Sabermetrician In Exile | Hacker News Reader

Sabermetrician In Exile

thepostgame.com

40 pointsrobg8 comments

I liked Moneyball a lot, very interesting read. I felt it encouraged me a bit to take more risks, regardless of what others think.

My two major takeaways were as follows:

1) value comes from exploiting market inefficiencies in many cases

2) separate results from process (cause from effect)

3) intangibles are generally a weasel term for fitting the data to a narrative

edit: formatting

Interesting that he was the man behind the Red Sox 2005 draft -- it would be fun to get a look at the stats and analysis he compiled for that. Michael Bowden came right out of high school so I wonder how the difference in stat keeping and obtaining at the high school level affected what he had to come up with.

(anecdote: We went to the same high school but didn't play together. He was on the freshman team when I was a senior on varsity. Stats were usually done by a volunteer student and only kept on paper.)

Everything coming out of the SABR community (SABR=society of american baseball researchers) says that high school stats are 99.9% worthless. The reasons are many, but mainly it's the following: inconsistent contexts (fields, opponents, level of play) and small sample size. Most analysis suggests a full season (~700 plate appearances) isn't really enough for a hitter; h.s. seasons are generally around 100AB.

On top of that, there is the difference in the pitching pool.

I used to do a bit of sabremetrics, back in the day, and I tried to do some projections of AA and AAA players into the majors, and found it almost impossible to get any decent correlations.

I think the minor league (and Japanese league) translations are pretty well nailed down these days. See http://www.baseballprospectus.com/statistics/minoreqa.php

Of course, there is a huge amount of volatility, due to the changing true talent level (most guys in the minors are young, and should be getting better). In addition, there's the fact that MiLB seasons are shorter, and so SSS abound.

But, in the community, these translations are seen as "pretty accurate".

Yep. Knowing HS stats are pretty much garbage, I wonder exactly what data, if any, they go off of with HS pitchers. There were always plenty of scouts around for his games, but they do the standard of watching two innings then leave.

What I've read and heard from talking to a few scouts, teams don't look @ HS stats at all. With few exceptions (Bryce Harper is #1 ATM), they want projectability rather than performance.

Reply on news.ycombinator.com