Which Player Benefited Most from Triple-Crown Stats?
This post was born out of a conversation randomly started by Ronit Shah on Twitter. He asked me:
“Is there a more interesting player page than this one? #CoorsEffect”
The page links to Dante Bichette and I did a little research, eventually concluding that Bichette’s slash line was .360/.397/.642 at home and .268/.303/.431 away from 1993 to 1999. He also had a HR/PA of 0.06 at home versus only 0.03 away and hit 1.5 times more extra base hits at home than he did on the road.
Satchel Price decided to chime in, asking the question I asked in the title: “Has any player ever benefited more from triple crown stats? Bad defense, rarely walked, played in Coors during the steroid era.”
I was interested to see what I could find on this, so I made up an analytical tool. I downloaded all player careers for all qualified batters from Fangraphs. Then I found the average and standard deviation for home runs, runs batted in and batting average for all of these players.
Finally, I used the formula (player HR – avg HR)/(SD HR) to find some sort of variance (I use a similar method for fantasy baseball rankings). Then, I added up all of these variances for each player and graphed them versus fWAR. I have basically no idea what this is measuring, but it’s measuring…something. I deleted all players who have a variance below negative 4, leaving me with this:
Finally, based on the linear relationship, I found which players have the greatest discrepancy between expected fWAR based on triple-crown stats and actual fWAR. Here are the top-10 players whose fWAR is much lower than it should be based on their triple crown stats:
Turns out Satchel was right. Obviously, this methodology is severely flawed; however, it worked out pretty well and was a fun exercise to do.
This post originally appeared at Steal of Home.
Nice article, I did something similar a few months back with OPS and WAR. What one would expect is the players that are hurt more in this respect either are bad defensively (Delgado, Manny…), or can’t take a walk (Bichette, Carlos Lee, GA)
Craziest line on his player page is 1999, when his .376 wOBA went along with a 98 wRC+ and -1.8 (!) WAR. Times really have changed.
You fool! This is obviously an exponential relationship. The r^2 is like .002 higher!
I agree that it’s an exponential relationship. If you don’t remove the variance below negative 4 people it really shows it. Unfortunately, I realized this once everything was already posted.