THE BOOK--Playing The Percentages In Baseball

Doesn’t Kincaid comment here? I’ve seen that name somewhere, maybe RAB.

Never posted under that name:

http://www.insidethebook.com/ee/index.php/site/members_list/

He posts on FanGraphs a lot.

Nick--

Didn’t realize you were vivaelpujols. I like the name

I like the writing and it is a pretty good effort, but I’d really like to know what they did to produce a “Marcel” projection for each pitcher.  It looks a little to me like no (or too little) regression was done for the projections.

It looks like the improvement we see in the STL pitchers (.2 in FIP, from 4.64 to 4.44) is merely regression, but I could be wrong.  Players who go from one team to another tend to be below average players and tend to be coming off below average years, so we will tend to see an increase in performance for all players who switch teams.  I’d also like to see the results of a control group - namely pitchers who were with a team or teams for 3 years and then went to another team other than STL.

Anyway, .2 in FIP in the NL and nothing in the AL (Oakland) is not a whole lot of evidence of anything for a sample like that.  The author does admit that of course.

I love his comments about the average fan and broadcaster’s understanding of statistics and what have you…

Thanks dan.

Did Scott Boras really post here?

Crap, I mispelled my own name 15 of 420 times.

Thanks for the link Tom.  I haven’t posted here before, but I often comb through this site and read through the archives to find various bits of information.

MGL, I more or less followed Tom’s write up on Marcels.  I didn’t apply an age adjustment because I wasn’t 100% sure I was doing it right, and because the only pitchers it would have made much difference for were the older ones that would have raised the projected ERA, so it was erring a bit against showing evidence of anything.  Not exactly the best way to go about it, I know.  For regression, I played around with the number of innings of the average to use until I got reliability figures to match up pretty closely to Tango’s published figures.  This was the first time I have attempted Marcels, so I may well have done something wrong.

I should have looked at a control to account for only looking at pitchers changing teams.  I’m still learning to pick out all the right factors to account for, and am constantly finding things I’ve missed.  Thank you for the suggestion.

I don’t feel like this study is really close to the standards of this or any other reputable site, and I mostly wrote up the article for the practice and experience even though it didn’t really say anything substantial.

Kincaid, while there are definitely some things you can do to clean up your study, I loved the approach and I particularly loved the way you couched the whole issue of the media and baseball insiders considering Duncan to be a genius, etc.

BTW, sometimes you can avoid the whole issue of adjustments, park issues, Marcel’s etc., just by using a control group and making sure that you have reasonably matched samples in each of the two groups.  This is a good opportunity for that.

I really liked the way you wrote up this piece!