MVN: Most Valuable Network

Not a lot to write today.

I have finished the work on my batter projections. There are 2447 projections currently available. Click here for an xls version or here for a csv version. I do need to finish filling in primary defensive position for many of the minor league players. I will repost the files when that is completed.

• It includes all batters appearing in 2008 at Class A or higher.
• For those batters, all levels that they played on in their career are included, including college.
• Stats are park neutral. To find how a batter would do for a given team, that team's factors would need to be applied.
• Also neutralized for level, including college, US minor leagues, and Japan.
• Projection is then regressed to the highest level the batter appeared at
• Also added an age adjustment

All these factors are empirically derived from comparing sets of batting data, described in detail in earlier posts.

Now I have 10 years of projections and actual performances to peruse, looking for any patterns, and analyzing the error distribution of actual data compared to the projections. This will allow me to give reliability scores and curves, like PECOTA's percentile breakdowns.

Probably this wekend I will post a list of the 2008 batters who most over or under achieved compared to the last projections, and are therefor the most likely to regress back towards the middle.