A Clever Rule

By David Roher

If you’re like me, you might have spent some time thinking about a peculiar rule in the seedings for the MLB Playoffs: if the #1 seed in a league (the best record among the 3 division winners) and the #4 (wild card) seed are in the same division, then the matchups in the first round are #1 vs. #3 and #2 vs. #4. If not, then the first round matchups are the normal 1-4 and 2-3.

That probably means you’re not like me. But on the off chance you are, or you just feel bad for me, you might find the following interesting.

As part of some other research, I compiled the average record by seed in both leagues for the last 9 years (2001-09). The results:

#1 Seed: 99.7 wins
#2 Seed: 95.1 wins
#3 Seed: 89.8 wins
#4 Seed: 93.4 wins

There’s a big problem here – the team with the best regular season record should be rewarded with playing the worst possible team as often as possible. This wasn’t an outlier screwing everything up either: in every year, the average #3 seed had a worse record than the wild card.

Back to that rule – before now, I had always assumed that it was to ensure interesting first-round pairings between teams who hadn’t played each other a lot. And maybe that is actually the intention. But the main result has been a very fair one: 14 out of 18 times, the #1 and #4 seed have been in the same division. Thanks to the rule, the #1 seed has played the #3 seed 78% of the time. Not bad. Granted, there is still the problem of home-field advantage when the wild card plays a division winner with a worse record. But short of basically abolishing divisions by changing the rule so that the best record played the worst no matter what (a likely disaster from a competitive-balance perspective), it seems as if the current system with this rule is a good call.

About the author


View all posts


  • I have always assumed that this rule was made to ensure that the Yankees and Red Sox meet for the pennant and not in a divisional series. I.e. its about ratings.

  • I did as well, David. That’s why this extra benefit was surprising to me. But I thought some more about this – would MLB really risk a guaranteed matchup between rivals for the relatively low chance (probably no more than 40% even if both were favored) that they would then face each other in a longer series?

    Case in point: Since the last matchup change was made to the Division Series (after 1997), there have been 12 postseasons including 2009. Had this rule not been implemented, we would have seen the following Yankees-Red Sox matchups:

    1998 (DS)
    1999 (DS)
    2003 (DS)
    2004 (DS)
    2007 (DS)
    2009 (DS)

    As opposed to what actually happened:

    1999 (CS)
    2003 (CS)
    2004 (CS)

    All division series matchups, yes, but twice as many.

Leave a Reply

Your email address will not be published. Required fields are marked *