Dave Bush has been dealing recently and after checking up on Rich Harden I wanted to see what Bush might be doing differently in his last few starts.  So I started by making a wear pattern plot for his fastball and saw something stranger than anything else I have seen since I started looking at the PITCHf/x data.  Here is the plot without his start yesterday sadly.

via baseball.bornbybits.com

Look at how much that horizontal movement jumps around.  That is a huge variation so I checked out his game logs to see is that is correlating to his good/bad starts and saw something I didn't expect.  When Bush pitches on the road his fastball has nearly 5 inches less of movement than his home starts.  Actually, it isn't a perfect fit, his first start of the year was in Wrigley and three starts ago (on this graph) he was in Atlanta.  The curious thing is he pitched well in both those games but basically poorly in all the other road starts.  At first, I thought this was a problem with my data corrections so I ran the numbers for Ben Sheets and got this:

via baseball.bornbybits.com

Sheets has some ups and downs but nothing like Bush and no real home/road split (again his last start is not shown).  Four of Sheets last six starts on this plot were road starts for Ben and all very close together.  Also Torres didn't display this when I looked at his sinker a while ago.

So what is going on with Bush?  Normally, when you see large changes in horizontal movement that indicates a lower arm angle and diminished velocity but Bush doesn't seem to show that at all.  He has been very consistently in the upper 80's all year.  Even look at how small the variation in his fastball speed is compared to Sheets or Harden or others.  That fastball speed just doesn't change much at all.

So how can his horizontal movement be changing so much and why is it happening?  This isn't an artifact of the data, this is real.  When he does have the good fastball on the road he pitches well but those starts have been few and far between.  Maybe a home/road platoon with Bush and McClung is the answer because I have no idea how you would go about fixing this.  The good news is this is something that seems to correlate extremely well to Bush's successful starts and his failures.  The Brewers should keep a close eye on the movement on his fastball because you should be able to tell almost right away if he has the goods or might need an early exit.

Edit:

Here is a shot of the home road splits with the road horizontal movement in an aqua color:

via baseball.bornbybits.com

Let me know if this makes some more sense.