Graphing performance by age and position

Clayton Kershaw -- $215 million over seven years. Giancarlo Stanton -- $325 million over thirteen years. Mike Trout -- $144.5 million over six years. These are just some of the bigger contracts signed in the past year or so, among the highest in average annual value in the history of baseball. How do teams know if they're safe in placing bets of this magnitude? One thing these three players have in common is signing these deals before entering free agency, making them younger than the typical free agent with more productive seasons ahead of them on average.

This Tableau data viz is a box plot graph showing how Baseball-Reference WAR (rWAR) values rise and fall as players age. It includes players since 1960 with at least 100 plate appearances in a season who played fifty percent or more of their games at a given position. All positions are shown, and sliding the age filter shows the gradual decline that occurs as players age. Hovering over individual data points shows additional information. I apologize for not including it in the body of this post, but it's well worth the time and effort to open it in a new window.

This is a screen grab for players at 25:

The data is broken into quartiles, with the two shaded boxes showing where the vast majority of players lie, the ones below the boxes are the busts (at least that year), and the ones above the boxes the best players. The players above the highest line (the whisker) are outliers:

These players run the gamut, players who were and are among the best to play their position mixed in with crash-and-burn flameouts (Marcus Giles) and injuries (Eric Davis).

Do not look for a dramatic drop-off in average rWAR as players near 30 -- players who don't produce tend to stop playing, which is demonstrated by the decrease in the number of players that begins around 30 or so for all positions. This graph shows this for second basemen:

As players mature, they get more playing time and reach a break point around age 26-27 in which they're either considered a legitimate major league player or cast aside for the next man up. As the winnowing process continues, those that continue to perform do so, and those that don't enter the next phase in their life. Average rWAR doesn't vary much as players age, but the number of players decreases dramatically once they reach 30.

Not every position ages the same, and this Google Docs sheet shows the aging patterns for all positions. It's no secret catchers wear down more quickly than other players, but the other positions have their vagaries as well. This is why I was intrigued by the signings of Brian McCann last year and Russell Martin this year -- they're entering the time when catchers start breaking down, and they're going to be owed some big paydays going forward. I wasn't particularly vexed when the Cubs lost out on signing Martin -- I could be wrong, but the odds are on my side.

As teams sign free agents to big money contracts, ask one very simple question -- what is the likelihood that player will be worth his contract:? The data viz has a line at the 5 rWAR mark because at around $5-6 million per WAR, a big money player needs to deliver around that to justify the contract (i.e. roughly what the highest paid players are receiving). As players near 30, the number that cross that threshold begins to diminish dramatically.

There is probability and actuality -- probability shows what has occurred over time and assumes the vast majority of future events will fall within very broad bands of expectation, and actuality recognizes not everything will be the same as in the past. This is the cut line between good-to-above-average players and generational talent, the kind worth every penny they're paid and destined for Cooperstown when their careers are over.

The smart GM recognizes trends and looks for the outlier, that Stanton, Kershaw, Longoria or Trout who really is head-and-shoulders above the rest and pays him accordingly. It's easy for us to identify and pay the superstars (especially since it's not our money), especially if they're young with plenty of their prime years ahead of them. The trick is when players near 30 and the GM has to decide if he's dealing with a typical player who will follow usual patterns of decline or someone who defies Father Time and delivers value beyond the due date. How well these decisions are made usually is the difference between success and failure.

All data from Baseball-Reference

Scott Lindholm lives in Davenport, IA. Follow him on Twitter @ScottLindholm.