Attempting to do some lineup analysis and I may have hit a snag.

by Jack Moore on Dec 2, 2008 3:20 AM EST

So I was looking into whether or not the Brewers were better off with Ned Yost's lineup or Dale Sveum's (both were bad, but that's a different discussion). I was trying to use Table 52 from The Book, which gives run values based on each event based on batting order, weighted by PAs as well.

Batting Order	1B	2B	3B	HR	NIBB	HBP	K	OUT
1	0.515	0.806	1.121	1.421	0.385	0.411	-0.329	-0.328
2	0.515	0.799	1.100	1.450	0.366	0.396	-0.322	-0.324
3	0.493	0.779	1.064	1.453	0.335	0.369	-0.317	-0.315
4	0.517	0.822	1.117	1.472	0.345	0.377	-0.332	-0.327
5	0.513	0.809	1.106	1.438	0.348	0.381	-0.324	-0.323
6	0.482	0.763	1.050	1.376	0.336	0.368	-0.306	-0.306
7	0.464	0.738	1.014	1.336	0.323	0.353	-0.296	-0.296
8	0.451	0.714	0.980	1.293	0.312	0.340	-0.287	-0.286

The problem is that this table is for the run environment from 1999-2002, which, after a couple quick glances at B-R, appear to be a not-insignificant amount higher. As a result, when I checked both lineups, they both (as well as the optimal lineup) come in around 50 runs below average over the course of 162 games. Either the problem is my elimination of RBOE data (since it's not in the BDB file in my SQL database) is big enough to account for 50 runs, or the run environment is so high that the Brewers are actually 50 runs below average using this data. The Brewers scored 750 runs this year (4.63 R/G) which is almost exactly average.

So my question is where did I go wrong? Is the problem as I think, that simply the difference in run environment is causing these values? Is not including the 9th batter screwing things up? Also, is there a way that I could find the proper run values for these events? Could just scaling them down from whatever the average runs/game was over 1999/2002 to the true NL 2008 value fix it?

Thanks in advance.

Tweet Comment 5 comments | 0 recs |

Do you like this story?

Comments

Runs fron pitcher accounted for?

Also the NL’s average ERA was 4.54 last year while during 1999 to 2002 it was 5, 5, 4.7and 4.45.

by Jeff Zimmerman on Dec 2, 2008 8:15 AM EST reply actions

Try here

In the comments section

Adopted Giant: Aaron King

Wearing the crown by 2011. Or at least the LOOGY hat

by baetown415 on Dec 2, 2008 12:39 PM EST reply actions

These are linear weights above average.

In order to reconcile them with team run scoring totals, you need to add .12*PA runs. 50 runs actually sounds like too small of a shortfall.

I has a blog. | Ask a sabermagician!

by cwyers on Dec 2, 2008 12:43 PM EST reply actions

I don't have a whole lot of time right now but I added that in

and it gave me 3.9 Runs/game for the Brewers, which I think makes a little more since given the higher run environment maybe when I get a chance I’ll post the rest of my analysis and I’ll find the problem.

Where does the .12 runs/PA come from, anyway?

---
Juuuust a bit outside!!
http://balkingtraditionalism.blogspot.com/

by Jack Moore on Dec 2, 2008 12:51 PM EST up reply actions

The out value

The quickest thing for you to do is to move the run value of the out closer to zero, the same amount, for each line.

If for example your overall total is -50 runs, and you have 4000 outs, and the total should actually be zero, then add 50/4000, or +.0125 to each of your out values.

by tangotiger on Dec 2, 2008 4:42 PM EST reply actions