THE BOOK--Playing The Percentages In Baseball

minor typo:

“On average, there are 0.3 runners on 1B, 0.2 runners on 2B and 0.1 runners on 1B. “

I assume should be:

“....and 0.1 runners on 3B.”

Thanks, fixed.

Is there a similar simple “theoretical” (as opposed to empirical) way to derive the (negative) value of an out?  Would you simply divide average R/G by the average number of outs per game (slightly under 27)?

I love posts like this one.

Very cool.  This is a great primer/refresher.

Tango, why not finish the presentation by analyzing the out value (absolute, not LW) in terms of bases lost.

David, that’ll be part of a SABR 101 - Outs post.

It was a nice refresher at any rate.

Well, for starters, runners on second and third base will sometimes NOT advance at all on a single!

For someone who is not all that familiar with SABR, this is awesome. Finally I know how you guys value hits and walks. (I assume HBP’s are valued the same as walks?) Thanks from overseas!

Craig/David: hopefully the SABR 101 and SABR 111 threads on outs handles your question.  If not, you can elaborate there.

I’m curious but how come you don’t use these values (1.35 HR, 1.10 3B, 0.81 2B, 0.45 1B, 0.36 BB, 0.36 BB) in the wOBA calculation?  More importantly, why is the recasting of the values to make them relative to outs preferable to using the actual values and calculating the runs created per PA?

Slightly off to the side:  is it worthwhile to note the difference in base advancement between 1B fielded by an outfielder, versus 1B fielded by an infielder?

Polo: your first question is answered here:

http://www.insidethebook.com/ee/index.php/site/article/sabr_111_base_values/

***

Doctor: sure, some analysts do that, notably MGL.

This all seems very nice, but was it checked against actual data?

If I take the runs each event is worth, and the average team MLB data from ESPN for 2009 singles, doubles, triples, HR and BB (BB + HBP + IB), it predicts that the average team should have scored 1162 runs in 2009.  The average team actually scored 747 runs!  That’s an overestimation of 156%.

Same for 2008 and 2007 data (since the number of each events and actual runs is nearly identical for each year).

Am I missing something important?  After all, I would have guessed that the formula should underestimate the runs scored due to not accounting for bases gained due to errors, stolen basis and sacrifices.  But the opposite appears to true, and in a huge way.

Thanks -

Gooseman

Goose: you are definitely missing something important: the outs.

You probably came in through Neyer’s blog.  You need to read the whole series:

http://www.insidethebook.com/ee/index.php/site/article/sabr_101_111_bases_and_outs/

Gooseman, the 1162 is what is called the ‘plus’ runs. To get the actual total you have to subtract the lost runs on batting outs, which is approx. 0.1 * (AB-H)

Cool thanks, folks.  And I did come in from Neyer’s blog.

So simply put, runners left on base is the majority of what’s left out.  In fact, if I take the ratio (from the Outs 101 article) of runners scoring to runners left on base (4.7 to 7.3), and multiply that times the number of runs predicted by the run values above, I get numbers that are within 1% of the actual number of runs scored.  I guess that must mean a few other things cancel each other out (like errors, and...)

That seems like a simple addition to the formula, and gives a more accurate picture of the number of runs each event is worth, right?

Thanks for the fast responses -

Gooseman