An ongoing and interesting discussion is how much a single player is worth to his team. The current standard for the statistical minded baseball community is to measure a players value in wins above replacement level (WAR). For example a player with a WAR of 5.0 contributes with five (5) more wins in a 162 game season than a replacement level player. More about WAR and replacement level can be found at FanGraphs, one of the best online resources for baseball.
Mike Trout is pretty much the complete baseball player and one of the brightest stars in MLB at the moment. If nothing radical happens he is going to be an elite performer for the Angels for many years to come. I decided to test how much Mike Trout is worth for the Los Angeles Angels by simulating a full MLB-season for the Angels with and without Mike Trout. The following setup was used for the simulation:
Each game was simulated 1000 times. Current depth-chart, projected playing time for each player and the schedule of each team was used in team construction. Each team consists of a five-man-rotation, a batting lineup of nine batters (eight + league average pitcher in games where NL-team is the home-team) and an eight man bullpen. So the Angels (and their opponents) consists of only 9 batters, whom are assumed to play in all 162 games. Two full seasons were simulated, the first one with Mike Trout and the second one with Collin Cowgill as center fielder and batting second. Collin Cowgill was chosen to replace Mike Trout because he is projected to be pretty close to replacement level and is the second in order for center field in the Angels depth chart. The results are shown below:
The results tells us that with Mike Trout in center field and batting second the Angels on average scored 64.4 more runs and allowed 1.8 more runs per season leading to an average run difference of +62.6 runs per season compared to the simulated season with Collin Cowgill player in center field and batting second. On average the Angels won 5.7 more games per season, which is pretty nicely in line with the theory of run value and wins (9 to 10 runs per win). More of runs and wins can also be found at FanGraphs.
I am going to do similar simulations in the near future and I am happy to take suggestions from readers on which teams and/or players should be tested.