Random Thoughts on WAR stats in Baseball
I've never put any weight on WAR stats and do not use them for anything. The hardest stat to determine is defensive range and that is always a part of WAR figures. Defensive range is figured by number of opportunities a player converts by assist or put out per nine innings. Of course the trouble is, a team could have the best defensive infielders in the league but they could also have a pitching staff that has a high strike out ratio. Then the defense gets less opportunities, thus a lower defensive range which is deceptive.
I know there are organizations that track defensive range by each play. They look at video and decide, for example, if the fly out that dropped in front of the center fielder would have been caught by an average center fielder. This introduces human judgement into what is an average center fielder. You also have to define what is a good jump for an outfielder. Maybe it looks like he had no chance to catch it cause he was 12 feet away, but maybe he got a bad jump on it instead.
I think defensive range has value if done with the opportunities per 9 innings method, but only over several seasons. Then using the average from at least 3 seasons. In a single season, its too easy for defensive stats to get scewed. Also, the 1st baseman gets so many put outs, that his defensive range value has little use.
In a way, any baseball sim designer that rates his players is doing a form of WAR calculations. In my version of 'True to Life Baseball' I give all players a 2 digit value ranging from 20-60. It isn't meant to be a WAR stat but could be so converted by finding the average player score, then use that as the baseline and assign a certain +/- run value to each point above or below the average player rating.
In my system I take offense, defense and base running into account for position players. Pitchers are only rated in pitching categories.
Please check out other topics in my baseball blog