A new way of ranking players

I am experimenting with a new way of presenting player rankings.  All stat systems discern a myriad of numerical information into a weighted number.  Is a 0.300 BA good?  What about a 3.0 WAR or WAA?  Numbers are meaningless unless put into context.  A 0.300 batting average is good in modern MLB because it exceeds what most batters can accomplish but simply throwing out that number without context  is pointless.  In some leagues a 0.300 BA might even be below average.

In past months I have demonstrated comparison between this data model and the WAR  system.  Although  WAR is computed very differently, the two systems can be compared by treating each as to how they rank players amongst each other.  In order to simplify this problem I chose  to only rank the top 300 and bottom 300 players for each system.  Players that don’t make either list are considered neutral.  It’s better to be neutral than part of the bottom 300 list however.

The top 300 is a sort from best to worst and the bottom 300 is a sort from worst to best.  The best player in the league will rank #1 in the top list and the worst player will rank #1 in the bottom list.  The number 300 was chosen to cover the top and bottom ten players of 30 teams.  (10X30=300).  When I go back in time to MLB seasons with less than 30 teams I will reduce this 300 figure accordingly.

I chose to use the following formula which spits out a single number that can be used for direct comparison and can be accumulated by a player from season to season.   A player will fall in either the top, bottom, or neutral lists.

The scoring is as follows:

  • Score = 300 – (TopRank -1 )  for players in top list
  • Score = (BotRank – 1) – 300 for players in bottom list
  • Score = 0 for players in neutral list

The highest ranked player in the top list will achieve a score of 300 and the highest ranked player in the bottom list will have a score of -300. Any system devised can rank players and be applied to this algorithm for comparison.

Top ten tables for 2014 have been updated to demonstrate this. The tables now allow for easy access to career numbers where this score is calculated for both this system and the WAR system.  Note that the top ten lists themselves currently show rank; red if that player is in the bottom 300, blue if he’s in the top 300, X if he’s neutral.  Soon Score will be calculated for this page as well.

tl;dr The new scoring system eliminates the vagaries of weighting factors used in various rating systems and allows for direct apples to apples comparison.