¶ 22 To determine whether a party has improved its position at trial after seeking a trial de novo from an arbitration award, we must decide whether that party's pre- and posttrial positions include only the damages awarded for the claim at each proceeding, or if the party's compared positions also include the statutory costs awarded for expenses incurred during litigation. Because the amount of costs awarded to a party generally has nothing to do with the merits of the dispute, costs should *454not be considered in determining whether a party improved its position at the trial de novo; only the arbitrator's award of damages and the damages awarded at the trial de novo should be compared. I respectfully dissent.
¶ 23 I agree with the reasoning of the Court of Appeals in both of its opinions in this case, and I have little to add.1 Bearden v. McGill , 193 Wash. App. 235, 372 P.3d 138 (2016) ( Bearden I ), adhered to on remand , 197 Wash. App. 852, 391 P.3d 577 (2017) ( Bearden II ). As the Court of Appeals held, whether a party improved its position at trial should be determined by "comparing comparables." See Bearden I , 193 Wash. App. at 239, 372 P.3d 138; see also Wilkerson v. United Inv., Inc. , 62 Wash. App. 712, 717, 815 P.2d 293 (1991) ; Haley v. Highland , 142 Wash.2d 135, 154, 12 P.3d 119 (2000). We should also adhere to a rule that is easily " 'understood by ordinary people.' " Niccum v. Enquist, 175 Wash.2d 441, 452, 286 P.3d 966 (2012) (whether a party improves his or her position at trial is meant to be understood by ordinary people) (quoting Cormar, Ltd. v. Sauro , 60 Wash. App. 622, 623, 806 P.2d 253 (1991) ). Here, the Court of Appeals recognized that in our leading cases interpreting this rule, Niccum v. Enquist , id. , and Nelson v. Erickson , 186 Wash.2d 385, 377 P.3d 196 (2016), this court calculated a party's posttrial position by looking only to the damages award in the trial de novo, exclusive of costs, and correctly concluded that the party's pretrial position must be established the same way: by looking at the initial arbitration award, exclusive of costs. Bearden II , 197 Wash. App. at 854, 391 P.3d 577 (citing Nelson, 186 Wash.2d 385, 377 P.3d 196 ); Niccum , 175 Wash.2d 441, 286 P.3d 966.
¶ 24 The purpose of mandatory arbitration is to discourage "meritless appeals from arbitration as well as [to] preserv[e] superior court resources for trial of larger claims."
*455Fernandes v. Mockr i dge , 75 Wash. App. 207, 212, 877 P.2d 719 (1994). At the same time, the "rules are not intended to divert from superior court, or burden with increased risk, those claimants whose legitimate damage claims have a reasonable possibility of exceeding the threshold amount." Id. In short, the rules are intended to discourage meritless requests for trial de novo, while not unduly burdening legitimate damage claims.
¶ 25 To balance the competing goals of discouraging meritless appeals and accommodating legitimate damage claims, we should adopt a rule that is based on the merits of cases and the legitimacy of the damage claims. The best bench marks for evaluating whether a party brought a meritorious request for trial de novo are the damages awarded at the arbitration compared to the damages awarded at the trial de novo. A higher award at the trial de novo obviously means that the plaintiff improved its position and a lower award that the defendant improved its position. This easily understood and applied rule automatically balances the goals of discouraging unnecessary trials de novo while encouraging legitimate claims.
¶ 26 In contrast, if we include costs in the calculation, we introduce an element that has nothing to do with the merits or legitimacy of the claims. This case provides a classic example. Plaintiff James Bearden was awarded $44,000 in damages by the arbitrator. At the trial de novo, plaintiff was awarded less, $42,500. If we do not include costs, it is clear *106that defendants Dolphus and Kelli Knox McGill improved their position because their liability was reduced by $1,500. But when costs are included ($1,187 at arbitration and $3,296 at trial), plaintiff Bearden's total award at arbitration becomes $45,187 and at trial de novo $45,796. Defendants' position after the trial de novo changed from a net reduction of liability of $1,500 on the merits to a net increase of liability of $609. This modest $609 difference triggered an award of actual attorney fees against the McGills in the amount of $71,800.
¶ 27 The costs that caused this significant swing in outcome consist of plaintiff's filing fees, service of process *456fees, witness fees, deposition costs, medical records, police report, expert report, and a statutory attorney fee of $200. None of these costs reflects the merits of the defendants' decision to seek a trial de novo.
¶ 28 Comparing an arbitrator's award of damages to the damages award in a trial de novo employs a meaningful measure of one's position pre- and posttrial: the merits of the claim. Not only does this approach "promote [ ] simplicity," as Judge Leach remarked, Bearden II , 197 Wash. App. at 859, 391 P.3d 577, it parallels the legislative purpose of the fee-shifting rule: to discourage meritless appeals. RCW 7.06.060 ; MAR 7.3.
¶ 29 Additionally, severing the merits of a trial de novo from a party's pre-and posttrial positions by allowing costs to be a thumb on the improvement scale is not only contrary to an ordinary person's understanding,2 it is inequitable: a defendant's request for a trial de novo can be meritorious (i.e., the damage award is more favorable than the arbitrator's award) and still result in the defendant's having not improved her position, while a plaintiff's request for a trial de novo can be fruitless (i.e., the damage award is less favorable than the arbitrator's award) and still result in the plaintiff's having improved her position.
Conclusion
¶ 30 Comparing an arbitrator's award of damages to a trial de novo damage award to determine whether a party improved its position at trial emphasizes the merits of a case by employing a meaningful measure of one's position, while still discouraging meritless requests for trials de novo. Including costs in this equation departs from our precedent, promotes uncertainty, and turns a meaningful *457measure into an arbitrary and inequitable determination. Therefore, I would affirm the Court of Appeals.
¶ 31 I respectfully dissent.
Madsen, J.
I disagree with the Court of Appeals in one respect: because I believe that the award of costs does not reflect the merits of the case, I would exclude all costs from the comparison of an arbitration damages award to a trial de novo damages award.
Ordinary defendants such as the McGills would understand that their position had improved if the jury verdict was lower than the arbitration award.