THE BOOK--Playing The Percentages In Baseball

Does that take into account their strength of schedule? I mean, the Pirates have to play the Cardinals & Brewers 33 times but they only face the Nationals & Padres for 12 games. That’ll make a big difference in how many games they can win.

They have to play the Cardinals more often than the Nationals.  But they also get to play the Astros more often than the Phillies.

Seems like a wash.

Big?  Can you convert your adjective to a number?  Is Big say 1 or 2 wins, or is Big say 5 or 6, or is Big say 10+?

FWIW, here’s how the Pirates did using Marcel through Diamond Mind over 100 iterations with rosters as of last week, in terms of # of times they won n games.

58: 1
59: 1
60: 2
61: 1
62: 3
63: 6
64: 3
65: 4
66: 4
67: 3
68: 5
69: 9
70: 7
71: 5
72: 8
73: 7
74: 1
75: 8
76: 4
77: 5
78: 4
79: 3
80: 3
81: 1
82: 1
83: 0
84: 0
85: 0
86: 1

Avg: 70.7 W

My average is a little higher, and we’ll see a slightly more normal distribution when the number of iterations goes up, but yep, they won 76+ 22 out of 100 times.

Tango #3, I hope he means “big” as in 1-2 games (at most), because that is all SOS will affect any team, IIRC from an earlier looksie I did.

SOS effects have been greatly exaggerated…

I would’ve thought it makes a bigger difference, since last year Pittsburgh went..

14-29 vs Cardinals, Cubs, and Brewers
11-11 vs Nationals, Padres, & Mets

43 games against those tough teams in the NL Central cost them a few games. if they played 43 games against the Nats/Pads/Mets at a .500 clip ...they’d go 21-22 or 22-21...giving them 10-11 wins more if the schedule had worked that way. That would make ‘em play the Cards/Cubs/Brewers for those other 22 games...which would project to about a 7-15 record there if we keep the winning percentage the same.

So instead of the combined record being 25-40, it’d be 28-37 (65 games). There’s 3 extra wins, just from those 65 games. If the schedule were flipped around more, and Pittsburgh (or any other team) faced other easier teams more often...it’s reasonable to think they could earn another 3 wins every 65 games...totalling 8-9 wins a season.

Yes, I’d call a possible 8 win difference a big difference.

So I’d think that if a projection system accounts for strength of opponent, it might be a little more accurate by at least 2-5 wins.

(of course, the Pirates biggest problem in ‘09 was everyone in the major leagues including their own midsummer youth movement...I’m just using the Pirates as an example since they were used in the post)

MGL, that’s why I asked, because I knew he did not mean 1-2 games, even though we KNOW it’s 1-2 games.

Devon: you are behind the curve on the strength of schedule adjustment.  There’s been a few articles recently on the subject.  I’ll see if I can scrounge them up for you…

Devon, you are not making sense.  No team gets to pick an choose only the weak teams from other divisions to play the bulk of the schedule.

You don’t like playing 15 games against the Cards?  OK, but if you move you also don’t get 15 games against the Astros.  You can’t pick up more games against the Padres or Nationals without also picking up more against Phillies and Dodgers.

There is very little difference in SOS within leagues.  Now take an NL team and put them in the AL, that would probably cost you 8 games or more.

The NL Central was the weakest division last year anyway:

East .488
Central .482
West .520

Is the difference between the al and nl really around 8 games?

I remember reading articles comparing the two leagues but i can’t remember the gap

slightly off topic but since it’s march madness does anyone know of studies done comparing the strength of conferences?we always have this talk on bball boards I go to but I haven’t read anything where people compared them.

The difference is at least .525 minus .475.  That’s 8 games.

That is, a .525 team facing a .475 team will win 55% of the time.  And, AL teams have a better than .550 record against NL teams over these past several years.  Apply a little regression to the observed results, and you’ll get something more than an 8 game difference.

8 is conservative.

Thanks Tango. I’ve been workin’ 2 jobs & tryin’ to keep up with all the readin’ can be tough.

@ Rally… I realize it evens out like that, but even in just the 65 games I mentioned, the Pirates seem to have lost 3 games last year due to the way the schedule works. I have no problem with the scheduling; I personally like it the way it is. I’m just saying that it seems that it would significantly affect the equations for projecting wins. So I figured I’d ask if the MARCEL system accounted for it, ‘cause I didn’t know.

I’ll go read up on the subject more.

Hey guys, thanks for the link. To answer the question about SOS, yes, it’s definitely accounted for. I used Marcel to establish a “true winning %” for each team, and ran it through a Monte Carlo sim 10,000 times using Log5 and the 2010 MLB schedule. The only issue I can see is that since Marcel is based on past raw numbers (unadjusted for SOS, obviously), the inputs already have SOS effects built in, and I’m compounding them by running them through the actual schedule.

Neil: well, great point!