Interesting lineup analysis

I will be the first to admit that this program has its flaws - primarily, speed can't be taken into the equation - but this ideal-lineup-generator I found on the Interwebs is pretty interesting.

Using OBP and SLG, this program determines runs created per game based on each possible batting order, and spits out the best and worst possible batting orders. Let's see how the way Vic lined them up matched what the analysis calls for.

I put in stats for the following lineup: c) trezza 1b) davilla 2b) roche ss) perez 3b) luis lopez dh) josue lopez lf) teilon cf) weed rf) torres. I used Teilon and Davilla's average career statistics instead of their season statistics, as the sample size this year was too small; for the rest I used '07 stats.

In the real world, the Spirit scored 4.978 runs per game. On average, the program predicted around 5.3 r/gm for the Spirit, although you have to remember that in real life, the Spirit lineup didn't always have Davilla, Teilon, or the Lopezes ... if they had those guys all season long they probably would have scored closer to 5.2-5.4 runs per game.

The following lineup was projected to score 5.401 runs per game:

R/gm
5.401	JLopez	Davilla	Trezza	Perez	Torres	Roche	LLopez	Teilon	Weed

Having Josue lead off obviously doesn't make a lot of 'conventional' sense; the first projection that had a reasonably normal leadoff hitter had Jerson Perez leading off.

R/gm
5.397	Perez	Davilla	Trezza	JLopez	Torres	Roche	LLopez	Teilon	Weed

Interesting in that I wouldn't really expect the 'ideal lineup' to have Trezza at #3; obviously Ks aren't taken into the equation here.

Using ONLY stats and projections from 1998-2002 (the original formula includes 1954-2004), the following lineup looks good:

R/gm
5.442	Torres	JLopez	LLopez	Davilla	Perez	Teilon	Trezza	Roche	Weed

That's pretty solid, but I might swap Josue with Perez.

Almost every single one of the 'best lineups' using 1954-2004 had Vic Davilla batting 2nd. Using 1998-2002 stats, one of the Lopezes was considered ideal for the no.2 hole with Vic usually in the 4 spot.

I couldn't find anything close to our actual lineup.

The 'worst' possible lineups were the following:

R/gm
5.120	Teilon	Weed	Trezza	Roche	LLopez	JLopez	Perez	Torres	Davilla

R/gm
5.090	Teilon	Roche	Perez	Weed	Trezza	LLopez	Torres	JLopez	Davilla

This demonstrates a few things to me: First of all, it's not all that surprising that this is considered a 'bad' lineup. Secondly, it also demonstrates that the Spirit, when at full strength, were a damn good offensive team --- the worst, dumbest, most insane possible batting order still would theoretically have averaged 5.090 runs per game. Unfortunately, injuries, major league purchases and the like prevented this team from reaching its full potential throughout the season.

BY THE WAY:
Another little tid-bit from baseballmusings.com, just to further piss off any Mets fans that might read this blog [sorry Mets fans (not really)]-

"According to Baseball Prospectus, the Mets collapse was the second biggest in history, just behind the 1995 Angels. That's based on fall from the highest probability of making the playoffs, based on BP's monte carlo method. On August 20th, 1995, the Angels had a 99.988% of making the playoffs. The Mets were 99.80 at their peak on September 12th. Now if you factor in time, I think the Mets fall was worse, but not much. The Angels peaking in late August says they were a great team in a weak division with weak wild cards. The Mets were a good team with a big lead late."

Statistical analysis

The other night at Fraser, I got into an argument with a nice but misguided older gentleman over the value of various baseball statistics; when I said that I'd rather have a pitcher who went 3-7 with a 2.60 ERA than one who went 7-3 with a 5.50 (Duh???) he told me that I "obviously don't watch baseball games," I just "look at box scores on my computer." Oh, OK. I don't watch baseball? Coulda fooled me :)

He also told me that Manny Ramirez is less valuable than Kevin Youkilis... I love Youk, but come on.

That argument reminded me that I've never posted links here to some of my favorite baseball stat websites.

This website is home to much of the empirical evidence used to complete The Book - Playing the Percentages in Baseball, which, for my money, is the best book about baseball strategy and stats ever written.

Some of my favorite tables and charts:

The classic run expectancy chart, which shows the average # of runs per inning based on how many runners are on base and how many outs there are

This chart is where it gets really interesting: run expectancy broken down by how many times exactly 0, 1, 2, 3, 4, or 5+ runs scored based on the situation. This chart is friendlier toward people who like sacrifice bunting all the time (but not much friendlier).

This one shows how many runs on average each result is worth based on the situation. I.E., a home run with nobody on and nobody out is always worth 1.000 runs and a strikeout with 2 outs is always worth 0.000 runs. It's kind of thick, but this page explains the methodology.

I got linked to The Hardball Times two years ago when they had a debate on the value of 'productive outs' (some interesting and mixed findings there, feel free to explore) and I have kept up with that site ever since.

Sadly, it's not really possible to find really detailed statistics on the CanAm League unless you have extensively [and I mean extensively] scored every game yourself.

Stolen Bases and Wasted Outs

One of my pet peeves all year has been how many runners the Spirit lose on the basepaths. It seems to me to be an inordinate amount - far more than I ever remember seeing from any team I've followed.

Interestingly enough, looking at the stolen base figures, we aren't doing THAT badly. Bearing in mind that Baseball Prospectus says stolen bases only give you a statistical advantage if you are successful at least 73% of the time, here's how the teams stack up:

Team: SB/CS/ Pct

New Haven 85/29 77.9%
New Jersey 96/30 76.2%
Grays 90/31 74.3%
Brockton 87/31 73.7%
Atlantic City: 92/33 73.2%
-------
Sussex 48/66 72.7%
Quebec 66/26 71.3%
North Shore: 70/29 70.7%
Worcester: 65/33 64.9%
Nashua: 60/51 59.4%

Focus on the Spirit:
Josue Lopez 1 SB 1 CS
Mike Torres 6 SB 4 CS

------

A few observations from those stats:

-The Spirit aren't hurting themselves as badly as I thought they were re: steals. 70.3% is statistically not helping you, but it's not hurting you that much, either. 70 successful SB is pretty good considering the fact that this isn't a team laden with speed top-to-bottom. But the fact that Josue Lopez has 2 attempted steals and Mike Torres has 10 is just silly. Those guys shouldn't be stealing. I hope that most of those CS's come from pick-offs. [Not that getting picked off is something to be happy about.]

-As I suspected, CanAm managers are way too over-reliant on trying to steal bases, to the point where it's statistically hurting half of the teams in the league. I'd suspect that sacrifice bunt stats would also show a similar over-reliance - and statistically speaking, bunting helps you even less often than stealing.

-Also, as I suspected, Butch Hobson's managing is terrible. A 59.4% stealing percentage is so bad that it makes serious dents in run totals. There's nothing wrong with trying to steal a lot if you're doing it well (see: Jersey) but continuing to run wildly when you get gunned down so often - for the sole purpose of being 'aggressive' [Nick Lopardo word of the day] - is downright foolish.

*

So, in looking at the actual numbers, our stolen base ratio isn't killing us, but it's not helping out much either, and we are certainly well within the bottom half of the league in terms of success stealing bases.

I suspect, however, that if I could find statistics about runners being gunned down at the plate, or trying for extra bases, we would probably on the top of that dubious list. If anyone knows where I could find such stats, let me know.