While discussing this year’s Rookie of the Year Award Voting in a post titled “Golden years,” I made use of Delta Graphs WAR figures, because A) their serious and objective, and B) I don’t have anything better to replace them with since I haven’t done Win Shares for a few years.
In that article I explain how, in my opinion, voters botched one of the awards that went to Hiroshima Carp closer Ryoji Kuribayashi and Orix Buffaloes starting pitcher Hiroya Miyagi.
I feel it is now necessary to cast doubt, not on WAR’s validity–I think it is likely very good at comparing players in the same league who play the same position.
But I would like to hear a good argument for the validity of a system that determined in 2021, that pitchers contributed 58.2 percent of the value above replacement in NPB, with fielding, batting and base running left to fight for the leftover 41.8 percent.
This is, from my point of view, is ass-backward.
The dominant factor in deciding any plate appearance is not the pitcher, but the hitter. No pitcher strikes out batters as often as Teruaki Sato strikes out at the plate. No pitcher in Japan allows batters to make contact as often as Japan’s best slap hitters do.
I haven’t studied the issue recently, but when evaluating players contributions consumed my life from 1992 to 1997, I calculated that along a broad range of plate appearance outcomes, the standard deviations among batters were about 38 percent larger than among pitchers.
If pitching were the dominant actors in every plate appearance, it would be the other way around, and the WAR totals produced by Delta Graphs lump the batting and defense together, making it even more one-sided.
For that reason, my principle issue is not the numbers are invalid, but that batting WAR and pitching WAR are apples and oranges, because the total sums simply don’t add up.
2021 NPB rookies above 0 WAR
Position players (min 100 PA)
Pitchers (min 30 IP)