A SUGGESTION FOR SABERMATRICIANS—STANDARDIZING WAR. WAR is controversial—among sabermatricians as well as among members of the old guard. One reason is that WAR takes fielding measurements into account—and there is no agreement at all about how that should be done. Another reason is that there is no agreement on how “replacement player” should be determined. This article by Dave Cameron on Fangraphs describes the process and links to this article by Seth Smith in The Hardball Times which elaborates on the calculations. Smith’s article indicates the assumptions required. A rough estimate is made of the average number of runs per year a replacement level player costs a team (20 per year below league average? or should it differ by position?). An effort is made to identify the 61st best catcher….
Since WAR is used to compare differences between player’s performances, the difference should be measured against a fixed, objective standard. (That is, if we are comparing how much better Miguel Cabrera does than a standard player to how much better Mike Trout does than a standard player, the standard player should be readily and objectively determined). Here are two standards which would meet these criteria by position: for a position—say third base—calculate the performance of all the third basemen for each team for a year. The 15th/16 th team gives you a standard for a major league average third baseman in a given year. The 30st team gives you a standard for a replacement level third baseman. If the team scrambled for a third baseman, their third base position statistics will probably include the performance of a number of marginal players. It is easier to determine the statistics for the marginal major league third baseman than for the marginal minor league third baseman.