THE BOOK--Playing The Percentages In Baseball

Nice article. Drama index neatly condenses games remaining and games ahead/behind into a single number. I wonder if there’s a way of accounting for games played vs direct rivals in the playoff hunt (ie heading into the last weekend of the season tied, the games are more important if the two teams are playing eachother). One way to do that while avoiding a sim would be to compute the team vs team binomial. So if the Mets and Phils had six games left, you’d figure out the chance of the Mets going 6-0, 5-1, 4-2 etc. I guess it would be complicated by the fact that if the Mets went 6-0, they’d arguably need fewer wins to reach the postseason (because the Phils went 0-6 in those games and will end the year with fewer wins).

The Rays didn’t have any drama in this system because, even though they battled the Red Sox for first place in the AL East, they had the wild card qualification to fall back on. Their postseason appearance was never seriously at stake.

I’m surprised a few games didn’t make the list. They were swept in a three-game set at Fenway in early May that dropped them three games off the division and two off the wild card. They began the series tied for first and 1 GB the wild card—you’d figure one of those games was important, even though it was early (although the Gnats had some games close to 1.00 early in the year according to the graph in the article). Also, heading into the all-star break they were half a game out of first and 2.5 ahead of Minnesota for the wild card (with the White Sox and Twins only a game apart, so the wild card was highly contested), and the closer you are to a tie, the greater the leverage/drama.

Looking forward to the player version. I hope the calculation is based on day-by-day rather than month-by-month. Would also be cool to see the entire batting line adjusted rather than just RC.

A more proper way to get this would be to use playoff odds, as Studeman touches on in his article (and which his method approximates). Given a playoff odds calculator, you can calculate Pr(make playoffs | win) - Pr(make playoffs | lose) and you have an index for drama. A 1 is the highest possible score (win and you’re in, lose and you’re out), and a 0 is the lowest (irrelevant).

As for the 55-50 issue: I think this stems from the fact that the upside of “making the playoffs” is not as good as “missing the playoffs,” since the playoffs is only an invite to go to the big stage. Biasing the probabilities reflects this view. With the above playoff odds method, you could try Pr(win World Series) instead of Pr(playoffs), with each team given a 12.5% shot to win if they make the playoffs.

--one of the two Andy Ls, I noticed there’s another…

Scratch that bit about Pr(win World Series), I realized it would make no difference to the number (it would just rescale it)

Early May is just too early to post a big index.  On May 7, the Rays were seven games behind their goal, and their index was .74.  No team reached a drama index of 1 or more until June.

Heading into the All Star break, the Rays’ index was just .375.  In that example, it doesn’t matter how close they were to first place, only to the wildcard team.

You’re right that the index should be higher when there are many teams close to you.  Unfortunately, my index misses that nuance.

Also, my system doesn’t pick up on teams that directly play each other.  I went through and manually adjusted some of those near the end of the season, but that’s all I could do.

For Playoff Probability Added, I’m just going to multiply each player’s game WPA by each game’s index.