Wrist Injuries and Power: A Quick Glance

It's another one of those baseball saws.  A wrist injury saps a batter's power and can continue to do so during a long recovery period. Spurred on by a question from Jeff Zimmerman, I decided to take a quick look into the truth of the matter.

I'll preface this article by saying my methodology was quick and dirty at best, and should by no means be taken as the final answer to the question.

Using Josh Hermsmeyer's wonderful injury database, I found all the batters with wrist injuries between 2002 and 2008.  For each, I calculated their projected isolated power (SLG - BA) upon return using Marcel projections.  As a simplification, if a player returned from their injury before July 1st, I used that year's projection.  If he returned after July 1st, I looked at the next season.  Finally, I compared the batter's actual ISO to the projected ISO to determine whether he was hitting with less power than expected.

Let me start off with some summary results, with the full data to follow below.

Of the 77 injuries (some players appear multiple times in the list), 32 bettered their projected slugging, while 45 did worse. 

The overall mean difference between the projected ISO and the actual was -0.030, which is pretty substantial.  To be fair, that includes 11 players who didn't play after their injuries.

Update: Daniel Moroz of Camden Crazies pointed out that if I remove those players, the average drop is -0.007 points of ISO, which is quite a bit less. I'm not sure whether those players who didn't return should be counted or not in this so I'll present it both ways.

Finally, the correlation between days on the DL and a drop in ISO was 0.044, which could mean a whole host of things ranging from no relationship at all, to some players trying to come back too quickly and decreasing the days on the DL, or even the reverse.

It appears there may be something to the idea that wrist injuries take away a player's power during his recovery. 

To do this right, there should be a lot more stringency around the choice of dates for the projections and the actual performance.  Also, the study should look at the effects over time and see when recovery does occur (if ever).  This quick look showed enough promise though, that it's probably worth investing the time for further study (hint, hint).

Without further adieu, here's the data: