A little while back, after a desperate plea for a balanced schedule, cisblog.ca chief cook and a number-cruncher Rob Pettapiece offered to improve this writer's half-assed math. Here's his explanation:
So I took the 2008 results and built the RPI/SRS rankings, just like I do for CIS football and basketball and so on. From that I get each team's "true" winning percentage.
Then I figured out the odds that each team would beat every other team (e.g., Toronto and Kansas City is about 60-40 for the 2008 Jays). The divisions are the same as in 1990, except Tampa Bay replaces Milwaukee, so we have 13 games against divisional teams and 12 against the other half of the league. Then we can use the odds and the number of games to get an expected number of wins (again using Toronto-Kansas City), it's 7.3 wins for Toronto and 4.7 for Kansas City).
Add up everyone's wins and you get their record for this new season, below the jump:
Below are the standings, with "balanced wins" on the left and the actual wins from 2008 on the right.
AL EastFrom 86 wins and fourth place to 87 wins and third place. Hooray.
94.2 BOS 95
92.9 TBA 97
87.2 TOR 86
86.8 NYA 89
80.2 CLE 81
72.8 DET 74
71.6 BAL 68
90.9 LAA 100
86.3 MIN 88 (in 163 games)
85.9 CHA 89 (in 163 games)
75.3 TEX 79
74.1 OAK 75
72.7 KCA 75
63.2 SEA 61
(Neate: Yes, but with the league's fourth-best record.)
The problem isn't with an unbalanced schedule, it is with unbalanced divisions.
The "true" W-L record (RPI/SRS-based) of the seven teams in the new AL East was .553. For the new AL West, it was .521. The reason both are above .500 is that the AL was better than the NL in 2008. That difference is huge: five wins over the course of a season. So, just because you balance the schedule doesn't mean you balance the quality of competition.
Moreover, moving back to reality for a second, the five real AL East teams from 2008 averaged 92.2 wins (in "true" terms). Not only is that five wins higher than their actual 2008 average of 87, it's way higher than every other division:
92.2 ALEAgain, the solution isn't simply balancing the schedule, because your W-L record is compared against better competition to determine who finishes first, second, etc.
And getting rid of that problem is easy: just get rid of the divisions.
Okay, so that's not going to happen. But you could realign divisions each few years based on quality: No. 1 to the East, Nos. 2 and 3 to the West, Nos. 4-5 to the East, and so on. Rename the divisions to "Robinson" and "Koufax," similar to the NHL once upon a time. If we did that based on this "true" record, the 2009 divisions would look like:
Robinson Division: Red Sox, Blue Jays, Yankees, Indians, Rangers, Royals, OriolesThe expected results don't change much, but at least they're more equal.
Koufax Division: Rays, Angels, Twins, White Sox, A's, Tigers, Mariners
And hey, if you took all 30 teams, divided them in the same way for the 2010 season, made the DH home manager's choice (to remove the AL-NL difference), you'd have these five divisions:
Robinson: Red Sox, Cubs, White Sox, Indians, Mets, NationalsEach team would play 13 games against your own division (65 games) and four or five against three of the other four (adding up to 75 more). The five division winners and an at-large wild card would advance to NFL-style playoffs (3 seed vs. 6 seed 4 vs. 5 in the first round, top two seeds receive byes). The season is down to 140 games, which you could run from April 19 to Sept. 19. Playoffs go Sept. 21 to Oct. 10 and the World Series would be Oct. 15-23.
Koufax: Yankees, Rangers, Cardinals, Mariners, Giants, Pirates
Ruth: Angels, Phillies, A's, Braves, Orioles, Padres
O'Neil: Rays, Dodgers, Tigers, Rockies, D-Backs, Reds
Clemente: Twins, Blue Jays, Marlins, Brewers, Astros, Royals
The first- and second-place teams would get about two weeks off before starting the playoffs, which is wonderful for them after a five-month season.
And that's a three-quarters-assed look at the situation.