voluntarily and unreasonably proceeded to use the defective product. Since the jury deemed Beck to be 50% at fault for the accident by his knowing confrontation of the risk of harm, the total amount of damages awarded was reduced by half.
Appellant Caterpillar contends in this appeal that the trial court erred by (1) improperly instructing the jury on the law of strict liability for alleged design defects and refusing to define “defect” for the jury; (2) improperly instructing the jury on the law of comparative negligence; and (3) refusing certain evidence offered by Caterpillar during the course of the trial to show Caterpillar’s marketing methods and decedent Beck’s knowledge of the availability of a ROPS from another source. Each of these issues raise important questions in the area of products liability. As we explain below, we agree with Caterpillar’s objection to some of the challenged instructions and evi-dentiary rulings and conclude that the judgment must be reversed.5
II. DESIGN DEFECT
Caterpillar contends that the trial court erred in instructing the jury that “[a] design defect is one in which the product, however perfectly manufactured, incorporates or fails to incorporate a design feature with the result that injury is proximately caused thereby.”6 Caterpillar argues that *877the instruction misconstrues the law of strict liability in that it essentially instructs the jury on absolute liability: it does not instruct the jury that they must first find a defect in the product nor does it give the jury guidance in determining whether a defect exists. In addition, Caterpillar urges that we define the term “defective” for cases involving design defect, preferably by use of a standard which balances the risk presented by the product in light of the product’s utility. Beck, in response, argues that the jury was properly instructed on the necessity of finding a defect before it could fix liability and that the term “defect” need not be defined beyond its inherent meaning.7 Because this issue poses so substantial a question in a nascent body of law, we believe it would be helpful, before turning to the particulars of this instruction, to briefly summarize the development of products liability.
Strict products liability evolved from two separate bases: implied warranty and strict liability in tort.8 In the process whereby the courts have combined these theoretical bases and have fashioned from them a new species of product liability, it has become apparent that several policy considerations have been at work.9 First, greater efficiency and justice would be achieved by eliminating the necessity for plaintiffs to prove fault by manufacturers of defective products. Section, frustration of consumer expectation would still be the underlying basis of recovery. Third, incidence of harm from unsafe products would be reduced by imposing strict liability and thus providing an incentive to produce safer products. Fourth, risks would be allocated and losses spread by requiring a manufacturing enterprise to accept responsibility in strict liability for harm attributable to a defect in its product. The concept of risk allocation has been the primary policy rationale convincing courts to adopt strict products liability.10
We adopted strict products liability in Clary v. Fifth Avenue Chrysler Center, Inc., 454 P.2d 244 (Alaska 1969), holding that:
“A manufacturer is strictly liable in tort when an article he places on the market, knowing that it is to be used without inspection for defects, proves to have a *878defect that causes injury to a human being.”11
Following the doctrinal approach taken by the California Supreme Court in Greenman v. Yuba Power Products, Inc., 59 Cal.2d 57, 27 Cal.Rptr. 697, 377 P.2d 897 (1962), California’s seminal products liability decision, we stated that:
“The purpose of imposing such strict liability on the manufacturer and retailer is to insure that the cost of injuries resulting from defective products are borne by the manufacturers that put such products on the market rather than by the injured persons who are powerless to protect themselves.”12
We reaffirmed this policy of risk allocation in subsequent products liability decisions. See Cloud v. Kit Manufacturing Co., 563 P.2d 248, 250 (Alaska 1977); Bachner v. Pearson, 479 P.2d 319, 327 (Alaska 1970).
Later, in Butaud v. Suburban Marine & Sporting Goods, Inc., 543 P.2d 209 (Alaska 1975) (hereinafter Butaud I,) we adopted the holding of the California Supreme Court in Cronin v. J. B. E. Olson Corp., 8 Cal.3d 121, 104 Cal.Rptr. 433, 501 P.2d 1153 (1972). Cronin, addressing the plaintiff’s burden of proof in showing a defect, rejected the standard proposed by the Restatement (Second) of Torts13 that the plaintiff demonstrate that the product’s defect rendered it “unreasonably dangerous.” Rather, it held that a plaintiff satisfies his burden of proof when he proves the existence of a “defect” and that such defect was a proximate cause of his injuries.14 Cronin, 104 Cal.Rptr. at 438-442, 501 P.2d at 1158-62. This standard applies to cases wherein the alleged defect is one of design as well as one of manufacture.15 Cloud v. Kit Manufacturing Co., 563 P.2d 248 (Alaska 1977); Hauter v. Zogarts, 14 Cal.3d 104, 120 Cal.Rptr. 681, 534 P.2d 377 (1975); Henderson *879v. Harnischfeger Corp., 12 Cal.3d 663, 117 Cal.Rptr. 1, 527 P.2d 353 (1974).
We turn now to Caterpillar’s first objection to the trial court’s instructions to the jury; i. e., that the trial court mistakenly instructed the jury in absolute liability.
Although courts and commentators have struggled with diverse approaches to strict products liability,16 most authorities appear to agree that manufacturers are not absolute insurers of their products.17 Strict liability will not impose legal responsibility simply because a product causes harm.18 A product must be defective as marketed if liability is to attach, and “defective” must mean something more than a condition causing physical injury.
The instruction at issue stated that plaintiff Beck alleged that a design defect caused the death of Deraid Beck and then proceeded to state that a product is defectively designed when an injury results from a design feature (or the lack of a design feature). Essentially, the jury was instructed that there must be proof of a defect in the product’s design in order for liability to attach, and that proof of an injury could suffice to prove a defect. If such defect proximately caused the injury, strict liability is imposed. This is a tautology and tantamount to an instruction of absolute liability.
Beck argues that the instruction was proper because it “simply advises the jury that a design defect is a defect resulting from a design feature.” However, it is not just any design feature yielding an injury which will result in the imposition of legal responsibility for the incidents of such injury. A product is not necessarily defective merely because an injury occurred. Baker v. Chrysler Corp., 55 Cal.App.3d 710, 127 Cal.Rptr. 745 (1976). As Chief Justice Traynor of the California Supreme Court explained:
“The reasons justifying strict liability emphasize that there is something wrong, if not in the manufacturer’s manner of production, at least in his product . A bottling company is liable for the injury caused by a decomposing mouse found in its bottle. It is not liable for whatever harm results to the consumer’s teeth from the sugar in its beverage. A knife manufacturer is not liable when the user cuts himself with one of its knives. When the injury is in no way attributable to a defect there is no basis for strict liability.”19
Likewise, it is not just the absence of any possible safety device which will render a product legally defective.
Beck contends that, even if the subject instruction were erroneous, the error was cured by other instructions which clearly required the jury to find that a defect existed. Other instructions did direct the jury that they must find that Beck had proven a defect in order to return a verdict in her favor. However, the only guidance the jury was given in how they should determine if the absence of a ROPS was a design defect was the above-quoted instruction. Even though the instructions as a whole required the jury to find a defect, the circular definition of a design defect was erroneous.
*880Caterpillar also contends that the jury was inadequately instructed on the meaning of design defect. Since a manufacturer’s strict liability depends upon the meaning of defective, Caterpillar urges us to develop particular standards for assessing a product’s defectiveness.20
Design defects present the most perplexing problems in the field of strict products liability because there is no readily ascertainable external measure of defectiveness. While manufacturing flaws can be evaluated against the intended design of the product, no such objective standard exists in the design defect context. Beck argues that, despite the lack of an external standard in design defect cases, the word “defect” has an inherent meaning; i. e., a flaw or incompleteness. Therefore, when jurors are instructed that they must find a defect to impose liability, they realize that the product must be faulty. However, we believe that the dictionary meaning of the word “defect” is insufficient to fully explain to the jury when a manufacturer should incur legal responsibility for his product’s design.
“[The] natural application [of the word ‘defective’] would be limited to the situation in which something went wrong in the manufacturing process, so that the article was defective in the sense that the manufacturer had not intended it to be in that condition. To apply it also to the case in which . . . the design turns out to be a bad one or the product is likely to be injurious in its normal condition . . . [and] [t]o use it without defining it to the jury is almost to ensure that they will be misled.” (footnotes omitted)
Wade, On the Nature of Strict Tort Liability for Products, 44 Miss.L.J. 825, 831-32 (1973). Since we are unwilling to view any product which causes an injury while being used properly as defectively designed, we must formulate a test which will guide the jury in design defect cases.21
In the strict liability context, the term “defect” is “neither self-defining nor susceptible to a single definition applicable in all contexts.” Barker v. Lull Engineering Co., Inc., 20 Cal.3d 413, 143 Cal.Rptr. 225, 235, 573 P.2d 443, 453 (1978). The varied purposes and the theoretical underpinnings of implied warranty and strict liability in tort have generated five main tests for product defectiveness: (1) the “deviation from the norm” test;22 (2) the “reasonable fitness for intended purpose” test;23 (3) the Restatement test; (4) the “risk/utility analysis” test;24 and (5) the recent test proposed by the California Supreme Court in Barker v. Lull Engineering Co., Inc., supra.
*881Before determining which test for defect is the most appropriate, we must address Beck’s claim that to define defect is inconsistent with our position in Clary25 and Butaud I.26 In many products liability cases, the approach taken in Clary will suffice for advising the jury when liability can be properly fixed. However, those decisions were neither all-encompassing nor intended to be read as precluding any further refinement of the defectiveness standard. We are also not restricted by the holding of Cronin,27 establishing the plaintiff’s burden of proof, which we adopted in Butaud I.
“Cronin did not purport to hold that the term ‘defect’ must remain undefined in all contexts, and did not preclude a trial court from framing a definition of defect, appropriate to the circumstances of a particular case, to guide the jury as to the standard to be applied in determining whether a product is defective or not.” [citation omitted]
Barker v. Lull Engineering Co., Inc., supra, 143 Cal.Rptr. at 234, 573 P.2d at 452. Thus, because of the additional problems presented by design defect cases, we must now add a further dimension to our previous products liability decisions.
Under the “deviation from the norm” test, the product is classified as defective because it does not match the quality of most similar products.28 While this test may be reliable in a case involving a manufacturing defect, it is inadequate for those cases where the plaintiff contends that an entire product line is defectively designed.29 Comparison with similar products manufactured by others, while helpful, would be improper if used as the definitive test for ascertaining a defect.30 The test may also be overinclusive in some fact set*882tings; e. g., where “unavoidably unsafe products” are involved.31 In addition, the burden of proving the deviation would fall on the plaintiff. Since one of the major goals of strict products liability is to relieve the plaintiff of the burdensome evidentiary requirements of the negligence cause of action, this test would unduly burden the plaintiff.
The “unfitness for intended purpose” test of defectiveness originated “as a basis for deciding when purchasers could recover from sellers for intangible financial and commercial losses” resulting from frustration of the purchaser’s expectations as to what the product would do.32 It was intended as a test to determine when a manufacturer would be liable for injuries caused by its products; but the adaptation of the test from commercial expectations suggests that the only basis for liability is consumer expectations.33 Accordingly, plaintiffs are more likely to recover when the danger is hidden than when it is patent. This is unacceptable because we have already determined that any distinction between patent and latent defects is unnecessarily restrictive to plaintiffs. See Butaud v. Suburban Marine & Sporting Goods, Inc., 543 P.2d 209, 214 (Alaska 1975); Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 448-449, 501 P.2d 1163, 1168-69 (1972).34 Also, this test operates to shield a defendant from liability as long as the product does not fall below the ordinary customer’s expectations as to the product’s safety. Consumer expectations are a factor to be considered in determining defectiveness, but the public policy supporting strict liability would be poorly served if consumer expectations were the sole boundary of liability.35
The Restatement (Second) of Torts, section 402A, provides a two-prong test for the seller’s liability. Defectiveness is established if the product leaves the seller’s hands in an “unreasonably dangerous” condition and such condition is not contemplated by the ultimate consumer.36 Thus, recovery would not be allowed for obvious or generally known dangers. Although the Restatement test improves upon the deficiencies of the above tests, we adhere to our previously expressed view that the “unreasonably dangerous” terminology of the Restatement unnecessarily limits the scope of *883liability and unduly increases the plaintiff’s burden of proof.37
Two commentators, Deans Page Keeton and John Wade,38 have proposed similar risk/utility analysis tests. Their approaches recognized that evaluation of product defectiveness involves multiple factors, and they attempt to fulfill the purposes of strict products liability by creating a framework for balancing the relevant factors without shifting the inquiry to what the manufacturer should have foreseen at the time the product was sold.
Dean Wade asserts that it is possible to use “a tort way of thinking and tort terminology” in relation to strict products liability — “to abandon the warranty way of thinking and its terminology” reflected in the “unfitness for intended purpose” test— without eroding the purposes served by strict liability and without retreating to negligence.39 He would have the trial court first determine, by comparing product risk with product utility, whether it would be reasonable to allow the jury to find for the plaintiff. If so, the jury would then be instructed as follows:
“A [product] is not duly safe if it is so likely to be harmful to persons [or property] that a reasonable prudent manufacturer [supplier], who had actual knowledge of its harmful character would not place it on the market. It is not necessary to find that this defendant had knowledge of the harmful character of the [product] in order to determine that it was not duly safe.”40
Dean Keeton’s approach is that a product is defective if it is unreasonably dangerous as marketed. It is unreasonably dangerous if a reasonable person would conclude that the magnitude of the scientifically perceivable danger as it is proved to be at the time of trial outweighed the benefits of the way the product was designed and marketed. Under the heading of benefits one would include anything that gives utility of some kind to the product; one would also include the infeasibility and additional cost of making a safer product.41
Because a conclusion of “defectiveness” represents a determination that the design at issue is “wrong” in the sense that we will impose legal responsibility for harm caused therefrom, we are in agreement with Deans Wade and Keeton that the final assessment of “defectiveness” necessarily implies a weighing of diverse factors related to the product’s desirability and to its dangerousness. Since evidence of these factors will be considered by the jury, we believe it is preferable to provide the jury with a proper framework in which to consider such evidence. However, we perceive no reason to reinsert negligence terminology into the analysis by instructing the jury, as suggested by Deans Wade and Keeton, to determine what a reasonable, prudent manufacturer would have done if aware of the harmful nature of the product. The focus of strict products liability is on the condition of the product, not on the manufacturing and marketing decision of the defendant. Bachner v. Pearson, 479 P.2d 319, 325 (Alaska 1970).42 Therefore,
*884indiscriminate use of reasonableness language will confuse the strict liability standard by introducing inessential attention to the conduct of the defendant, even if actual knowledge of the harmful condition is presumed rather than proven. Nor do we think it necessary to have a two-tier system where the trial judge, before giving the case to the jury, must first find that it would be reasonable for the jury to find for the plaintiff. Given proper instructions, we believe a jury can determine from the evidence whether a defect exists or not.
The California Supreme Court held in Barker v. Lull Engineering Co., Inc., supra, that:
“[A] trial judge may properly instruct the jury that a product is defective in design (1) if the plaintiff demonstrates that the product failed to perform as safely as an ordinary consumer would expect when used in an intended or reasonably foreseeable manner, or (2) if the plaintiff proves that the product’s design proximately caused his injury and the defendant fails to prove . . . that on balance the benefits of the challenged design outweighed the risk of danger inherent in such design.”
143 Cal.Rptr. at 239, 573 P.2d at 457-58.43 This dual approach was conceived in recognition of the multiplicity of “injury-produe-ing deficiencies” encompassed within the word “defect”. Id. at 235, 573 P.2d at 453. Following the California court’s decision in Cronin, wherein the court did not purport to define “defect” but rather left the problem to be resolved by “resort to a cluster of useful precedents,”44 the California Courts of Appeal devised many different tests to define product defectiveness45 Synthesizing the various approaches taken by the Courts of Appeal in dealing with particular “defect” situations, the California Supreme Court devised the above-quoted two-prong definition of defectiveness. Again, as in Cronin, the court was not attempting to formulate a mechanistic definition applicable to all situations. Rather, Barker directs the trial courts to formulate instructions which would elucidate the meaning of “defect,” within the contours of the two-prong definition, in the circumstances of the cases presented to them.
In view of the diversity of product deficiencies which could fall within the notion of defect, we are persuaded that the Barker two-prong test provides the most comprehensive guidelines for instructing juries, without compromising any of the goals of strict liability.
The Barker test represents a composite of the most workable features of each of the other tests.46 The first prong of the Barker *885test — that a product is defectively designed if it fails “to perform as safely as an ordinary consumer would expect when used in an intended or reasonably foreseeable manner”47 — incorporates notions of the implied warranty of fitness for reasonable use, a primary concept in the evolution of strict products liability,48 but eases the burden of proof on the plaintiff. In a cause of action based upon a warranty of fitness and merchantability, the plaintiff would have to prove the breach by pinpointing the actual defect.49 Under the Barker test, the plaintiff need only show, for strict liability to apply, that he used the product in an intended or reasonably foreseeable fashion and the product failed to perform in that capacity as safely as expected. Thus, this test, while giving content to the meaning of “defect,” preserves the policies we expressed in Clary.50
The second prong of the Barker definition encompasses those situations, such as the lack of a safety device which is presented here, where the product satisfies ordinary consumer expectations as to its general use but is still “defective” in that its design exposes the user or bystander to “excessive preventable danger.”51 What is excessive preventable danger must turn on the facts of each case. With the focus off the ordinary consumer’s expectations, the trier of fact balances the risk and the social utility of the product. This requires a weighing of various factors. Therefore, where necessary, the trial court may instruct the jury that if “on balance, the benefits of the challenged design outweigh the risk of danger inherent in such design,” no liability will attach. Barker, supra, 143 Cal.Rptr. at 240, 573 P.2d at 458.
Although we believe that a balancing process is inevitable in certain design defect cases, we in no way intend to diminish our adherence to the goals of strict products liability. Therefore, we also agree with the position taken by Barker regarding the allocation of the burden of proof:
“[I]nasmuch as . . .a manufacturer who seeks to escape liability for an injury proximately caused by the product’s design on a risk-benefit theory should bear the burden of persuading the trier of fact that its product should not be judged defective, the defendant’s burden is one affecting the burden of proof, rather than simply the burden of producing evidence.” (citations omitted)
143 Cal.Rptr. at 237, 573 P.2d at 455. We hold that the plaintiff need only show that he was injured and that the injury was proximately caused by the product’s design. The defendant may then avoid liability for a defectively designed product by proving by a preponderance of the evidence that, “on balance, the benefits of the challenged design outweigh the risk of danger inherent in such design.” Barker, supra, 143 Cal.Rptr. at 240, 573 P.2d at 458. This will require the fact-finder to consider and compare a number of competing factors, including but not limited to,
*886“the gravity of the danger posed by the challenged design, the likelihood that such danger would occur, the mechanical feasibility of a safer alternative design, the financial cost of an improved design, and the adverse consequences to the product and to the consumer that would result from an alternative design.”
143 Cal.Rptr. at 237, 573 P.2d at 455. Besides lessening the burdens of the plaintiff’s prima facie case, this allocation puts the burden of producing the relevant complex and technical evidence on the party who has the most access to and is the most familiar with such evidence.52
Beck protests that adding further content to the meaning of product defectiveness will be a retreat to negligence concepts, will increase the plaintiff’s burden of proof, and will create a distinction between manufacturing and design defects. As we have stated above, we intend no retreat from our holdings in Clary and Butaud I. Negligence concepts will not dilute the plaintiff’s case because the trier of fact will concentrate on the nature of the product in determining defectiveness rather than upon the conduct of the defendant.
“Thus, the fact that the manufacturer took reasonable precautions in an attempt to design a safe product or otherwise acted as a reasonably prudent manufacturer would have under the circumstances, while perhaps absolving the manufacturer of liability under a negligence theory, will not preclude the imposition of liability under strict liability principles if, upon hindsight, the trier of fact concludes that the product’s design is unsafe to consumers, users, or bystanders.” (citation omitted)
Barker, supra, 143 Cal.Rptr. at 239, 573 P.2d at 457. The plaintiff’s prima facie case in strict liability will not be made more onerous. The plaintiff’s prima facie case essentially remains the same whether the defect is one in manufacturing or in design. Our holding merely reflects that the concept of defectiveness covers a diversity of product characteristics and, because we are not willing to affix liability solely on the basis of an injury-producing product, we will allow the trial courts to formulate instructions which elucidate the concept of “defect” in the particular circumstances presented.
In summary then, we hold that the instruction approved in Clary will in many cases suffice. However, in those cases where the meaning of “defect” will be unduly vague, particularly in design defect cases, the trial court may formulate instructions which define the legal concept of “defect” for the jury. Following the guidelines set by the Barker court, we hold that the trial court may instruct the jury that a product is defectively designed if:
“(1) the plaintiff proves that the product failed to perform as safely as an ordinary consumer would expect when used in an intended or reasonably foreseeable manner, or (2) the plaintiff proves that the product’s design proximately caused injury and the defendant fails to prove, in light of the relevant factors, that on balance the benefits of the challenged design outweigh the risk of danger inherent in such design.”
143 Cal.Rptr. at 234, 573 P.2d at 452. Because the jury here was erroneously instructed on causation as the primary definition of defect, the judgment entered against appellant Caterpillar must be reversed.
Caterpillar also contends the trial court erred in refusing seven of its proposed instructions. Although Beck argues that Caterpillar waived its objection to the court’s refusal to give the requested instructions, our review of the record indicates that Caterpillar did object. Three of these proposed instructions advised the jury of the various factors to be considered in a risk/utility *887analysis of a product’s defectiveness. We need not address the issue of the propriety of the trial court’s rulings because our holding herein will guide both the court and counsel on retrial in drafting appropriate instructions in accord with this opinion.
Three other proposed instructions concerned a manufacturer’s duty in designing a product. Because these issues may recur on retrial, we will briefly discuss them. Proposed instruction 3 provided:
“A manufacturer of a product does not have a duty to design the product to make it accident proof, but only to make it safe for the use for which it is intended.”
This instruction was properly rejected because a manufacturer must consider any reasonable foreseeable use of the product as well as its intended use. Cronin v. J. B. E. Olson Corp., 8 Cal.3d 121, 104 Cal.Rptr. 433, 437, 501 P.2d 1153, 1157 (1972). Proposed instructions 4 and 8 read:
“A manufacturer is not an insurer of his product, and he has no duty to make it accident proof.”
“A manufacturer is not required in the design or manufacture of its product to incorporate only those features which represent the ultimate in design safety.”
Although it is true that a manufacturer is not the insurer of his product, it was not error for the trial court to have refused Caterpillar’s instructions. Both requested instructions, while proper in a negligence action, are improper in an action based on strict liability because they tend to focus the jury’s attention on the manufacturer’s conduct rather than the condition of the product.
Lastly, the trial court instructed the jury as follows:
“In determining whether a product is defective for the purpose of strict liability, you are instructed that it is no defense for the manufacturer to claim that his product was within the state of the art at the time of manufacture or that it was designed and manufactured in a manner similar to that of other manufacturers at the time of production.”
Caterpillar concedes that the instruction is a correct statement of the law, but contends that its proposed instruction 23 should have been given to avoid misleading the jury. Requested instruction 23 provided:
“In considering whether or not the manufacture of this [m]odel 944 front loader in 1964 by Caterpillar Tractor Company without a ROPS was a defect, you may take into consideration the state of the art existing at that time. This means you may consider what other manufacturers of construction equipment were doing so far as the installation of [ROPS] at the time of manufacture was concerned.”
Again, we find no error by the trial court. In determining whether a product is defective, the jury may consider the technological infeasibility of a different design as one factor in its analysis. However, the state of the art, or conformity with industry-wide practices will not protect a manufacturer from liability. See Olson v. A. W. Chesterton Co., 256 N.W.2d 530, 540 (N.D.1977). The jury may consider what other manufacturers were producing insofar as it indicates feasibility. See Collins v. Ridge Tool Co., 520 F.2d 591 (7th Cir. 1975); Traynor, The Ways and Meanings of Defective Products and Strict Liability, 32 Tenn.L.Rev. 363, 367-70 (1965). Since Caterpillar’s requested instruction did not make this distinction clear, the trial court properly refused it.
III. COMPARATIVE NEGLIGENCE
The trial court ruled that it would adopt the doctrine of comparative negligence.53 The judge instructed the jury that if decedent Deraid Beck “voluntarily and unreasonably assumed” a known risk of injury because of the alleged design defect,54 the *888jury should then reduce his damages by comparing “in percentage terms decedent’s own fault which has contributed to his death with the fault of the manufacturer in placing a defective product on the market.”55
We held in Butaud I56 that the defense of contributory negligence “depends on proving the user’s actual awareness of the product’s defect and his voluntary and unreasonable encounter of the risk known to him.” 543 P.2d at 211.57 See Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 501 P.2d 1163 (1972). The trial court, in instructing the jury as it did, was following the mandate of Butaud I within the framework of the recently adopted doctrine of comparative negligence which reduces plaintiff’s damages rather than bars his recovery, as did the doctrine of contributory negligence. Five months after the completion of the trial in the case at bar, we held *889in Butaud II58 that pure comparative negligence applies to products liability suits:
“We feel that pure comparative negligence can provide a predicate of fairness to products liability cases in which the plaintiff and defendant contribute to the injury. The defendant is strictly liable due to the existence of a defective condition in the product. On the other hand, the plaintiff’s liability attaches as a result of his conduct in using the product. It is appropriate, therefore, that the parties’ contribution to the injury be apportioned. The defendant is strictly liable for the harm caused from his defective product, except that the award of damages shall be reduced in proportion to the plaintiff’s contribution to his injury.” (footnotes omitted)
555 P.2d at 45-46.59
Caterpillar first contends that the instructions given were erroneous because the manufacturer’s “fault” is not at issue in strict products liability, and, therefore, cannot be compared with decedent’s “fault.” Use of the word “fault” may engender some theoretical and semantic confusion when comparing the manufacturer’s strict liability with the plaintiff’s negligence, and, consequently, it should be avoided in instructing the jury.60 Nevertheless, we do not believe the jury was misled in this case for two reasons.
First, as we have stated above, strict products liability is not absolute liability but stems from the existence of a defect. The production and marketing of a defective product is tantamount to “fault” in the sense that we will impose legal responsibility for it. Similarly, we will impose responsibility on the plaintiff for his contribution to the loss. Secondly, the jury was instructed that once it found that decedent did unreasonably confront a known risk of harm, it must apportion the decedent’s own unreasonable conduct which contributed to his death in light of the manufacturer’s total liability for placing a defective product on the market. This instruction, although five months prior to our decision in Butaud II effectively achieves the result intended by Butaud II:
“The manufacturer is still accountable for all the harm [caused by his] defective *890product, except that part caused by the consumer’s own conduct.”
555 P.2d at 46.
Caterpillar next claims that the trial court erred in limiting the defense of comparative negligence to the assumption of a known risk,61 when the decedent’s general negligence in using the product may also be considered by the jury. Beck, on the other hand, contends in her cross-appeal that it was error to submit the issue of decedent’s comparative negligence to the jury. She argues that because decedent used the loader as designed and marketed by Caterpillar, his knowledge of the lack of a ROPS cannot be considered by the jury. If it were, it would defeat the policy of strict products liability by allowing the manufacturer to escape liability when the user fails to cure the defect. Beck urges us to hold that the defense of comparative negligence is disallowed in those cases where the alleged design defect is one going to the absence of a safety feature.
We held in Butaud II that:
“The defense of comparative negligence is not limited to those cases where the plaintiff uses the product with knowledge of the defective condition, but also extends to those cases where the plaintiff misuses the product and that misuse is a proximate cause of his injuries.”
555 P.2d at 46. Therefore, Caterpillar is correct in stating that the consumer’s misuse of the product may be considered by the jury. However, as we have previously noted, Butaud II was decided five months subsequent to this trial. For reasons which we will explain below, we need not address to what extent Butaud II will be given retroactive application.
Product misuse is by necessity a broad notion capable of application in many circumstances. It encompasses all those situations where the plaintiff’s own volitional conduct contributed to the incident in concurrence with the defective product. Beck interprets product misuse to mean that there must be misuse of the “injury producing mechanism,” i. e., the lack of a ROPS. Since there was no ROPS to misuse, the doctrine of Butaud II is inapplicable, according to Beck. In other words, since her cause of action is founded upon the aggravation of injuries due to the lack of a safety feature, it would be improper to consider decedent’s general negligence in causing the accident. Caterpillar responds that such a rule would preclude any defenses at all in design defect cases where the alleged defect consists of the lack of a safety device.
We do not agree with Beck’s analysis because any negligence of the plaintiff which contributes to the chain of events leading to his injury may be considered by the jury. We will not distill the defect or the loss into piecemeal components. See Daly v. General Motors Corp., 20 Cal.3d 725, 144 Cal.Rptr. 380, 393, 575 P.2d 1162, 1175 (1978). Rather, in applying comparative negligence, we will consider each parties’ responsibility in light of the total loss suffered in the particular incident as a whole. In any event, in this case the issue of decedent’s general negligence in causing the accident is not before us, for the thrust of Caterpillar’s assertion of decedent’s comparative negligence is that Beck perceived (or should have perceived) the danger of operating the loader without a ROPS and was himself negligent in failing to install a ROPS.
We do not believe that a consumer who uses a product as it was intended to be used, and who knows or should know of the lack of a safety device, can be deemed to have misused the product within the meaning of Butaud II. If the jury finds that a product is defective by virtue of its lack of a safety feature, plaintiff’s failure to install such a device will not reduce his recovery *891based upon his mere knowledge of the inadequate safety features on the product. This is not to say that a manufacturer of such a product has no defense. Rather, as was instructed here, the standard of Butaud I would apply.
A similar issue was presented in Bexiga v. Havir Manufacturing Corp., 60 N.J. 402, 290 A.2d 281 (1972).62
In Bexiga the lack of a safety device on a punch press was held to render the product “unreasonably dangerous” to the user. The manufacturer contended, however, that since it was impractical for him to equip all the presses with protective devices, the user was contributorily negligent for failure to supply safety features himself. The court held that the manufacturer could not absolve himself of liability by expecting the purchaser to provide safety features.
“Where a manufacturer places into the channels of trade a finished product which can be put to use and which should be provided with safety devices because without such it creates an unreasonable risk of harm, and where such safety devices can feasibly be installed by the manufacturer, the fact that he expects that someone else will install such devices should not immunize him. The public interest in assuring that safety devices are installed demands more from the manufacturer than to permit him to leave such a critical phase of his manufacturing process to the haphazard conduct of the ultimate purchaser. The only way to be certain that such devices will be installed on all machines — which clearly the public interest requires — is to place the duty on the manufacturer where it is feasible for him to do so.”
290 A.2d at 285.
Various policy considerations support this approach. First, the general policy of strict liability demands that responsibility for placing the defective product on the market should not be shifted to those in no position to realistically assess the danger. If the focus is on the nature of the product as defective, and the jury has found the lack of safety device to render the product defective, it is inconsistent to turn around and reduce the user’s recovery merely because he bought and used the product as marketed. See 2 Frumer & Friedman, Products Liability, section 16A [5][g] (1960 & Supp. 1978). Second, accidents are inevitable costs, and it is better to place the burden of such costs on the party best able to spread the losses. Third, the public interest in reducing the level of accidents makes it reasonable to put the cost burden on the party best able to prevent accidents.
Caterpillar was in a better position than the Beck family business to ascertain the risks and the means to avoid such risks. Caterpillar had access to statistics which afforded it the opportunity to balance risks and costs. Beck’s business was family run with very little business sophistication. They operated with large deficits and lacked sufficient objectivity to evaluate whether their own expense in installing ROPS would be warranted. See Note, Bexiga v. Havir Manufacturing Corp., 86 Harv. L.Rev. 923 (1973). It ignores economic reality to assume that many users will make safety alterations at their own expense.
“With the case in this two-party posture [injured employee and manufacturer], even reasonable expectations that an operative product which has been processed through all its manufacturing stages will be altered before its use seem inappropriate as a defense to strict liability, for the policies underlying strict liability continue to apply where such expectations exist.”
86 Harv.L.Rev. at 925-26.
The Bexiga court stated that different considerations might lead to a different result in the case of a product designed *892for multiple functions. This is how Caterpillar characterizes the loader involved herein. Although that analysis will also be a factor in the jury’s decision on defect, it would be proper to reduce the plaintiff’s recovery for an unreasonable failure to correct a patently unsafe product. However, comparative negligence cannot be used when the plaintiff’s negligence was nothing more than knowledge of the defect.63 Were it otherwise, we would be resurrecting, sub silentio, the distinction between patent and latent defects which was abolished in Butaud I. See Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 501 P.2d 1163 (1972).64
Therefore, we hold that the instructions given by the trial court were not erroneous. When the design defect is the lack of a safety device, the jury may be instructed that the plaintiff may be comparatively negligent in the knowing use of a defective product only if he voluntarily and unreasonably encounters the known risk.
IV. EVIDENTIARY AND PROCEDURAL RULINGS
Caterpillar asserts error in the exclusion of evidence which would have tended to show its marketing arrangements with various equipment dealers, its efforts (after the manufacture and sale of the 944 model) to educate dealers about the danger of operating loaders without ROPS, and the efforts undertaken by dealers to reach purchasers with information regarding ROPS.
The bulk of this evidence was properly excluded either because it was irrelevant or because the foundation for the evidence was not firmly established. To the extent that any portions of the excluded evidence could show in a direct manner the comparative fault of the decedent, they may be offered again at any retrial. We find no reason to review the excluded evidence in detail.
Similarly, we do not reach the question of whether the court erred in requiring Caterpillar and Nenana Excavators, since dismissed as a party, to share three peremptory challenges between them.
REVERSED and REMANDED.65
BURKE, J., concurring and dissenting.
DIMOND, Senior Justice, dissenting and concurring.
MATTHEWS, J., not participating.. It is unclear whether Northern Commercial or the prior owner of the loader attached the weather canopy.
. It is customary to operate the loader in reverse because the vehicle is geared through the rear wheels and attains its highest speeds in reverse gear.
. Jack Hasten, a manager of sales development for Caterpillar, testified that his study of the accident pictures indicated that Beck was “roading” the loader at approximately 15 miles per hour near the shoulder of the road. Beck then veered sharply to the left and the shoulder of the roadway caved in, sending the loader into a violent roll rather than the more gentle roll described by the state^trooper.
. There were two earlier regulations requiring ROPS devices. The first was issued in 1960 by the Army Corps of Engineers. This regulation affected only the Pacific Northwest and was not widely known. The first state to require ROPS was California in 1965. It was this regulation, while ROPS was still a relatively novel concept, which motivated manufacturers to begin ROPS testing and development.
. Caterpillar also alleges error by the trial court in its refusal to give both defendants, Caterpillar and Nenana Excavators, three peremptory challenges each. Rule 47(d), Alaska R.Civ.P. This issue will not recur upon remand because Nenana Excavators has been dismissed from the case.
. The challenged instruction reads in full:
*877“At this point in the instructions, I would like to explain the general law of strict products liability to you.
A manufacturer is strictly liable in tort when an article he places on the market proves to have a defect that causes injury to a human being.
A defendant manufacturer may be liable under a theory of strict liability without proof of its own fault or negligence in bringing about the injury of which a plaintiff complains. In this case, the [p]laintiff has alleged that a design defect proximately caused the death of the deceased. A design defect is one in which the product, however perfectly manufactured, incorporates or fails to incorporate a design feature with the result that injury is proximately caused thereby.”
. Beck also contends that Caterpillar cannot raise its risk/utility definition on appeal because it was not properly raised below. Caterpillar objected to the challenged instruction at trial on the same grounds as stated above. Also, although Caterpillar did not request a risk/analysis instruction by name, three of its proposed instructions, which Caterpillar contends on appeal were erroneously rejected by the trial court, would have had the jury consider risk/utility factors in determining if the lack of a ROPS was a design defect. The error now at issue was properly preserved. See Spruce Equipment Co. v. Maloney, 527 P.2d 1295, 1301 (Alaska 1974); Reader v. Ghemm Co., 490 P.2d 1200, 1202 n. 1 (Alaska 1971).
. See, e. g., Greenman v. Yuba Power Products, Inc., 59 Cal.2d 57, 27 Cal.Rptr. 697, 377 P.2d 897 (1962) (discussing both theories); Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 161 A.2d 69 (1960) (warranty). See generally, Daly v. General Motors Corp., 20 Cal.3d 725, 144 Cal.Rptr. 380, 383-384, 575 P.2d 1162, 1165-66 (1978); W. Prosser, Law of Torts § 96, at 641-44 (4th ed. 1971).
. See generally, Keeton, Product Liability and the Meaning of Defect, 5 St. Mary’s L.J. 30, 34-35 (1973).
. See, e. g., Greenman v. Yuba Power Products, Inc., supra, note 8; Henningsen v. Bloomfield Motors, Inc., supra, note 8; Goldberg v. Kollsman Instrument Corp., 12 N.Y.2d 432, 240 N.Y.S.2d 592, 191 N.E.2d 81 (1963).
. Clary v. Fifth Avenue Chrysler Center, Inc., supra, 454 P.2d at 247; quoting Greenman v. Yuba Power Products, Inc., supra, 27 Cal.Rptr. at 700, 377 P.2d at 900.
. Clary v. Fifth Avenue Chrysler Center, Inc., supra, 454 P.2d at 248.
. Restatement (Second) of Torts § 402A (1965). Section 402A provides in part that one selling “any product in a defective condition unreasonably dangerous to the user or consumer or to his property” is strictly liable in tort. After Greenman, several California products liability decisions had utilized the “unreasonably dangerous” terminology. See, e. g., Jiminez v. Sears, Roebuck & Co., 4 Cal.3d 379, 93 Cal.Rptr. 769, 772, 482 P.2d 681, 684 (1971); Pike v. Frank G. Hough Co., 2 Cal.3d 465, 85 Cal.Rptr. 629, 636, 467 P.2d 229, 236 (1970); Putensen v. Clay Adams, Inc., 12 Cal.App.3d 1062, 91 Cal.Rptr. 319, 325 (1970).
. In Butaud I, we also expressly adopted the rule of Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 501 P.2d 1163 (1972) which abrogated the distinction which had previously been made between patent and latent defects. Luque held that a plaintiff suing in strict liability need not prove that he was unaware of the defect in order to recover.
. Products may be defective for at least three reasons:
1.Flawed fabrication or construction (i. e., deviating from the condition intended by the manufacturer),
2. Improper design (i. e., conforming to the design intended by the manufacturer but producing unacceptable consequences), and
3. Misinformation or inadequate information about risks involved in using the product or about minimizing or avoiding harm from such risks.
Manufacturing defects and design errors are not mutually exclusive, however. As the California court recognized in Cronin, whether a defect results from manufacture or design is not always obvious. 104 Cal.Rptr. at 443, 501 P.2d at 1163. One commentator illustrates the problem as follows:
“For example, a workman has in front of him several sized screws. In the first situation, a number seven screw is called for by the plan. He, however, inadvertently takes and inserts a number four screw which proves inadequate. This mistake is one of construction. If, however, the plan is silent as to the size of the screw, and the workman in exercising his judgment chooses a screw which is not adequate, he has made a mistake more similar to the design defect. It might also be said that the failure of the planner to designate the size screw was in itself a design defect.”
Torke, Foreseeability in Product Design and Duty to Warn Cases — Distinctions and Misconceptions, 1968 Wis.L.Rev. 228, 233 n. 23. The categories not only overlap, but they may also operate simultaneously or be alleged in the alternative.
. See generally, Seattle-First Nat’l Bank v. Tabert, 86 Wash.2d 145, 542 P.2d 774 (1975); Dickerson, The ABC’s of Products Liability With a Close Look at Section 402A and the Code, 36 Tenn.L.Rev. 439 (1969); Keeton, Manufacturer’s Liability: The Meaning of "Defect’’ in the Manufacture and Design of Products, 20 Syracuse L.Rev. 559 (1969).
. Butaud v. Suburban Marine & Sporting Goods, Inc., 555 P.2d 42 (Alaska 1976); Cronin v. J. B. E. Olson Corp., 8 Cal.3d 121, 104 Cal. Rptr. 433, 501 P.2d 1153 (1972); Turner v. Manning, Maxwell & Moore, Inc., 216 Va. 245, 217 S.E.2d 863 (1975); Dippel v. Sciano, 37 Wis.2d 443, 155 N.W.2d 55, 63 (1967).
. See Keeton, supra note 9, 5 St. Mary’s L.J., at 33; Lascher, Strict Liability in Tort for Defective Products: The Road to and Past Vandermark, 38 S.Cal.L.Rev. 30, 47 (1965); Traynor, The Ways and Meanings of Defective Products and Strict Liability, 32 Tenn.L.Rev. 363, 366-67 (1965); Wade, Strict Tort Liability of Manufacturers, 19 Sw.L.J. 5, 13 (1965).
. Traynor, supra note 18, 32 Tenn.L.Rev., at 366-67.
. “Design issues present problems which carry over not only to matters of evidence and proof but also to the instructions to the jury. In matters involving design, the main issue may well focus upon how much design safety is enough. If courts are to render such judgments, ‘they must either accept and apply particularized standards developed elsewhere or develop standards by themselves.’ ” (footnote omitted)
Hoenig, Products Designs and Strict Tort Liability: Is There a Better Approach?, 8 Sw.U.L. Rev. 109, 121 (1976); See Beron v. Kramer-Trenton Co., 402 F.Supp. 1268, 1276 (E.D.Pa.1975).
. “If a jury in determining liability for a defect in design is instructed only that it should decide whether or not there is ‘a defective design,’ it may reach to the extreme conclusion that the plaintiff, having suffered injury, should without further showing, recover; on the other hand, it may go to the opposite extreme and conclude that because the product matches the intended design the plaintiff, under no conceivable circumstance, could recover.”
Barker v. Lull Engineering Co., Inc., 20 Cal.3d 413, 143 Cal.Rptr. 225, 239, 573 P.2d 443, 457 (1978). Since neither of these results is desirable, we must strike a balanced approach.
. See, e. g., Elmore v. American Motors Corp., 70 Cal.2d 578, 75 Cal.Rptr. 652, 451 P.2d 84 (1969); Vandermark v. Ford Motor Co., 61 Cal.2d 256, 37 Cal.Rptr. 896, 391 P.2d 168 (1964).
. See, e. g., Dunham v. Vaughan & Bushneii Mfg. Co., 42 Ill.2d 339, 247 N.E.2d 401 (1969); Ginnis v. Mapes Hotel Corp., 86 Nev. 408, 470 P.2d 135 (1970); Lunt v. Brady Mfg. Corp., 13 Ariz.App. 305, 475 P.2d 964 (1970).
. See Phillips v. Kimwood Machine Co., 269 Or. 485, 525 P.2d 1033 (1974); cf. Seattle-First Nat’I Bank v. Tabert, 86 Wash.2d 145, 542 P.2d 774 (1975) (a hybrid test involving some elements of risk/utility analysis).
. Clary v. Fifth Avenue Chrysler Center, Inc., 454 P.2d 244 (Alaska 1969). As we mentioned earlier in this opinion, in Clary we announced that we would follow the strict liability test articulated by the California Supreme Court in Greenman v. Yuba Power Products, 59 Cal.2d 57, 27 Cal.Rptr. 697, 377 P.2d 897 (1962):
“A manufacturer is strictly liable in tort when an article he places on the market, knowing that it is to be used without inspection for defects, proves to have a defect that causes injury to a human being.”
Clary, supra at 247, quoting Greenman, supra 27 Cal.Rptr. at 700, 377 P.2d at 900. We also adopted in Clary the extension of the Green-man strict liability doctrine to impose liability upon retailers, as held by the California Supreme Court in Vandermark v. Ford Motor Co., 61 Cal.2d 256, 37 Cal.Rptr. 896, 391 P.2d 168 (1964). Our opinion in Clary, however, did not set forth a specific definition of defect.
. Butaud v. Suburban Marine and Sporting Goods, Inc., 543 P.2d 209 (Alaska 1975). We reaffirmed in Butaud the strict liability test adopted in Clary, supra, n. 25, but again set forth no specific definition of defect.
. Cronin v. J. B. E. Olson Corp., 8 Cal.3d 121, 104 Cal.Rptr. 433, 501 P.2d 1153 (1972).
. Jiminez v. Sears, Roebuck & Co., 4 Cal.3d 379, 93 Cal.Rptr. 769, 772, 482 P.2d 681, 684 (1971). As Chief Justice Traynor explained the test, “If a normal sample of defendant’s product would not have injured plaintiff, but the peculiarities of the particular product did cause harm, the manufacturer is liable for injuries caused by this deviation.” Traynor, supra note 18, 32 Tenn.L.Rev., at 367.
. One commentator views injury-producing design characteristics as being of two kinds: inadvertent design errors and conscious design choices. Inadvertent design errors are similar to manufacturing flaws in that they are unintended, are usually hidden, and often defeat the very purposes for which the products were produced and marketed. Conscious design choices are intended, are more likely to be obvious to users and consumers, and are rarely in direct conflict with the product’s intended functions. Henderson, Judicial Review of Manufacturers’ Conscious Design Choices: The Limits of Adjudication, 73 Colum.L.Rev. 1531, 1548-50 (1973). Professor Henderson argues that both manufacturing flaws and inadvertent design errors are appropriately evaluated by courts because external standards are readily available. However, he asserts that conscious design choices involve “polycentric” problems which courts are institutionally ill-equipped to evaluate.
. Consistency with industry-wide practices has never been conclusive in strict liability or negligence actions. Olson v. A. W. Chesterton Co., 256 N.W.2d 530, 540 (N.D.1977); The T. J. Hooper, 60 F.2d 737, 740 (2d Cir.1932); Phillips, The Standard for Determining Defectiveness in Products Liability, 46 U.Cin.L.Rev. 101 (1977). Liability may attach even if dangerous design characteristics are the norm in a particular industry. See Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 501 P.2d 1163 (1972).
. See Traynor, supra note 18, 32 Tenn.L.Rev., at 367.
. Keeton, supra note 9, 5 St. Mary’s L.J., at 36.
. Id. at 37. A consumer’s reasonable expectations may be frustrated in several ways:
(1) The product fails to perform its intended function (e. g., brakes that fail);
(2) The product creates a danger against which it was supposed to guard (e. g., polio vaccine that causes polio);
(3) The product performs its intended function but creates a bad side effect (e. g., cigarettes that cause cancer);
(4) The product fails to minimize avoidable consequences in the event of an accident (e. g., automobile without head rest).
Dickerson, The ABC’s of Products Liability— With a Close Look at Section 402A and the Code, 36 Tenn.L.Rev. 439, 454 (1969).
. If manufacturers escape liability when the dangerous characteristics are obvious, two purposes of strict liability are undercut: risk allocation and risk avoidance. Note, Reasonable Product Safety: Giving Content to the Defectiveness Standard in California Strict Products Liability Cases, 10 U.S.F.L.Rev. 492, 504 n. 82 (1976).
. Consumer expectations cannot be the exclusive test because “[i]n many situations . the consumer would not know what to expect, because he would have no idea how safe the product could be made.” Wade, On the Nature of Strict Tort Liability for Products, 44 Miss. L.J. 825, 829 (1973).
. Restatement (Second) of Torts § 402A, Comments g and i (1965).
Courts and commentators have disagreed on whether the Restatement’s requirements of “defective” and “unreasonably dangerous” establish one or two standards. Dean Wade traces the history of section 402A and concludes that the Restatement’s handling of the two terms produced essentially synonymous terms. See Seattle-First Nat’l Bank v. Tabert, supra note 24, 542 P.2d at 779. Wade, supra note 35, at 829-33. See Borel v. Fibreboard Paper Prods. Corp., 493 F.2d 1076, 1087 (5th Cir. 1973). Compare Polelle, The Foreseeability Concept and Strict Products Liability: The Odd Couple of Tort Law, 8 Rut.-Cam.L.J. 101, 110 (1976).
. See Butaud I, 543 P.2d 209 (Alaska 1975); Butaud v. Suburban Marine & Sporting Goods, Inc., 555 P.2d 42 (Alaska 1976) (hereinafter Butaud II).
. See Keeton, supra note 9, 5 St. Mary’s L.J., at 37. An earlier formulation appears in Kee-ton, Manufacturer’s Liability: The Meaning of “Defect" in the Manufacture and Design of Products, 20 Syracuse L.Rev. 559, 568 (1969); Wade, On the Nature of Strict Tort Liability for Products, 44 Miss.L.J. 825, 834-35 (1973); Wade, Strict Tort Liability of Manufacturers, 19 Sw.L.J. 5, 15-17 (1965).
. Wade, supra note 38, 44 Miss.L.J., at 834— 35.
. Id. at 839-40.
. Keeton, supra note 9, 5 St. Mary’s L.J., at 37-38. Keeton notes that his test of defect makes no allowance for the manufacturer’s excusable ignorance of the risk. Id. at 39.
. “The change in the substantive law as regards the liability of makers of products and other sellers in the marketing chain has been from fault to defect. The plaintiff is no longer required to impugn the maker, but he is required to impugn the product.”
Keeton, supra note 9, 5 St. Mary’s L.J., at 33. See Note, Reasonable Product Safety: Giving *884Content to the Defectiveness Standard in California Strict Products Liability Cases, 10 U.S.F. L.Rev. 492, 516-17 (1976).
. Other courts have adopted a similar test. See, e. g., Welch v. Outboard Marine Corp., 481 F.2d 252, 254 (5th Cir. 1973); Henderson v. Ford Motor Co., 519 S.W.2d 87, 92 (Tex.1974).
. Cronin v. J. B. E. Olson Corp., supra, quoting Traynor, supra note 18, 32 Tenn.L.Rev., at 373.
. See, e. g., Buccery v. General Motors Corp., 60 Cal.App.3d 533, 132 Cal.Rptr. 605, 614 (1976); Sabich v. Outboard Marine Corp., 60 Cal.App.3d 591, 131 Cal.Rptr. 703, 706 (1976); Baker v. Chrysler Corp., 55 Cal.App.3d 710, 127 Cal.Rptr. 745, 748-49 (1976); Dosier v. Wilcox & Crittendon Co., 45 Cal.App.3d 74, 119 Cal.Rptr. 135, 138-39 (1975); Self v. General Motors Corp., 42 Cal.App.3d 1, 116 Cal.Rptr. 575, 578 (1974); Balido v. Improved Machinery, 29 Cal.App.3d 633, 105 Cal.Rptr. 890, 896 (1963). See also, Jackson, Cronin and Luque: The Search for a Standard in Strict Liability in Tort, 10 U.S.F. L.Rev. 725 (1976); Note Reasonable Product Safety: Giving Content to the Defectiveness Standard in California Strict Liability Cases, 10 U.S.F. L.Rev. 492, 504 N. 82 (1976).
. Other models of “defectiveness” standards have been suggested, but we have not discussed them either because of their overlap with other approaches or because of their inconsistency with the orientation previously expressed by this court. See, e. g., Hoenig, Product Designs and Strict Tort Liability: Is There a Better Approach ? 8 Sw.U.L.Rev. 109 (1976) (arguing for a design defect test which openly uses negligence concepts); Phillips, The Standard for Determining Defectiveness in Products Liability, 46 U.Cin.L.Rev. 101 (1977) (arguing for a test based on foreseeability and feasibility rather than due care); Polelle, The Foreseeability Concept and Strict Products Liability: The Odd Couple of Tort Law, 8 Rut.-Cam.L.J. 101 (1976) (proposing two specific formulations based heavily on risk spreading).
. Barker v. Lull Engineering Co., Inc., 20 Cal.3d 413, 143 Cal.Rptr. 225, 239, 573 P.2d 443, 457 (1978).
. See Clary v. Fifth Avenue Chrysler Center, Inc., supra, 454 P.2d at 252 (Rabinowitz, Justice, dissenting in part, concurring in part); Greenman v. Yuba Power Products, Inc., supra, 27 Cal.Rptr. at 701, 377 P.2d at 901.
. The Uniform Commercial Code’s implied warranty of merchantability requires that goods “be fit for the ordinary purposes for which the goods are used.” AS 45.05.-096(b)(3). Recovery is permitted for “injury to persons or property proximately resulting from a breach of warranty,” /. e., for consequential damages. AS 45.05.222(b)(2).
. “The purpose of imposing such strict liability on the manufacturer and retailer is to insure that the cost of injuries resulting from defective products are borne by the manufacturers that put such products on the market rather than by the injured persons who are powerless to protect themselves. Sales warranties serve this purpose fitfully at best.”
Clary v. Fifth Avenue Chrysler Center, Inc., supra, 454 P.2d at 248.
. Barker v. Lull Engineering Co., Inc., supra, 143 Cal.Rptr. at 236, 573 P.2d at 454; Self v. General Motors Corp., 42 Cal.App.3d 1, 116 Cal.Rptr. 575, 578 (1974).
. The established rule is that the product must be shown to be defective when it left the defendant’s possession or control. Preissman v. Ford Motor Co., 1 Cal.App.3d 841, 82 Cal.Rptr. 108 (1969); Simmons v. Koeteeuw, 5 Wash.App. 572, 489 P.2d 364 (1971). See generally, Hursh & Bailey, American Law of Products Liability (1974), Sec. 1:10, Sec. 4:16. It follows that design defects must be measured by the knowledge and information which existed when the product left the hands of the manufacturer.
. We adopted the doctrine of pure comparative negligence in Kaatz v. State, 540 P.2d 1037 (Alaska 1975). Under this standard, the plaintiffs damages are reduced proportionally to his contribution to the injury, whether that proportion be 99 percent or 1 percent.
. Instruction No. 14 stated:
*888“Defendant Caterpillar Tractor Company has raised the affirmative defense that the deceased, Deraid Allen Beck, voluntarily and unreasonably assumed a known risk of injury from the alleged design defect in the Caterpillar product which Plaintiff contends caused his death.
To establish this defense, the Defendant has the burden of proving by a preponderance of the evidence the following propositions:
1. That the deceased, Deraid Allen Beck, had an actual awareness of the alleged design defect in the Caterpillar product; that is, he had an awareness of the absence of a roll bar or roll over protection structure;
2. That the deceased, Deraid Allen Beck, knew of the risk and of the danger in operating the front end loader without roll bar protection;
3. That the deceased, Deraid Allen Beck, voluntarily and unreasonably encountered this risk known to him.
In order for the user to have assumed such risk, he must have had actual knowledge of the defect and an appreciation of the risk or danger involved in using the defective product together with an understanding of the magnitude of such risk and must thereafter have voluntarily and unreasonably proceeded to use the product to his injury.
In determining whether the deceased, Der-ald Allen Beck, unreasonably proceeded to use the product, you should consider what a reasonably prudent person with the same knowledge would have done under the same or similar circumstances.
For a person to act voluntarily, he must have freedom of choice. This freedom of choice must come from the circumstances that provide him a reasonable opportunity to safely refuse to expose himself to the danger in question.
In determining whether the deceased, Der-ald Allen Beck, assumed such risk, you may consider his maturity, intelligence, experience and capacity, along with all the other surrounding circumstances as shown by the evidence.
This defense operates to reduce a plaintiff’s recoverable damages in the proportion to which the plaintiffs fault bears to the total fault of himself and the fault of the manufacturer in placing such defective product on the market.”
. Instruction No. 38 states in part:
“If you should find from the preponderance of the evidence in the case that the deceased, Deraid Allen Beck, knew of such design defect and voluntarily and unreasonably proceeded to use the defective product and to confront a known risk of harm, then you must compare and indicate in percentage terms decedent’s own fault which has contributed to his death with the fault of the manufacturer in placing a defective product on the market which you may find has concurred with the decedent’s fault in causing his death. The combined fault of plaintiff and defendant, if any, must total one hundred percent. It would then be my duty to reduce the total recovery of damages you may award the plaintiff by an amount equal to the percentage of fault attributable to the decedent, Deraid Allen Beck.”
. Butaud v. Suburban Marine & Sporting Goods, Inc., 543 P.2d 209 (Alaska 1975).
. The defense of assumption of the risk was abolished in Leavitt v. Gillaspie, 443 P.2d 61 (Alaska 1968) to the extent that it was distinct from the concept of contributory negligence. 443 P.2d at 67-69. However, the notion of an unreasonable encounter with a known risk rather than the mere assumption of a known risk, was held to be still viable in the products liability context as a variant of contributory negligence. See Butaud I, supra, note 56, at 214 n. 22; Bachner v. Pearson, 479 P.2d 319, 329-30 and n. 24 (Alaska 1970). Likewise, it will now be properly considered as subsumed within the concept of comparative negligence. See Kaatz v. State, 540 P.2d 1037 (Alaska 1975); Daly v. General Motors Corp., 20 Cal.3d 725, 144 Cal.Rptr. 380, 385, 575 P.2d 1162, 1167 (1978).
. Butaud v. Suburban Marine & Sporting Goods, Inc., 555 P.2d 42 (Alaska 1976).
. We will not reiterate our rationale for extending comparative negligence principles to actions in strict products liability, but we note that many of the states which have adopted comparative negligence, either judicially or by statute, have also made it applicable in strict liability suits. See, e. g., Edwards v. Sears, Roebuck and Co., 512 F.2d 276, 290 (5th Cir. 1975); Sun Valley Airlines Inc. v. Avco-Lycoming Corp., 411 F.Supp. 598, 602-603 (D.Idaho 1976); Hagenbuch v. Snap-On Tools Corp., 339 F.Supp. 676, 681-83 (D.N.H.1972); Daly v. General Motors Corp., 20 Cal.3d 725, 144 Cal. Rptr. 380, 390, 575 P.2d 1162, 1172 (1978); West v. Caterpillar Tractor Co., Inc., 336 So.2d 80, 90 (Fla. 1976); General Motors Corp. v. Hopkins, 548 S.W.2d 344, 351-52 (Tex. 1977); Dippel v. Sciano, 37 Wis.2d 443, 155 N.W.2d 55, 64 (1967); contra, Melia v. Ford Motor Co., 534 F.2d 795, 802 (8th Cir. 1976); Kirkland v. General Motors Corp., 521 P.2d 1353, 1367-68 (Okl.1974); Kinard v. Coats Co., Inc., 37 Colo. App. 555, 553 P.2d 835, 837 (1976). In addition, the majority of commentators considering the issue have urged the application of comparative negligence to strict liability. See, e. g., Epstein, Products Liability: Defenses Based on Plaintiffs Conduct, 1968 Utah L.Kev. 267, 284; Fleming, The Supreme Court of California 1974-1975, Foreword: Comparative Negligence at Last — By Judicial Choice, 64 Cal.L.Rev. 239, 269-71 (1976); 2 Frumer & Friedman, Products Liability § 16A [5][g] (1960 & Supp.1978); Schwartz, Strict Liability and Comparative Negligence, 42 Tenn.L.Rev. 171 (1974); Schwartz, Comparative Negligence § 12.1 et seq. (1974 & Supp.1978); Wade, A Uniform Comparative Fault Act — What Should It Provide?, 10 U.Mich.J.L.Ref. 220 (1977); contra Levine, Strict Products Liability and Comparative Negligence: The Collision of Fault and No-Fault, 14 San Diego L.Rev. 337 (1977). The Uniform Comparative Fault Act also extends comparative principles to strict liability actions and states that the term “fault” encompasses causes of action in both negligence and strict liability. Uniform Comparative Fault Act, § 1(a) and (b) (1977).
. See the discussion in Butaud II, supra, 555 P.2d at 45-47, and in Justice Rabinowitz’s concurring opinion, 555 P.2d at 47.
. The jury found by special verdict that decedent Beck did “know of such design defect and voluntarily and unreasonably [did] proceed to use the defective product and to confront a known risk of harm.” Nevertheless, Caterpillar claims that it was harmed by the instructions because the trial court did not allow it to offer evidence on the question of its fault or lack of fault.
. Although Bexiga presents a similar issue, there are two important distinctions to be kept in mind. First, New Jersey follows the Restatement test for defect, unlike this court. Second, and more importantly, the New Jersey court was concerned with the applicability of contributory negligence which would have barred recovery rather than just reduced it as in comparative negligence.
. This is especially true where there is a large used equipment market. The buyer in that market is even less likely to have the data on the dangers of operating without ROPS and the means to forestall such danger.
. See also Buccery v. General Motors Corp., 60 Cal.App.3d 533, 132 Cal.Rptr. 605 (1976), which involved a design defect for lack of a head restraint device. The court of appeal refused to apply comparative negligence to the plaintiff’s use of the vehicle even though the plaintiff testified that he knew it quite dangerous to drive without such a device. The court reasoned that it was doubtful that Li v. Yellow Cab Co., 13 Cal.3d 804, 119 Cal.Rptr. 858, 532 P.2d 1226 (1975) (the California comparative negligence decision) would supersede the position taken in Luque v. McLean, 8 Cal.3d 136, 104 Cal.Rptr. 443, 501 P.2d 1163 (1972), which did away with the plaintiffs burden of proving that he was unaware of the defect. Luque further held that the only defense to strict liability could be a voluntary and unreasonable encounter of a known danger.
. We commend all counsel for the thoughtful advocacy presented in both the briefs and oral argument.