Tuesday, May 5, 2009
Updated: May 8, 3:30 PM ET
Explaining the MLB Player Ratings
Who is the No. 1 player in baseball right now? The MLB Player Ratings seek to answer that question, featuring four statistical metrics for evaluating player performance. The metrics come from the following sources: ESPN, The Elias Sports Bureau, Inside Edge and The Baseball Encyclopedia.
An overall top-100 ranking is provided based on the players' average rankings among the four metrics. Thus, the player with the highest average ranking is listed as the top-rated player. In addition to viewing the overall rankings, users can view rankings for each individual metric.
Here are explanations for each of the four metrics, which are updated daily:
Click here for a full explanation of ESPN's player ratings from Jeff Bennett, ESPN's senior director of research.
Each plate appearance is evaluated on the game situation before and after the PA. The difference in the team's chance to win the game before and after is attributed to the batter/pitcher involved in the PA. The differences are accumulated over the course of the season, in the context of game situations and stadium effects. A rating is calculated on a percentile scale, relative to a typical major league "replacement player." The win probabilities are based on a computer simulation over several hundred million innings for all parks done by the noted statistician, Jeff Sagarin, whose ratings are published in USA Today and are used by the NCAA to help determine postseason participation in both football and men's basketball.
Every pitch is given a value, whether it be a first pitch taken for a strike or a 2-0 pitch driven for a double. The value placed on each pitch is derived empirically, based on how much it ultimately affects the outcome of the plate appearance. The first-pitch strike, for example, shaves off about 20 points off a batting average and 50 points off a slugging percentage (on or after that pitch) compared to a first-pitch ball. So the pitcher receives a certain amount of positive "pitch points" or "edge points" for that pitch, and the hitter receives negative points for that pitch. The bulk of the IE ratings are determined by averaging the pitch points and then normalizing them to a scale of 0-100. Pitcher and hitter points are the inverse of each other. For example, if a hitter's edge points average out to 60, the combined average for the pitchers he faced would come out to 40.
The IE rating system also rewards consistency, so players who uphold a certain standard over several months going back to 2007 receive bonus points. In addition, the ratings adjust for plate appearances so players with minimal PAs can receive ratings but not qualify for top spots until they have amassed a reasonable amount of playing time to prove they are worthy of their scores. By drilling down to the pitch level, IE's ratings boil the many aspects of performance down to one number. Some of the classic IE metrics such as chase percentage, miss percentage of swings and well-hit average all factor into the equation, and provide what IE feels is the most complete measurement of player performance available.
The Player Overall Win Rating (POWR) used by the ESPN Baseball Encyclopedia is based on a system of linear weights that assigns a value to every event in the course of a baseball game. These discrete event values, expressed in terms of runs, are summed and then converted to batting, pitching, fielding and baserunning wins. Player wins in each category are combined to create POWR, expressing each player's value relative to an average major league player. An above-average player has a positive value; a below-average player has a negative value. Players with positive values help their teams win games and put otherwise-average teams above .500; players with negative values cause otherwise-average teams to lose games and thus fall below .500.