A week ago I proposed a contest for readers to construct the best possible lineup from the 2002 Giants roster; the rules are here and here. The winner of the contest gets 3 Shadow-points; second place gets 2; and third place gets 1. In addition everyone whose entry beats my own entry gets an additional Shadow-point. Every two months the person with the most Shadow points gets their name on in the Shadow Hall of Fame, $2, and the ability to write any one article for the blog.

-----------------------------------------------

Today, I'm announcing the winnners of the contest. The average submission scored somewhere around 4.9 runs per game, and the best scored above 5. I have more comments to make, but without further ado, the three winners:

In third place, submitting a lineup that scored an average of 5.044* runs per game, is Neil Lender. His lineup is:

1: Yorvit Torrealba (C)

2: Kenny Lofton (CF)

3: Barry Bonds (LF)

4: Jeff Kent (2B)

5: Reggie Sanders (RF)

6: Damon Minor (1B)

7: David Bell (3B)

8: Jason Schmidt (P)

9: Ramon Martinez (SS)

In second place, submitting a lineup that scored an average of 5.080 runs per game, is Colin. His lineup is:

1: Kenny Lofton (CF)

2: David Bell (3B)

3: Barry Bonds (LF)

4: Jeff Kent (2B)

5: Damon Minor (1B)

6: Reggie Sanders (RF)

7: Ramon Martinez (SS)

8: Benito Santiago (C)

9: Jason Schmidt (P)

The first place finisher,The Ralph M. Parsons professor of law at Stanford Law School, was Alan J. Bankman, who sumitted a lineup scoring an average of 5.179 runs; however, he was later disqualified when tests confirmed he shared a last name with myself.

And so instead, finishing in first place with a lineup that scored an average of 5.126 runs is Matthew Nass. His lineup is:

1: Damon Minor (1B)

2: David Bell (3B)

3: Barry Bonds (LF)

4: Jeff Kent (2B)

5: Reggie Sanders (RF)

6: Kenny Lofton (CF)

7: Ramon Martinez (SS)

8: Benito Santiago (C)

9: Jason Schmidt (P)

And how did the submittted lineups compare to mine? Well, I submitted the following lineup:

1: Kenny Lofton (CF)

2: Ramon Martinez (SS)

3: Barry Bonds (LF)

4: Jeff Kent (2B)

5: Reggie Sanders (RF)

6: Damon Minor (1B)

7: David Bell (3B)

8: Jason Schmidt (P)

9: Yorvit Torrealba (C)

It scored an average of 5.114 runs per game--beating all submissions except for Matt's.

So, the Shadow-point payout is:

Matthew Nass--4

Colin--2

Neil Lender--1

The results should be reflected in a widget to the right of the blog.

So, what are my reacitons to the submissions?

First, it looks like basically everyone tried to do the same thing: make a Barry Bonds machine, batting high on base percentage guys before him and power hitters after him to knock him in if he walks.

Second, one thing that almost everyone forgot to look at is double plays. Yorvit Torrealba has a great on base percentage but is really slow and basically only hits ground balls--meaning that he grounds into a double play 24% of the time he's able to, so if there's one out and someone on first his OBP is really more like .2002 instead of .355. (Note that in my lineup I batted him after the pitcher to try to minimize the chances of this happening.) There were some submissions which I think would have done very well but were significantly hurt by their double play rates. Kenny Lofton and Ramon Martinez were good choices for second hitters who don't hit into very many double plays.

Third, playing good players helps. The top submissions basically all agreed on which players to use (except for catcher, where Torrealba and Santiago are roughly equally good). Other lineups tried using J.T. Snow, Rich Aurilia, and Marvin Benard, and it hurt them.

Fourth, no one tried batting Bonds anywhere but third or fourth. For fun, I tried rotating my lineup so that Bonds batted first (i.e. cyclicly rotating it by 7 slots); it turned out to do much worse--scoring 5.048 runs per game. This, I think is partially due to Minor grounding into a lot of double plays when batting fourth--the lineup averaged .833 double plays per game, instead of the .766 that my actual submission did.

The lineups, generally, ranged from grounding into a double play .7 times per game to .9 times per game. For fun, I tried taking the average lineup accross all teams from the year and playing with the double play rates.

Trial 1: unchanged. Double Plays per Game: .823; Runs per Game: 4.266

Trial 2: multiply double play rate of all players by 1.15. Double Plays Per Game: .924; Runs Per Game: 4.216

Trial 3: divide base double play rate of all players by 1.15. Double Plays per Game: .722; Runs Per Game: 4.318

So, it looks like the variation in double play rates from rearanging a lineup cna have an impact of roughly +- .05 runs per game, or roughly 0.8 wins/year--not that much, but it explains about half of the variation within how to order a lineup; not bad for a largely overlooked stat.

For those who were wondering, Joe Bankman's lineup that scored 5.179 runs per game:

1: Kenny Lofton (CF)

2: Ramon Martinez (3B)

3: Barry Bonds (LF)

4: Jeff Kent (2B)

5: Reggie Sanders (RF)

6: David Bell (3B)

7: Benito Santiago

8: Damon Minor (1B)

9: Jason Schmidt (P)

Stay tuned later today for this week's contest.

_________________________________________________________________________________

*: The error on these numbers inherent to the fact that the lineups were only simulated 1,000,000 times is roughly +- 0.003 runs per game

Note: a few typos on the positions people play were fixed.

I'm glad you disqualified Joe. Neither baseball genius, nor random luck could explain his winning, so the fix must have been in.

ReplyDelete