Dave Bush's peripherals

[ESPNOwen]

You are focueing on the wrong stats. Bush's K/9 is 5.81, BB/9 is 2.07 and HR/9 is 1.36 for a FIP of 4.62. Sheets has a K/9 of 7.5, BB/9 of 2.04 and a HR/9 of 0.75 for a FIP of 3.19. They aren't close. Bush is solid but he is getting lucky with a BABIP of .254. He's average which is far from a bad thing but he really isn't comparable to Sheets.

Shouldn't we focus solely on Bush in the second half of the season? I think the real Dave Bush appears from July on.

[rluzinski]

We are calling sample error, "luck". Flip a coin 10 times and you may get 7 heads but that doesn't mean the odds are 70/30 of flipping heads going forward. Sample error can let 10 meatballs turn into 10 outs. With a small sample, the sample error is so large that it completely overwelms the underlying true probabilities. Moeller can even hit for cycle because of it!

 

That's the fundamental limitation of sample statistics and it.s a concept that needs to be understood if you want to properly analyze baseball stats.

[MJLiverock]

I guess that is my problem with calling it all luck. A coin flip has two outcomes, heads or tails, there really isn't a lot going on to influence which one will occur. A pitched baseball is affected by so many differing variables, many of which are a direct result of how the pitcher is actually peforming his job rather than just some sample variance of potential outcomes. The difference between one pitch and the next as well as differing hitters can be so different they aren't even part of the same sample let alone comparing a pitcher's ability to repeat perfect pitches.