While looking at stats to see where the Astros have room for improvement in 2010, I noticed something odd.  Kazuo Matsui has bizarre splits (insert tasteless anal fissure joke here).  For his career, Kazuo has basically hit right-handed pitching as well as he has hit left-handed pitching.  Here are his nearly identical stats against RHP and LHP:

Kazuo Matsui's Career Batting Splits

This is the sort of split profile we might expect from a switch hitter like KazMat.  But look at the oddities when you break those numbers down into his last three years (2007 in Colorado, 2008 and 2009 in Houston):

Kazuo Matsui's Splits - Last 3 Years

• In 2007, Matsui was better against RHP than he was against LHP, to the tune of +20 AVG, +37 OBP, and +23 SLG.
• In 2008, Matsui's numbers were basically the same whether he was facing RHP or LHP, though he still had an edge against RHP in the power department.
• In 2009, Matsui's splits suddenly go the opposite direction...he hit LHP better than RHP by margins of +26 AVG, +55 OBP, and +69 SLG.

What's going on here?  The variation between Matsui's career, 2007, and 2008 seasons can probably all be waved off as just typical statistical fluctuation.  But normal variation doesn't seem like it would explain 2009's sudden left turn at Albuquerque.  My first instinct is to look at two additional sets of stats, BABIP and PA, to see whether we might have luck or sample size issues.  

Matsui's BABIP and PA Splits - Last 3 Years

The first thing that jumps out at me is that Matsui's BABIP this season was basically the same against LHP and RHP, so that isn't the reason why he was so much stronger against lefties this year.  It probably does explain his better stats against righties in 2007, but it doesn't tell us much about 2009.

The second thing I see is that we might not actually be looking at large enough sample sizes here.  I'd hate for the analysis to boil down to just "need a bigger sample size", though...can anyone else think of any other reasons why Matsui suddenly hit LHP much better than RHP in 2009?

The third thing I see is something that gives me a little bit of hope.  It looks like Kazuo had a run of bad luck in 2009.  His BABIP was 20-30 points lower than his career average this season, and 40-50 points lower than what he was getting in the two previous years (after he left New York).  So unless this dip was related to a permanent injury, we can reasonably expect Kaz to bounce back some next season.