THE BOOK--Playing The Percentages In Baseball

Thursday, January 22, 2009

Sabermetric Neophytes: Meet Jeff Kent

Bill Price:

But I say Kent cannot be looked at as a second baseman, especially in this day and age where middle infielders all over baseball are putting up numbers usually posted by outfielders and third baseman. We have to start judging players - other than catchers - as all part of the same pool.

Yes!  Well, he goes way too much to the other side, but “yes” to the idea that you don’t separate secondbasemen from thirdbasemen.  As readers of this blog knows, the fielding talents of 2B and 3B, over (at least) the last decade, are fairly equivalent, much the same way that you would simply look at corner outfielders as one pool of players, and not think to separate them as LF and RF. 

So, big kudos to Bill Price for swimming in my sabermetric pool.

Now, what about Jeff Kent?  How bad (or good?) was his glove? Well:

According to the STATS-based UZR (sUZR), since 2003, he was -11 runs.  According to the BIS-based UZR (bUZR), since 2003, he was -42 runs.  I’ve run WOWY (With Or Without You) on all fielders since 1993.  If I look at Jeff Kent since 2003, I have him at 57 plays below average which (if we count each play as .75 runs for middle infielders, .80 runs for corner infielders, and .85 for outfielders) comes in at -43 runs.  I like that bUZR and WOWY agree so strongly.  The reason is that we don’t have any bUZR data prior to 2002, so I can focus on WOWY for that time period.  (In 2002, bUZR has him at +1 runs, while WOWY is at -1 plays.)

From 1993-2002, Jeff Kent was +1 plays in WOWY, meaning a perfectly average fielding 2B.

Indeed, if we focus solely on the last 4 years, bUZR has him at -47 runs, and WOWY has him at -56 plays (-42 runs).  This means that from 1993-2004, he was a league average 2B.  So, Kent has gotten a hugely bad rap for his fielding, based strongly on his diminished abilities at the tail-end of his career.

So, to go back to Bill Price’s original point, I would consider what Jeff Kent has done as mostly as an average-fielding 2B, with the last 4 years as if he were an average-fielding corner outfielder.

(He also happens to have the equivalent of 0.5 seasons at 1B, at which he played league-average, and 0.5 seasons at 3B, at which his numbers show him to have played very well.)

When you add it all up, we have Jeff Kent as being -26 plays in his career, relative to the average of each position he played.  And the % of his playing time breaks down as:
91% 2B
5% 3B
4% 1B

When we use my fielding spectrum positional adjustments (giving +2.5 runs credit to 2B/3B, and -12.5 runs penalty to 1B), it comes in at +1.9 runs per season.  He played the equivalent of 12 full seasons (since 1993), so he gets +23 runs in positional adjustment, which pretty much cancels the -26 plays he was relative to his positional peers!

So, Jeff Kent should be considered a league average fielder at a neutral position.  He gets NO additional consideration for being a “middle infielder”.  As a hitter, he was very good, but around a career +30 wins above average.  That’s a shade below the borderline level to make the Hall of Fame.  You can make a reasonable argument for him (as a hitter), but in no way can you use his position as a mild plus, without considering his glove as an equally mild minus.