*416In this class action against defendant Kia Motors America, Inc. (KMA), plaintiff class of 8455 Kia Sephia owners and lessees represented by Regina Little proved at a jury trial that the Sephia, model years 1997 through 2000, had a defective front brake system, which caused premature brake pad and rotor wear. Concluding that the defect amounted to a breach of express and implied warranties, and that all owners had suffered damage due to the defect, the jury awarded each member of the class $750 ($6.3 million total) in repair damages.
Determining for the first time post-trial that repair damages could not be awarded on a class-wide basis because they were dependent upon individual factors, the trial court granted KMA's motion for judgment *505notwithstanding the verdict (JNOV) on the repair damages award, decertified the class for purposes of damages, and ordered a new trial on repair damages only, to proceed by way of claim forms. With the advantage of recent case law unavailable to the trial judge, we now reverse, reinstate the jury award and remand for determination of counsel fees.
I.
We recount only the facts and procedural history relevant to this appeal. We begin with the procedural history. On June 26, 2001, Little filed an amended class action complaint on behalf of herself and others similarly situated, against defendant, a California corporation with offices in New Jersey. The putative class alleged that the Sephia had a defective front brake system and asserted causes of action for: fraudulent business practices in violation of California law and the New Jersey Consumer Fraud Act (CFA), N.J.S.A. 56:8-1 to -210; breach of an express warranty; breach of the implied warranty of merchantability; and failure to comply with the federal Magnuson-Moss Warranty Improvement Act (MMWA), 15 U.S.C. §§ 2301 to 2312.
In August 2003, the court granted class certification. Prior to trial, the trial judge heard a number of pretrial motions on the *417admissibility of evidence. Defendant moved unsuccessfully to exclude as net opinions the class expert testimony of Raymond King, on repair damages, and John Matthews, on diminution of value damages.
After a month-long trial, in June 2008 the jury returned a verdict finding that defendant had breached the express and implied warranties as well as the MMWA, but that it had not violated the CFA. The jury found that the class had suffered damages and awarded each member repair damages. It awarded no damages for diminution in value.
In a November 24, 2008 written decision, the trial judge granted defendant's motion for JNOV as to repair damages only, decertifying the class for purposes of damages only based on the finding that individual factors predominated, and ordered a new trial on repair damages to proceed by way of claim forms.
In a January 2011 decision, another judge granted plaintiff's motion to recertify the class, explaining that individual damages issues did not require decertification. This judge appointed a special master. In an August 12, 2011 order, without having read the record and based on the special master's recommendation, the motion judge vacated the zero diminution in value jury award to allow the master to consider damages for all class members on any applicable theory of recovery.
In a published decision dated April 2, 2012, we reversed the August 12, 2011 order because the motion judge had improperly vacated the jury's finding of no diminution in value damages without first canvassing the record to determine whether that aspect of the verdict resulted in a manifest denial of justice. Little v. KIA Motors Am., Inc., 425 N.J. Super. 82, 89-91, 39 A.3d 930 (App. Div. 2012). Further, the motion judge's decision was inconsistent with the law of the case doctrine, since the trial judge's decision on the limited new trial had controlled the proceedings for nearly three years. Id. at 93, 39 A.3d 930.
*418On remand, the motion judge appointed a new special master to adjudicate the claims. In August 2013, she accepted the new special master's finding that only 150 claimants had proven their damages, and his recommendation of a total award of $46,197. Little was not among the members for whom he recommended recovery.
*506In January 2015, class counsel requested an award of $6,055,916 in attorney fees and $481,850 in costs of suit, with pre- and post-judgment interest, pursuant to the MMWA. After reducing the class's attorney fee award based on the paucity of damages it recovered, on May 6, 2015, a new motion judge ordered defendant to pay: $200,000 for the class's attorney fees, plus $19,113 in prejudgment interest; $481,850 in fees and costs of suit; and $5000 to Little as an incentive award.
II.
At trial, plaintiff demonstrated a defect in the Sephia's brakes. Defendant began selling the Sephia in New Jersey in 1997. Raymond King, plaintiff's expert in mechanical engineering and repair damages, explained that when a driver presses the brakes, hydraulic pressure forces brake fluid into a brake caliper, which causes the brake pads to squeeze against the rotors and decrease the spinning of the wheel. The pressure of the brake pads against the rotors causes friction, which produces heat. The hotter the brake system becomes, the faster the brake pads and rotors wear.
Based on the documents from defendant that King had reviewed, as well as deposition testimony from defendant executives, King concluded that the Sephia's front brake system had a systemic design defect that did not allow for the proper dissipation of heat. This defect caused a premature wear of the brake pads, pulsating or grinding brakes, warped or prematurely worn rotors, and shaking or vibration (also called shudder or judder) when the driver applied the brakes. Repairs or replacement of the brake pads and rotors failed to correct the problem.
*419To reach this conclusion, King reviewed a standardized industry report; Quality Assurance Field Product Reports and District Parts and Service Manager Reports, drafted by defendant's mechanics and managers throughout the United States; defendant's Technical Assistance Center Incident Reports; Technical Service Bulletins; and defendant's warranty brake claims data.
The parties stipulated that from 1997 to 2000 a total of 8455 Sephias were sold in New Jersey. Defendant's warranty repair data showed that the total number of warranty repairs to front brake components on the Sephia in New Jersey was about 8400. Defendant sold 42,713 model year 1997 Sephias in the United States. The warranty claim rate nationally for that model's brakes was 92%. King testified that he had never before seen a warranty claim rate that high. In his view, it "screamed" that there was a problem with the brake system. The following years the model had similarly high claim rates.
In January 2002, Kia Motors Corporation (Kia Motors), KMA's parent company based in South Korea, issued a technical services bulletin introducing newly designed brake pads and rotors, known as the "field fix." The updated pads were not compatible with the original rotors; thus, both had to be replaced as a set. This was an improvement, King said, but it failed to meet the 20,000-mile standard. At most, the field fix brake pads lasted 14,000 to 15,000 miles.
In addition to reviewing Kia Motors' documents, King inspected the cars belonging to Little and Samuel-Basset (the named plaintiff in a Pennsylvania class action against defendant, Samuel-Bassett v. Kia Motors Am., Inc., 613 Pa. 371, 34 A.3d 1 (2011) ). King found nothing remarkable about either car in general, or the brake system in particular, that would have caused premature brake pad and rotor wear, and nothing to suggest that driving *507habits had caused the premature brake wear.
After surveying five Kia dealerships, King estimated that an owner would spend about $250 for a brake repair. Defendant's documents showed brake replacements when cars had as little as *4202000 miles, and others at more than 10,000 miles. On average, a Sephia would need a brake replacement every 10,000 miles. In King's experience, and based on industry data he reviewed, cars typically lasted 100,000 miles, or seven to eight years.
Based on a life of 100,000 miles, and the need for a brake repair every 10,000 miles, King estimated an owner would incur ten brake repairs over the life of the car, doubling the normal repair expense due to the defective brake system. As a result, the owner would incur $1250 in additional repair expenses (five times $250) due to the defective brake system.
On cross-examination, King conceded that the $1250 brake repair costs would not apply to someone who had brake replacements at 20,000-mile-or-more intervals, or to someone who had each brake replacement paid under warranty. He also admitted that his damages model did not conform exactly to Little's experience.
Little testified that in January or February 1999, she purchased a new Sephia for $13,288. Her constant brake problems began within two weeks. She testified that for the three years she owned the car, a set of brakes lasted no more than six to seven months.
Plaintiff read into the record a portion of the deposition testimony of several individuals, including defendant KMA's Director of Technical Operations, Timothy McCurdy, who testified that defendant had been aware of the brake issue based on the rate of repairs, and that it had taken steps to address it by relaying the complaints to Kia Motors.
A "major cause" of these problems was improper dissipation of heat. While there was no set standard for the life of brakes, McCurdy said that consumers typically expected them to last 20,000 miles. One study from Kia Motors reported that the 1999 Sephia had a brake pad life of 16,000 miles. Defendant notified Kia Motors that 16,000 miles was not acceptable, since brake pads should last at least 20,000 miles.
*421Defendant did not cover the brake pads under warranty, but it did cover defects in the brake system under the three-year or 36,000 mile warranty. In model year 2002, Kia Motors replaced the Sephia with the Kia Spectra. The Spectra was "the same basic car," but with a different brake system. None of Kia Motors' vehicles, including the Spectra, had brake repair rates as high as the Sephia's.
Kia Motors Deputy General Manager Young Sun Sohn's deposition testimony revealed that when Kia Motors developed the Sephia, the specification for the brake pads was that they achieve a life of 20,000 kilometers, or just under 12,500 miles.
Lee Sawyer, defendant's Senior Vice President of Fixed Operations, testified at deposition that some Sephias had brakes that wore prematurely. Typically, brake pads lasted 20,000 to 25,000 miles before needing replacement. "Some of the Sephia owners were experiencing brake pad life in the [ten] to [twelve thousand] mile range."
Defendant became aware of the Sephia brake problem within the first year of sales based on warranty claims and brake pad orders from the parts department. After the first year, defendant also noticed an increase in part orders for rotors, which usually last 50,000 to 75,000 miles.
*508While defendant's policy was to exclude brakes from the warranty, some dealers covered brake pad replacements as warranty repairs or as goodwill repairs, both at no charge to the owner. Dealers did this because they knew that there were problems with the brakes.
Sawyer said that McCurdy had an engineer investigate the brake issue and send a report to Kia Motors' headquarters. At some point, a South Korean engineer met with someone at KMA and said the brakes had to be redesigned with better quality material.
Michelle Cameron, defendant's Manager for Consumer Affairs, testified that people who answered complaints through defendant's *422call center were trained to notify callers that brakes were not covered under the warranty.
Plaintiff presented expert testimony from John Matthews, a professor at the University of Wisconsin School of Business, on diminution damages. In Matthews's opinion, Sephia owners paid about $2000 more for their Sephia than the car was worth as a result of the defective brake system.
Matthews computed the diminution in value based on Sephia's value retention at the time of resale. The 1998 model's initial sale price was $10,000, but it retained only forty percent of that value, or $4000, while comparable cars retained fifty percent of value, or $5000, at resale. Matthews concluded that double the actual resale value, or $8000, was thus the true value of the Sephia at the time of purchase.
Matthews testified that the diminution in value was a result of the defective brake system. Not only were owners aware of the problems, but defendant's dealers were also aware of the problems, based on the Technical Service Bulletins that Kia Motors had distributed. This knowledge drove down the price that people were willing to pay for a used Sephia.
On cross-examination, Matthews said that the diminution in value was not dependent on the number of brake repairs the car had, but rather, on the market perception of a car with a faulty brake system.
III.
The defense did not deny the Sephia had brake problems. Donald Pearce, defendant's Vice President of Parts and Service, testified that a database recorded the following dates of brake repairs to Little's car as warranty repairs: September 1999, April 2000, and June 2000. He said defendant had difficulty addressing the brake system complaints because the complaints differed: some related to noise, others had to do with judder, and some related to premature wear. Based on data he had seen, Pearce *423said that $250 was a reasonable charge to replace brake pads and rotors.
Larry Douglas Petersen, defendant's expert in auto engineering and design and warranty data analysis, testified that he was not aware of any industry standard or expectation for the life of brake pads. He believed the rate at which brake pads wore was dependent on environmental conditions, driving habits, type and size of the car, and design and construction of the brake pads and brake system.
Petersen did not believe that the problem with the Sephia brake system was due to heat and the system's inability to dissipate it. Instead, he opined that the problem was the result of low-quality rotors provided by the vendor. He based this primarily on the warranty data he had seen, which suggested that the problem related primarily to the rotors.
*509On cross-examination, Petersen said that McCurdy's goal was that brake pads would last 20,000 miles. Petersen was not aware of any test where the Sephia brake pads achieved that goal.
Petersen testified that defendant did not provide him, nor did he request, information on brake repairs not covered under warranty. He believed, however, the dealerships would have had that information. With respect to recalls, Petersen believed auto companies only issued them for safety concerns. He also testified that if a car had a design defect, all owners would experience the problem.
Bruce Strombom, defendant's expert in statistics, economic analysis, loss causation, and damage calculation, disagreed with Matthews's conclusion that the faulty brake system caused excessive depreciation or diminution in value. Strombom's opinion was that Matthews's overpayment formula failed to account for the difference in purchase price and length of ownership, and the comparison groups that he used resulted in an overstatement of damages. Matthews also failed to account for other explanations for the Sephia's increased rate of depreciation, such as decreased *424purchase price, problems with the fuel pump and seatbelt, and low quality-rating score. Thus, a major flaw in Matthews's analysis was that he failed to establish a link between depreciation and the brakes.
Defendant also called as witnesses three Sephia owners who had opted out of the class because they were all satisfied with their cars and had experienced no problems with the brakes.
IV.
The jury returned a verdict finding that defendant had breached its express warranty, the implied warranty of merchantability, and the MMWA. In answering the question, "Did the class sustain damages?" the jury answered "yes." The verdict sheet then asked the jury to specify the amount of damages that each class member had incurred for "the difference in value, if any, of the Sephia as warranted compared to the Sephia as delivered," and the amount for "repair expenses reasonably incurred as a result of defendant's breach of warranty." The jury answered zero for diminution in value, and $750 for repair expenses.
In a November 2008 written decision, the trial judge found the evidence supported the jury's finding of liability, but that she had erred in submitting the question of repair damages to the jury because those damages were dependent upon individual factors, and thus, could not be awarded on a class-wide basis. She granted defendant's motion for JNOV as to repair damages only, decertified the class for purposes of damages, and ordered a new trial on repair damages to proceed by way of claim forms.
The trial judge cited Kyriazi v. Western Electric Co., 647 F.2d 388, 392 (3d Cir. 1981), in ordering the claim-form proceeding. Kyriazi is a class action sex-discrimination employment case, where the trial court found in the liability phase of the litigation that the employer had a policy of discrimination against women. Id. at 390. In the damages phase, the court ordered the class members to submit claim forms to a special master who would presume each claim valid and consider any employer challenges at *425a hearing. Id. at 390-91. Here, after significant motion practice over the exact parameters of the claim-form process, the result was a hugely reduced total damages amount of $46,197.
V.
Plaintiff contends that the trial judge erred in granting defendant's motion for *510JNOV and vacating the $750 repair expenses award for each class member. It argues that the award was (a) consistent with the Uniform Commercial Code (UCC), N.J.S.A. 12A:2-714, which allows for a reasonable estimate of damages in a breach of warranty case; (b) consistent with breach of contract damages in a class action; and (c) supported by the evidence.
In considering a motion for JNOV, a trial court and a reviewing court apply the same standard: "[I]f, accepting as true all the evidence which supports the position of the party defending against the motion and according [it] the benefit of all inferences which can reasonably and legitimately be deduced therefrom, reasonable minds could differ, the motion must be denied ...." Boyle v. Ford Motor Co., 399 N.J. Super. 18, 40, 942 A.2d 850 (App. Div. 2008) (quoting Verdicchio v. Ricca, 179 N.J. 1, 30, 843 A.2d 1042 (2004) ).
Pursuant to Rule 4:49-1(a), a trial court may grant a new trial "as to all or part of the issues" decided at trial. "The trial judge shall grant the motion if, having given due regard to the opportunity of the jury to pass upon the credibility of the witnesses, it clearly and convincingly appears that there was a miscarriage of justice under the law."R. 4:49-1(a).
A jury verdict is entitled to considerable deference and "should not be overthrown except upon the basis of a carefully reasoned and factually supported (and articulated) determination, after canvassing the record and weighing the evidence, that the continued viability of the judgment would constitute a manifest denial of justice." That is, a motion for a new trial "should be granted only where to do otherwise would result in a miscarriage of justice shocking to the conscience of the court." In fact, in Carey v. Lovett, 132 N.J. 44, 66 [622 A.2d 1279] (1993), we expressly stated that a "trial court should not disturb the amount of a verdict unless it constitutes a manifest injustice ...." [
*426Risko v. Thompson Muller Auto. Grp., Inc., 206 N.J. 506, 521, 20 A.3d 1123 (2011) (first quoting Baxter v. Fairmont Food Co., 74 N.J. 588, 597-98, 379 A.2d 225 (1977) ; then quoting Kulbacki v. Sobchinsky, 38 N.J. 435, 456, 185 A.2d 835 (1962) ).]
The trial judge vacated the $750 repair damages award on the ground that only diminution in value applied to the class as a whole. She believed she had erred in submitting repair damages to the jury for consideration, explaining that those damages were dependent on the actual expenses incurred by each class member. The judge said:
The jury determined that [p]laintiff had not proven a diminution in value of the Kia automobiles. Such a finding would result in damage throughout the class. The jury instead determined that class members suffered losses of $750 due to the defective brake system. This court is convinced this finding was based upon an erroneous submission by the court of the jury question and accompanying instructions. The damages suffered by each class member are dependent on numerous variables, such as brake life, frequency of repair, driving habits and length of time the car was owned. These damages cannot be ascertained on a class wide basis, and the court's decision to submit same to the jury was error.
We note that where theories of damages may impact one another, the court must order a new trial on damages as a whole and may not order a new trial on only one theory of damages. Donovan v. Port Auth. Trans-Hudson Corp., 309 N.J. Super. 340, 353, 707 A.2d 171 (App. Div. 1998) (ordering a new trial on damages as a whole where future wage loss could impact *511pain and suffering damages, even though the jury's pain and suffering award was supported by the evidence and law). Here, the judge instructed the jury it could award damages based on a diminution of value or repair expenses, or both. The judge did not tell the jury of any ramifications if only repair damages were awarded. A court may not grant a new trial on a discrete issue unless that issue is "fairly separable from the other issues" in the case. Corridon v. City of Bayonne, 129 N.J. Super. 393, 398, 324 A.2d 42 (App. Div. 1974). As explained to the jury, the two types of damages were not fairly separable from each other.
On appeal, plaintiff contends the judge erred in finding that repair damages could not be awarded on a class-wide basis. All class members purchased a car with the same faulty brake system, which, according to defendant's records, required a brake repair *427about every 10,000 miles. As King testified, based on the average life of a car (100,000 miles) and the undisputed cost of a brake repair ($250), one could accurately estimate each class member's average repair damages resulting from defective brakes.
Plaintiff contends that King's method of estimating damages was consistent with the UCC standard for determining contract damages, as that standard allows for the computation of damages based on any reasonable method that places the plaintiff in the position he or she would have been in had the defendant not breached the contract, or warranty, as in this case. See N.J.S.A. 12A:2-714.
Defendant argues that proof of a product defect does not equate to proof of damages, and it disputes that New Jersey law allows for the averaging of damages. It argues that the class must prove damages for each individual class member and the relatively few claims submitted during the claim-form process show that King's estimate was incorrect and was nothing more than a net opinion, unsupported by fact.
Because the jury found the claims for breach of express and implied warranties were established, the UCC provides the starting point for assessing damages. N.J.S.A. 12A:2-313 (express warranties); N.J.S.A. 2A:2-314 (implied warranty of merchantability); N.J.S.A. 2A:2-315 (implied warranty of fitness for a particular purpose); N.J.S.A. 12A:2-714 (buyer's damages for breach of accepted goods); N.J.S.A. 12A:2-719 (modification of remedy).
The trial judge's interpretation of those statutes as applied to this case are entitled to no deference, as they entail matters of law. Manalapan Realty, L.P. v. Twp. Comm. of Manalapan, 140 N.J. 366, 378, 658 A.2d 1230 (1995) ("A trial court's interpretation of the law and the legal consequences that flow from established facts are not entitled to any special deference.").
The jury in this case found no diminution in value damages. In most breach of warranty cases, diminution in value, or the difference between the goods as delivered and as warranted, *428provides the proper measure of damages. Perth Amboy Iron Works, Inc. v. Am. Home Assurance Co., 226 N.J. Super. 200, 219, 543 A.2d 1020 (App. Div. 1988). However, repair damages may be appropriate, depending on the facts of the case. Ibid.; accord Furst v. Einstein Moomjy, Inc., 182 N.J. 1, 13, 860 A.2d 435 (2004) (applying UCC principles to a consumer fraud case and concluding that the cost of replacing a damaged carpet was the appropriate measure of damages, as that method put the buyer in the position he would have been in if he had received a non-defective carpet). Diminished resale value or the cost of *512a retrofit repair may also provide a reasonable measure of damages. In re GMC Pick-Up Truck Fuel Tank Prods. Liab. Litig., 55 F.3d 768, 816-17 (3d Cir. 1995) (emphasizing, in a class action suit, that section 2-714(1) of the UCC allows for damages "as determined in any manner which is reasonable").
Our Supreme Court explained that breach of contract compensatory damages are intended to put the injured party in the position he or she would have been in had the goods been delivered as promised. 525 Main St. Corp. v. Eagle Roofing Co., 34 N.J. 251, 254, 168 A.2d 33 (1961). "Although specific rules are formulated for sundry situations, they are subordinate to this broad purpose"; thus, "a given formula is improvidently invoked if it defeats a common sense solution" to computing damages. Ibid.
Breach of contract damages need not be established with exact certainty:
[M]ere uncertainty as to the quantum of damages is an insufficient basis on which to deny the non-breaching party relief. Although it complicates the precise calculation of damages, our courts have long held that "[p]roof of damages need not be done with exactitude. ... It is therefore sufficient that the plaintiff prove damages with such certainty as the nature of the case may permit, laying a foundation which will enable the trier of the facts to make a fair and reasonable estimate."
[ Totaro, Duffy, Cannova and Co., LLC v. Lane, Middleton & Co., LLC, 191 N.J. 1, 14, 921 A.2d 1100 (2007) (quoting Lane v. Oil Delivery Inc., 216 N.J. Super. 413, 420, 524 A.2d 405 (App. Div. 1987) ).]
As explained in the Restatement (Second) of Contracts: Alternative to Loss in Value of Performance § 348 (Am.
*429Law Inst. 1981), a small windfall to the injured party based on an inability to prove exact damages should not defeat recovery. Also, the injured party need not prove that he or she actually spent the money to repair the defect in order to recover for the breach. Cox v. Sears Roebuck & Co., 138 N.J. 2, 22, 647 A.2d 454 (1994).
Those principles support the class's argument that breach of contract damages are not limited to the actual out-of-pocket expenses that each class member incurred, but rather, are proper if based on any reasonable method that places the class members in the position they would have been in if KMA had provided a car free of a defective brake system. King and Matthews provided two ways the jury could reasonably compute damages. King's formula estimated the additional repair expenses the class would incur as a result of the defective brake system, while Matthews's formula estimated the diminution in value between the Sephia as warranted and the Sephia as delivered.
A net opinion is an expert opinion that is not supported by facts and data. Polzo v. Cty. of Essex, 196 N.J. 569, 583, 960 A.2d 375 (2008) (discussing the requirements of N.J.R.E. 703 ). King's opinion, which is the only one that defendant claims on appeal is a net opinion, was based on the average life of the car (100,000 miles), the expected life of brake pads (20,000 miles), the average cost of a brake repair ($250), and defendant's warranty repair data. King's trial testimony provided examples of the claims, which supported his conclusion that, on average, the Sephia needed a brake repair every 10,000 miles.
Defendant did not dispute that the average life of a car was 100,000 miles, and it conceded the average cost of a brake repair ($250) and the expected life of brake *513pads (20,000 miles). While it did not concede that the Sephia needed a brake repair every 10,000 miles, it provided no evidence to challenge that figure, other than the testimony of three class members who had opted out of the class because they had experienced no brake problems with their cars. The experience that those three owners had, out of *430an 8455-person class, does not outweigh the voluminous evidence that supported King's 10,000-mile brake-repair average.
It is worth noting that during the jury charge conference, the trial judge considered the propriety of the class's two theories of damages for a second time. In disagreeing with an argument raised by defendant that the jury could award diminution in value damages, or repair damages, but not both, the judge said that the jury was free to determine damages in any reasonable manner to make the aggrieved party whole. The jury could find both that the class members overpaid for their cars, and that they incurred additional repair expenses as a result of the faulty brake system. Both theories of damages presented by the class were proper considerations for the jury.
In ruling JNOV that class-wide damages could not be calculated pursuant to King's model, the trial judge relied primarily on our decision in Muise v. GPU, Inc., 371 N.J. Super. 13, 851 A.2d 799 (App. Div. 2004). In Muise, two groups of GPU customers filed suit against GPU after experiencing power outages during a heat wave. Id. at 18-20, 851 A.2d 799. The plaintiffs sought damages for the loss of power based on consumer fraud, negligence and breach of contract. Id. at 18, 851 A.2d 799. The trial court initially granted class certification to all New Jersey customers of GPU, and about 2500 customers submitted claims ranging in value from $1 to $150,000. Id. at 23, 851 A.2d 799.
Only the negligence claim survived summary judgment. Id. at 20, 851 A.2d 799. GPU moved to decertify the class, and the class moved for a declaration of admissibility of their experts' survey-based method of computing damages for the class. Ibid. The class-wide damages model valued hypothetical power outages based on surveys of electrical customers in California and Canada. Id. at 24, 851 A.2d 799. The experts "adapted this data to New Jersey in order to 'estimate customer damages resulting from GPU service interruptions from July 4 to 10, 1999.' " Ibid. (quoting the experts' report).
*431The experts recognized that their survey model of damages was limited, and admitted that their "results were predictions and estimations based on survey scenarios" that reflected what some class members had experienced. Ibid. The exact cost of power loss was dependent upon socioeconomic, demographic and geographic factors. Id. at 24-25, 851 A.2d 799.
The trial court concluded that the class-wide model of damages was not reliable, as it was based on hypothetical estimates that did not reflect actual damages in the case. Id. at 28, 851 A.2d 799. The trial judge recognized that individualized proofs of damages were not necessary so long as the class-wide damages model was based on a reliable mathematical formula. Id. at 47, 851 A.2d 799. This formula, however, was hypothetical and unreliable. Ibid. Further, the class had failed to prove actual damage, or harm as a proximate cause of power loss; experiencing a break in power did not necessarily result in actual harm. Ibid.
We agreed. As the trial court had recognized, proof of individual damages was the norm, and a court should depart from that norm only if (a) there was proof, or at least a presumption, that the class members sustained damage, and (b) there was a *514reliable mathematical formula to compute the estimated or aggregated damages. Ibid. (citing In re Prudential Ins. Co. of Am. Sales Practices Litig., 962 F.Supp. 450, 517 (D.N.J. 1997), aff'd in part, vacated in part on other grounds, 148 F.3d 283 (3d Cir. 1998), where "[t]he court acknowledged that the possible methods of proving the amount of damages included expert testimony estimating aggregate damages sustained by the class, or a formula to be applied to individual class members" so long as "sufficient facts [are] introduced so that a court can arrive at an intelligent estimate without speculation or conjecture").
While we agreed with the trial court that class certification was not appropriate for all New Jersey customers, we concluded that certification was appropriate for a more limited class, namely, customers in Red Bank "whose outages directly resulted from the alleged negligence in delaying the replacement of the transformers *432in the Red Bank substation." Id. at 64, 851 A.2d 799. We recognized that individual damages issues might still exist with respect to the limited class, but concluded that those damages issues could be addressed through a number of techniques, including claim forms, interrogatories, requests for admissions, and "statistical interpretation of sampling data from the relevant universe, established based on competent data." Ibid.
Here, King based his class-wide model for damages on: actual brake repairs that Sephia owners had experienced, as evidenced by defendant's warranty repair documents; studies that defendant produced on the duration of brake pads and rotors in the Sephia; reports on defendant's efforts to improve the duration of the Sephia brake system; and testimony from defendant's executives on the brake system. As the Muise decision makes clear, so long as a plaintiff-class establishes proof of damage, or at least a pretrial presumption of damage, the class need not prove individual damage, but may instead present class-wide average damages based on a reliable mathematical formula.
The trial judge did not have the advantage of reviewing the 2011 Pennsylvania Supreme Court opinion, supporting the propriety of the verdict in that sister-case. Samuel-Bassett, 34 A.3d at 11-13. King also testified as the plaintiff's expert in Pennsylvania. The Samuel-Bassett jury found that KMA had breached its express and implied warranties, and awarded each of the 9400 Pennsylvania class members $600 (about $5.6 million for the class) in repair damages. Id. at 13. After the trial court denied KMA's motion for JNOV, it awarded the class $4,125,000 in counsel fees and $267,513 in costs of suit. Ibid.
In finding that KMA's challenges lacked merit, the Samuel-Bassett majority began by discussing the facts that established class-wide liability. Id. at 35-36. The majority said that deposition testimony from the same KMA executives who testified or whose deposition testimony was presented in this case (McCurdy, Sawyer, Pearce and Sohn), along with KMA's internal documents, established that all Sephias had the same defective front brake *433system that did not allow for proper dissipation of heat. Id. at 35-36. This resulted in premature brake pad wear and warping of the rotors, which required excessive repairs. Id. at 36.
The majority found that the class had "adduced sufficient evidence to prove" that each member suffered damages. Id. at 37. Two of KMA's executives (Pearce and Cameron) had testified that while KMA had paid for some repairs under warranty, a coupon program and good will replacements, KMA's policy was to exclude brake repairs from the warranty as wear-and-tear items. Ibid."As a result [Sephia] owners *515sustained out-of-pocket repair costs estimated by [King] at approximately $1005 over the life of their Kia Sephia." Ibid.
King had "estimated that each vehicle underwent five extra repairs in addition to wear-and-tear replacements of brake pads and rotors." Ibid. The majority said:
This calculation, of course, does not account for factors such as: whether class members owned their vehicles over 100,000 miles, whether each class member experienced exactly five additional repairs, and whether any additional repairs were covered under warranty. Indeed, warranty data introduced at trial reflected that KMA covered some of the brake component replacements under good will and brake coupon programs, which suggested that a number of the estimated repairs for the class did not in fact cause class members out-of-pocket expenses.
[ Ibid. ]
The majority noted that KMA's defense had centered on undermining the allegation of a defective system and questioning whether the alleged defect had affected each owner in the class, particularly since brake wear depended upon many variables. Id. at 37.
The majority found unpersuasive the one dissenting judge's criticism that the class's aggregate approach to damages "blur[ed] the substantive requirements of the law of damages," which generally required proof of individual damages. Id. at 40 (citing Scottsdale Mem'l Health Sys., Inc. v. Maricopa Cty., 224 Ariz. 125, 228 P.3d 117, 133 (App. 2010) (rejecting an argument that calculating damages based on a statistical sample violated due process); In re Pharm. Indus. Average Wholesale Price Litig., 582 F.3d 156, 197-99 (1st Cir. 2009) (holding: "Aggregate computation of class *434monetary relief is lawful and proper. Courts have not required absolute precision as to damages."); Hilao v. Estate of Marcos, 103 F.3d 767, 784-86 (9th Cir. 1996) (finding no due process violation in a formula that aggregated damages based on a sample) ).
The majority noted that the United States Supreme Court had also rejected the notion that a jury may not estimate damages. Ibid. (quoting Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 124, 89 S.Ct. 1562, 23 L.Ed.2d 129 (1969), for the proposition that "[a]lthough the factfinder is not entitled to base a judgment on speculation or guesswork, the jury may make a just and reasonable estimate of the damage based on relevant data, and render its verdict accordingly"). Thus, the majority affirmed the trial court's decision denying KMA's motion for JNOV. Ibid.
Samuel-Bassett majority's decision is consistent with New Jersey law on breach of contract damages in a class action. King's testimony here provided a reasonable basis for an aggregation of damages award, and the jury accepted that basis in computing damages. It did not award the full amount of damages that King had estimated ($1250), but instead decreased the amount by $500, or two brake repairs. That decision was also supported by the record, which showed that defendant had provided some repairs under warranty or goodwill, and it had increased the life of the brake pads to 14,000 or 15,000 miles with the field fix. Thus, the jury could reasonably have decreased King's estimate to reflect those two factors.
The trial judge's concern that the jury award would provide a windfall to class members who did not actually spend $750 in additional repair costs was inconsistent with principles of contract damages, and resulted in an unjust windfall to defendant, *516which produced no evidence to defeat King's estimated damages.
The 2016 United States Supreme Court decision in Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. ----, 136 S.Ct. 1036, 194 L.Ed.2d 124 (2016), also supports the conclusion that King's aggregate method of computing damages was appropriate. The Tyson Foods decision primarily addresses whether a class may *435prove liability for failure to adequately compensate employees based on statistical data of average uncompensated work time.
In Tyson Foods, a class of 3344 employees filed suit against their employer, Tyson Foods, claiming that it had failed to pay them for time spent changing into and out of a safety protective suit needed for butchering. Id. at 1041-43. The class claimed entitlement to overtime pay pursuant to the Fair Labor Standards Act, 29 U.S.C. § 207(a). Id. at 1042.
Because the employees sought overtime pay, each member had to establish that he or she had worked more than forty hours per week, inclusive of the time spent changing into and out of the protective gear, in order to establish a claim. Id. at 1043. Due to Tyson Foods' lack of records on the amount of time the employees spent changing, the plaintiffs relied on what they called "representative evidence" of the average time it took an employee to don and doff the gear. Id. at 1043.
This evidence included employee testimony, video recordings of donning and doffing at the plant, and, most important, a study performed by an industrial relations expert, Dr. Kenneth Mericle. Mericle conducted 744 videotaped observations and analyzed how long various donning and doffing activities took. He then averaged the time taken in the observations to produce an estimate of 18 minutes a day for the cut and retrim departments and 21.25 minutes for the kill department.
[ Ibid. ]
An employees' expert's testimony supported an aggregate award of about $6.7 million in unpaid wages; however, the jury awarded the class only $2.9 million. Ibid. Tyson Foods moved to set aside the verdict claiming that the district court had erred in certifying the class based on the variation in donning and doffing time. Ibid.
On appeal to the Supreme Court, Tyson Foods first argued that the trial court had erred in certifying the class "because the primary method of proving injury assumed each employee [had] spent the same time donning and doffing protective gear, even though differences in composition of that gear may have meant that, in fact, employees [had taken] different amounts of time to don and doff." Id. at 1041. Tyson Foods claimed that the class's *436method of averaging was an inherently unfair way to (1) avoid the differences that made the group inappropriate for class certification, and (2) relieve the employees of proving individual damages. Id. at 1046. Tyson Foods requested the Court to announce a broad rule prohibiting the type of representative evidence relied upon by the class to establish liability. Ibid.
In rejecting Tyson Foods' request for a broad rule excluding representative evidence to establish liability, the Court said:
A categorical exclusion of that sort, however, would make little sense. A representative or statistical sample, like all evidence, is a means to establish or defend against liability. Its permissibility turns not on the form a proceeding takes - be it a class or individual action - but on the degree to which the evidence is reliable in proving or disproving the elements of the relevant cause of action.
*517[ Ibid. ]
Whether statistical evidence can establish liability in a class action depends on the purpose of the evidence and the underlying cause of action. Ibid.
The Court underscored that the employees had used the representative evidence to "fill an evidentiary gap created by the employer's failure to keep adequate records." Id. at 1047. Once the district court found the evidence admissible, "its persuasiveness [was], in general, a matter for the jury." Id. at 1049.
Defendant here, like Tyson Foods, failed to present evidence to rebut the class's formula for computing aggregate damages. Defendant had access to documents that would have established the actual rate of brake repairs by defendant's dealers. But the only documents on repairs that they produced related to warranty repairs. We see no reason to disturb the jury's considered determination of class damages.
VI.
On cross-appeal, defendant contends that the court erred in charging the jury on breach of the implied warranty of merchantability. In charging the jury, the court must "set forth in clearly understandable language the law that applies to the issues *437in the case." Toto v. Ensuar, 196 N.J. 134, 144, 952 A.2d 463 (2008). The charge serves as "a road map that explains the applicable legal principles, outlines the jury's function, and spells out 'how the jury should apply the legal principles charged to the facts of the case.' " Ibid. (quoting Viscik v. Fowler Equip. Co., 173 N.J. 1, 18, 800 A.2d 826 (2002) ).
We review the charge as a whole to determine whether it correctly sets forth the law as applied to the facts. Ibid. With respect to the implied warranty of merchantability, the trial judge charged the jury:
An implied warranty of merchantability is a promise by the sellers that the goods, in this case, Mrs. Little's and the class vehicles, were fit for their ordinary purpose for which automobiles are used.
The implied warranty of merchantability does not require that the goods, in this case the vehicles, be defect free, but the implied warranty of merchantability, provides that [ ] K[M]A warrants the cars that are fit for their ordinary purposes for which cars are used.
Thus, to prove a breach of implied warranty, the [p]laintiff must show an implied promise that the vehicles were fit for their ordinary purpose for which a vehicle is used.
Two[,] that there was a defect that substantially impaired the value or use of the car.
Three, the defect caused a lost [sic] to the purchaser.
And, four, damages with regard to the instructions I will give you.
Defendant contends the charge was erroneous because the correct standard was not whether the Sephia was fit for the ordinary purposes for which cars are used, but rather, whether the cars provided safe and reliable transportation. Defendant contends that it did not breach the implied warranty of merchantability because Little conceded that her car provided safe transportation, and she drove her car 75,000 miles prior to surrendering it.
The charge was proper and consistent with N.J.S.A. 12A:2-314(1)(c), which provides that goods are merchantable if they "are fit for the ordinary purposes for which such goods are used." See also Model Jury Charge (Civil), 4.22, "Breach of Implied Warranty of Fitness for Particular Purpose Under UCC" (1984) (explaining *438*518in the commentary that when the issue does not deal with fitness for a particular purpose, the standard is one of fitness for the ordinary purpose). The implied warranty standard was not limited to whether the Sephia was safe and reliable, but rather, whether it was fit for the ordinary purpose for which cars are used. A driver would fairly assume that an ordinary car would have a brake system that does not cause premature wear of brake pads and rotors. The court correctly charged the jury on the implied warranty of merchantability.
VII.
The class was certified in 2003 based on Rule 4:32-1(a) and (b)(3). Defendant contends that the court erred in denying its many motions to decertify the class because Little's experience was not typical of the class, individual issues predominated, and the vast majority of the class members suffered no damages as demonstrated by the claims form process. In order to certify a class, the court must find that (1) the large number of members makes joinder impractical, (2) the members raise common questions of law or fact, (3) the representative's claims or defenses are typical, and (4) the representative will fairly and adequately protect the interests of the group. R. 4:32-1(a).
The court must also find that the group has satisfied at least one of the three requirements set forth in Rule 4:32-1(b). Rule 4:32-1(b)(3) requires a finding that common questions of law or fact predominate, and that a class action is superior to other methods of resolution.
Defendant first argues that the court erred in denying its request for decertification because Little's experience was not typical of the class. Defendant relies on (1) King's testimony that Little's repairs did not fit exactly within the average repairs his model computed, and (2) the testimony of the three Sephia owners who had opted out of the class because they had experienced no problems with their cars.
*439This argument lacks merit because a class representative need not establish her experience was exactly the same as every class members' in order to establish her claims are typical of other members' claims. Like the class members, Little purchased the same type of car with the same defective brake system that required repairs about every 10,000 miles. To require more similarity than that would defeat the purpose of class actions.
As our Supreme Court said in Iliadis v. Wal-Mart Stores, Inc., 191 N.J. 88, 104, 922 A.2d 710 (2007) :
When making certification determinations, "the best policy" is to interpret the class-action rule "so as to promote the purposes underlying the rule." 5 James W. Moore et al., Moore's Federal Practice, § 23.03 (3d ed. 1997). Unitary adjudication through class litigation furthers numerous practical purposes, including judicial economy, cost-effectiveness, convenience, consistent treatment of class members, protection of defendants from inconsistent obligations, and allocation of litigation costs among numerous, similarly-situated litigants.
The class action in New Jersey also helps to equalize adversaries, a purpose that is even more compelling when the proposed class consists of people with small claims. In such disputes, where the claims are, in isolation, "too small ... to warrant recourse to litigation," the class-action device equalizes the claimants' ability to zealously advocate their positions. [ In re Cadillac V8-6-4 Class Action, 93 N.J. 412, 435, 461 A.2d 736 (1983) ]. ... In short, the class action's equalization function opens the courthouse doors for those who cannot enter alone.
*519Class certification furthered those purposes here where each class members' damages were relatively small.
Finally, defendant contends that the court should have decertified the class during the claim-form proceeding because repairs were unique to each member and the majority of the class members did not fit the claim-form criteria. This argument lacks merit because the claim-form proceeding should not have occurred. Further, failing to return a claim form does not prove that the class member incurred no damage. The court correctly certified the class. The cross-appeal is without merit.
We reverse the trial judge's grant of a JNOV and remand for a determination of counsel fees consistent with this opinion. We do not retain jurisdiction.