A balancing act

The Bud Selig Era has brought several major changes to league scheduling. These changes aren’t talked about a whole lot when reviewing team records because we don’t know how much the changes affect. When the changes are discussed, the focus is usually the 18 or so games from interleague play. Rarely do people complain about the unbalanced schedule, even though it arguably affects as much or more. Consider that, since 2001 when the unbalanced schedule debuted, NL Central teams have played anywhere from 79 to 90 games against their divisionmates; for the bulk of those seasons, any team in the majors would have paid good money to face the NL Central more often (“More games against the Pirates? If you must…”).

To see just how much effect these changes might have had, I adjusted every schedule since 2001 as if it were the 1996 schedule, because 1996 was the only year with three divisions, 162 games played and no interleague play. I’m using the 1996 schedule as it stands for the American League, because divisions haven’t changed much, and modifying the schedule for the National League’s extra teams.

So rather than this breakdown from 2007:

AL        East     Central     West     Interleague
East        72          37       35              18
Central     37          72       35              18
West        44          43       57              18
East        72          41       34              15
Central     33          79       35              15
West        33          42       72              15

we have this breakdown:

AL        East     Central     West
East        52          61       49
Central     61          52       49
West        62          61       39
East        52          61       49
Central     54          54       54
West        49          61       52

Using the latter schedule, I took every team’s winning percentage against a certain division and multiplied it by the theoretical number of games that they would have to play against that division to produce the balanced schedule. Interleague totals were completely ignored (for which the National League thanks me).

In essence, all I’m doing is seeing how team records would change if 2001 and forward had been played on a 1996 slate. This is somewhat like a strength-of-schedule adjustment, only its focus is more on how the teams actually performed against their foes rather than the winning percentages of those opponents. Of course, this method is not without its flaws. A team could get hot regardless of opponent and get skewed accordingly; plus, any team that gets hot in September will be downgraded since it was all against division opponents (2007 Rockies, anyone?). Still, in generalities this exercise should help us see the range of impacts for all the weird Seligian changes and how much confidence we can have comparing one year’s teams to those of another year.

The biggest winners: Most gain from the unbalanced schedule

             Projected     Actual     Gain
2006 Red Sox        76         86       10
2001 Pirates        54         62        8
2004 Rockies        60         68        8
2006 Tigers         87         95        8
2003 Astros         80         87        7
2003 Royals         76         83        7

Most of those teams didn’t exactly tear up the standings, but they certainly got some benefit from the schedule. The Tigers are the only American League playoff team that, in this exercise, should have missed the postseason—87 wins would have put them five wins behind the Twins for the division and six wins behind the Angels for the wild card (the Angels also picked up a few wins). The Bengals had a .600 winning percentage against the AL Central but a sub-.500 against the AL East; that’s the general split of most of these teams. The Pirates managed an amazingly low .233 versus the NL East in 2001 but were a semi-respectable .429 against the NL Central, which gave them a less embarrassing record.

The biggest losers: Most pain from the unbalanced schedule

             Projected     Actual     Loss
2004 Pirates        82         72       10
2001 Rockies        80         73        7
2003 Rangers        78         71        7
2005 Devil Rays     74         67        7
2006 Cardinals      90         83        7

Two obvious oddities: the Pirates of the Lloyd McClendon era feasibly having a winning record (they were .567 against the NL East and .533 against the NL West, but only .416 against the NL Central), and the Cardinals coming off as not a bad team, mainly off their ridiculous .676 against the NL West. Once the Cards stopped facing their own division, they were actually pretty good, a bias that could partly explain the postseason success that few predicted for them.

Other affected races

We have already looked at the Tigers, but here are possible outcomes following a balanced schedule in the last few years. All are National League races.

2001 East (Braves and Phillies tied at 89 wins): This was close anyway, but the Phils gain from being able to play any division above .500, while the Braves stumbled against the NL West (.469).

2003 Wild Card (Phillies, Expos, and Dodgers all tie at 87 wins, with Marlins at 86): If it had been the Expos headed to October, would the Expos still exist to face the Washington Marlins? The Marlins get a 5-win adjustment, as they had a .632 against the East but no better than .487 against anyone else (though their .667 record against the Yankees in the World Series papers over any flaws).

2004 West/Wild Card (Dodgers and Giants both in with 91 wins, Astros out at 87): The pre-Red Sox steamroller that was the 2004 Cardinals would have smashed in the face of whichever team made it, so this one’s not so earth-shattering.

2005 East/Wild Card (Phillies take East with 92 wins, Braves and Astros tie for wild card at 91): This isn’t much news, as the Phillies and Astros battled to the last day anyway. It’s still one win of difference.

2007: Last year’s National League was bizarre from a number of angles, but two divisions and the wild card would be different in a balanced schedule. The Mets would have been four games up on the Braves and Phillies, 92 to 88, mainly by feasting on the NL Central; the Mets were .486 against the NL East and NL West but .700 in between. The NL Central race would be even more limp than it was, with the Brewers triumphing with an 80-82 record and the Cubs just one game behind them. As for the wild card, the Rockies lose only three games in their record, but the Padres gain one and would have outraced the Phillies, Braves, and Dodgers to the postseason.

Using Recurrent Neural Networks to Predict Player Performance
Technology is rapidly advancing possibilities in decision-making.

Very few of those results are surprising, but the last one highlights a key feature of unbalanced play: If all the teams in a division are bad, they will gain some wins from playing each other more often. In that sense, the division becomes a league of its own. In this way, the unbalanced schedule inoculates against the potentially embarrassing 1994 season, where the four worst teams by winning percentage in the American League were all in the West. Who knows if Selig had that sort of disaster in mind when unbalancing things, but it at least can serve that purpose.


I expected the results to swing more than they did, with 10 games being more the norm than the exception. Apparently, interleague play and an unbalanced schedule have not reshaped things as much as they sometimes seem to.

That’s one good thing about baseball: The long season tends to even out some of these things, with the result that weird scheduling ideas can’t be entirely to blame for a team’s problems. A swing of three or four games is still in the hinterland that Pythagoras controls, and most teams with severe differences weren’t in line for the playoffs, anyway. Maybe we’re just lucky that it hasn’t all collided yet, but at least the question of whether the schedule is balanced has not destroyed all semblance of order and useful statistical comparison across divisions.

References & Resources
Forman’s gift to humanity.

Print This Post
Sort by:   newest | oldest | most voted
LV iPhone 6 Wallet Cases

The Android 5.1.1 Lollipop update was initially released in March as a form of maintenance as it is aimed to improve the entire system’s stability and also to fix bugs.