THE BOOK--Playing The Percentages In Baseball

While I know Jeff picked those 2 seasons for Lewis because they were in Japan, effectively it’s ‘selective endpointing’. Ff you take Lewis’ whole career, MLB, Japan, minors, and MLE them vs. MLE’ing Darvish’s whole career, I don’t think you end up with pitchers that look all that similar except for a season or maybe two.

I don’t have play by play yet for Japan, but Darvish’s consistently low BABIP combined with extremely low HR% suggests he’s has a very high GB%. A commenter on Jeff’s article did point out the difference in HR’s, as in Colby Lewis with half the HRs is a very fine pitcher.

What’s with the missing 2011 numbers.?  There season is done, right?

Maybe Jeff didn’t have the NPB batters faced for 2011.

885 BFP, 36 BB, 276 SO
.041 BB% .312 SO%

I have the exact same SO% for Darvish, but my BB% is lower

year age  pa  bb  so  bb%  so%
2007  20 790  49 210 .062 .266
2008  21 764  44 208 .058 .272
2009  22 701  45 167 .064 .238 
2010  23 805  47 222 .058 .276
2011  24 885  36 276 .041 .312

#3 - Sorry, I got the number from b-ref and the 2011 season hasn’t been updated.  I could find the TBF for the % numbers.

So when Colby came back over, did he under perform his projection.

Here’s how Oliver (THT Forecasts) projected Colby Lewis after 2009, his last of two seasons in Japan. This projection also relied on data from his 2007 season in the PCL and AL.

Projected
Year  ERA wOBA   BA   OB   SA  BH%  HR%  BB%  SO%
2010 3.30 .292 .251 .285 .392 .302 .033 .039 .224
2011 3.49 .299 .257 .291 .401 .305 .033 .038 .220

Actual
Year  ERA wOBA   BA   OB   SA  BH%  HR%  BB%  SO%
2010 3.72 .291 .228 .290 .369 .276 .035 .078 .236
2011 4.40 .316 .244 .297 .439 .267 .055 .066 .205

Lewis allowed a lower BABIP (perhaps due to a lower gb%), more walks (where he was outstanding in Japan), more HR’s in 2011, and the same SO%.

His gb% for Texas were .39 and .35, which helps to explain the low BABIP but high HR/Con. I don’t have the GB% for Japan to see if the rate had changed.

His projected ERA, from components, was 3.30 and 3.49, while the actuals were 3.72 and 4.40. However, the component ERA for those two years were 3.21 and 3.79, an average of 3.50, where the two year projection average 3.40. The difference could be due to pitching worse than expected with men on base (Oliver currently assumes men on/no runners split is same for all pitchers).