Open Thread: Cubs vs. Padres, Sunday 5/7, 3:05 CT

First, a team has to go on a winning streak.  Then, the division leader has to lose (or, more correctly, not win) at a rate that's greater than the other team's winning streak of 3 or more games out of, say 12 to be generous (since there's huge difference between the end points of your range).  To get the probability, you'd multiply the likelihood of these two events together.

The first event is far more likely than the second, because you get to choose from every team not in first place having a winning streak.  The second event, however, is far less likely because it's just one team and any team good enough to be leading the division isn't going to have many losing streaks during the season.

Finally, you get to cherry-pick the time frame.  Let me do the same and I'll find week after week after week when no division leader lost 3 games to any other team in its division.  Shouldn't all these failure be included in the overall probability?  Of course they should.

In his exhuberance to make it clear that the probability of coming back from 5 games down to win the division is pretty low, Gauis may have used adjectives that overstate the case.  But he's absolutely correct that the probability is low, lower than, say, drawing to an inside straight after the flop in HE.  Does it happen?  Sure.  Should you bet on it?  Not without great pot odds.  And the pot odds for the Cubs are horrid right now.

The math is really pretty simple.  How many teams each year are 5 games or more out of the division lead on, say, June 1st?  How many of those teams come back to win their division?  Divide the two.  Repeat for each year, then average your results.  If it's greater than 10%, I'll push a peanut down Clark Street with my nose.

If you (the general you) want to hang you hat on a 1-in-10 shot, got for it.  I'm rooting for the Cubs to be the 1-in-10 exception, but my money's firmly in my pocket on them doing so.