Sunday, January 31, 2010

Simulating The Entire 2010 Season


What started as an exercise for my baseball simulator to play every game on the Dodgers 2010 schedule has now turned into an exercise to play every teams entire 2010 schedule. I had to take some steps, which may cut into the efficiency of the exercise to prevent this from becoming too time consuming.

Basic methodology: Each teams starting rotation for this exercise will be going 7 deep. The top starters will be getting more starts, but occasionally I have the 6th and 7th starters on the depth charts subbing in for one of the other starters. Each team uses the same algorithm for determining which pitcher starts. I am using each teams projected starting lineup for each game. Bench players are only used in pinch hitting roles. In reality there will be injuries and days off to regulars that would force bench players to start. I went the easy route here to save time. I don't think it should effect the results too much, but feel free to dock or add a few wins to teams you think have superior benches.

The results will be posted in fractions of wins and losses as well as winning percentage up to the date of games that I have finished simulating. I am simulating each game 10,000 times.

Results: Season is complete!

2010 Season Simulated
Place Team Wins Losses PCT
NL West
1 Dodgers 93.8 68.2 .579
2 Rockies 86.7 75.3 .535
3 DBacks 82.7 79.3 .510
4 Giants 75.6 86.4 .466
5 Padres 72.3 89.7 .446
NL Central
1 Cubs 87.9 75.2 .543
2 Cardinals 84.6 77.4 .522
3 Brewers 78.0 84.0 .482
4 Reds 75.1 86.9 .464
5 Pirates 72.5 89.5 .447
6 Astros 71.9 90.1 .444
NL East
1 Braves 90.3 71.7 .557
2 Phillies 90.1 71.9 .556
3 Marlins 81.2 80.8 .501
4 Mets 75.7 86.3 .467
5 Nationals 74.9 87.1 .462
AL West
1 Mariners 82.4 79.6 .509
2 Angels 82.2 79.8 .507
3 A's 81.5 80.5 .503
4 Rangers 80.9 81.1 .499
AL Central
1 Twins 86.5 75.5 .534
2 White Sox 82.1 79.9 .507
3 Tigers 77.9 84.1 .481
4 Royals 72.2 89.8 .446
5 Indians 69.9 92.1 .431
AL East
1 Yankees 102.0 60.0 .630
2 Red Sox 91.0 71.0 .561
3 Rays 82.2 79.8 .508
4 Blue Jays 74.7 87.3 .461
5 Orioles 71.5 90.5 .441
March Madness Picks


12 comments:

Levi said...

Definitely seems reasonable. Thanks for doing this, I'll be keeping an eye on it.

Jesse said...

Great stuff

El Lay Dave said...

I will be very pleased if the Dodgers beat the Rockies by seven games. Should I dock the Dodgers a couple wins for Manny's 30 games not starting?

I see the sim has the AL West as very competitive and very mediocre.

Xeifrank said...

Dave, I would dock the Dodgers around 0.025 wins per start missed by Manny. But other teams will be missing starters at times too, so attacking the problem like this gets more complicated. Good question though.
vr, Xei

Chone Smith said...

What stats are you using for this? Projected stats? Which system?

What simulation? Diamond mind?

Xeifrank said...

Stats/Projections are coming from my own proprietary system (Xizzles). Yeah, I know that sucks as far as open sourcing everything for peer review. All I can say is that it is probably similar to many of the other popular ones. It has a logistic weighting system with aging curves and a regression technique. Another good question though.

I suppose I could run with Chone, Zips, Marcel etc..., but I just don't have the time to do so.
vr, Xei

El Lay Dave said...

So 0.025 * 30 = 0.75 wins docked from the Dodgers for the 30 off days I estimate Manny will need. Not too bad really, knocks the Dodger wins down to about 93 even.

I picked on Manny because of the known amount of non-starts he is due to have in order to have sufficient rest at his advanced age. I don't think there is a similar player in the NL West.

Linus said...

If the Yankees are eleven games better than the Red Sox, I'll eat my hat.

mb21 said...

Xei, do you publish your player projections anywhere?

Xeifrank said...

mb21, I do not publish them. If there is one or two particular players you'd like to see please email me (xeifrank@yahoo.com). Usually all I run my simulator for is individual games, so all I am really concerned about are rate stats. ie - how many HR per PA. For simulating "a particular" game, that is all you need to know (rate stats). More than just the HR portion obviously. :)
vr, Xei

Xeifrank said...

El Lay Dave, yes that is how I would calculate the Manny penalty, without actually taking him out of 30 or so randomly picked games. It's an estimate, but should be pretty accurate. The margin of error for all teams should be around 0.8 wins.
Of course, it is all based off of the input projections I used.
vr, Xei

ppljones said...

AWESOME! Go Mariners!!!