2010 NL East - Virtual Reality

I quoted was from the output of the sim only.

sqrt((p * (1- p)) / n) * g

p = teams win percentage
n = number of games simulated (10,000)
g = games per season (162)

If you run the 2010 season through my sim, with the same projections, you would have around a 68% chance that the Braves win total would be within around 0.8 wins of the total I quoted.

You are quoting a different margin of error. Good luck calculting that. :)
vr, Xei

I think the problem you’re having is related to projection methodology. It’s almost impossible to project playing time accurate to each individual player without a crystal ball to tell who will be injured. You play the odds a bit with guys who have nagging health issues, but for the rest, you simply project that someone is going to miss some playing time.

And you do this universally across each team. The odds suggest that each team is going to have at least one starter who misses time for various reasons, and therefore, they will have a 6th pitcher make starts, and a 7th as well. The Braves were extremely fortunate to avoid any significant injuries to that starting rotation last year. And here’s how the starts broke down

1) 34
2) 34
3) 32
4) 25
5) 21
6) 7
7) 5
8) 4

As you can see, in addition to our principle 5 starters, we had an additional 16 starts, and that is with no remarkable injuries. Part of this is Tim Hudson coming back for the end of the year to make 7 starts, but despite that, the overall distribution matters more than the individual elements since it fits within the league wide trend. For example, the Phillies:

1) 32
2) 31
3) 25
4) 23
5) 12
6) 10
7) 9
8) 7
9) 5
10) 5
11) 2
12) 1

And Mets:

1) 31
2) 25
3) 23
4) 17
5) 15
6) 14
7) 10
8) 8
9) 7
10) 7
11) 5

That’s just the first two teams I pulled up. The Mets had 46 starts from pitchers #6 and up, and the Phillies had 39. The Braves should count themselves extremely fortunate if they’re only needing 26 (or fewer) starts tocome from the outside the rotation they’re starting the season with.

And the excess starters don’t have a great impact on the projections because all such pitchers are going to be very near replacement level, and every team is going to have them. You’ll mention that we have Kris Medlen, who should be well above replacement level, but that’s not a big impact either since he’s already a part of the projections. His numbers will project him to throw more innings than anyone else in the bullpen, which implies he’ll be involved in some of those vacated innings. If you take the 150 innings that the 6th and 7th guys are throwing, you increase Medlen’s innings by maybe 40, which is not enough to make a large difference on the entire season. Maybe it’s significant since we’re talking about a .2 win margin, but this still a projection.

a little bit. I want to be careful to not “play god” as much as possible and inject any personal biases. What I did if I felt a pitcher was going to make less starts or perhaps join the team later in the season due to a season ending surgery in 2009, is that I bumped him down on the depth chart so that he made less starts. I believe for the Braves I had Derek Lowe @ #1 making 33 starts. And all things being equal I would place a younger pitcher who might have his innings watched closely (ie – Hanson) moved down.

Looking at my files, here was the Braves (top 5) distribution I used.

1. Lowe (33 starts)
2. Jurrjens (31)
3. Hudson (28)
4. Hanson (24)
5. Kawakami (19)

Phillies
Halladay-Hamels-Blanton-Happ-Moyer-Bastardo-Kendrick