We get a lot of questions and requests related to player salaries, and we’ve got some changes in the works. The first of those relates to improving our LWTS calculation.
A player’s base salary is derived from four factors – current overall rating, potential overall rating, player ‘greed’, and in-game performance. We use Linear Weights (LWTS) as our standard performance metric, evaluated for the last three seasons of the league.
Linear Weights are a way to assign values to each batting event (walks, doubles, strikeouts, etc..) and scale players against league averages. By the end of a season, the sum of all LWTS in the league will approximate 0, and a completely average player will have LWTS = 0. (For a math-heavy description of LWTS, check out Fangraphs article on Linear Weights.)
Until now, all of our LWTS have been calculated with static values for batting events taken from the MLB. But this is not ideal – we know circumstances of leagues can vary widely across different sim models and salary caps and levels of competition.
To get a more accurate measure of performance, we are moving away from those static values, and will begin independently measuring the run scoring environment for each league every season, and calculate batting LWTS from that data.
The impact of this change will be subtle, as LWTS continues to total 0 across all players in the league. We grabbed data from two leagues in each sim model, and the median change in each league to player’s LWTS was between 0.75 – 1.17. The maximum change for all players was 12.98, which translates to a ~$400K difference.
The new formulation won’t begin to impact a player’s salary until their next season, and as performance is based on a player’s last three seasons, it will take three seasons to see the full impact on salaries.
_____________________________
More accurate LWTS calculation is a building block towards further advanced stats.
It is also only one step towards a better salary model, with future plans to move towards:
*Integrating defensive impacts
*Improved overall ratings
*Reducing the impact of potentials more quickly as a player ages
*Increasing the relative impact of performance on salary