They trotted out two recent examples: the Orioles' 10-run comeback to beat the Red Sox, and the Marlins' football-score victory from July 9.
The team on MLB Tonight described snatch-'em-backs them as games you should have lost. That's good enough for them but woefully short on specifics for a guy like me. If the internet teaches us nothing else about baseball, it teaches us that virtually everything in a baseball game can be captured by a number.
So we turn this time to the marvelous FanGraphs Baseball, which will tell you exactly how improbable those victories were.* It uses a statistic called Win Expectancy (WE) which indicates how likely it is for the home team to win the game given every score, inning, baserunner, and outs situation.** And since the WE for any baseball game will always be between 0% (you lost) and 100% (you
won), it's extremely easy to compare any two baseball games.
*The Infinite Improbability Drive of Baseball, for those Douglas Adams fans out there.
**The basis of the work is an actual historical assessment of every game in which a particular situation (e.g., home team up 5-3 in the bottom of the 6th with a runner on second and one out) had occurred, but the method FanGraphs uses has some more math behind it.
So, those two games. Well, they're both very improbable, but the Orioles victory is far less so. The Marlins, at their lowest point - when Chris Young grounded out to score Justin Upton - were expected to win that game 1.4% of the time. The Orioles, at their lowest point - when Dustin Pedroia singled to score Jeff Bailey - were expected to win that game 0.4% of the time: 4 out of 1000 times!
Funny thing. Some people referred to this Orioles game as revenge for the Mother's Day Miracle. You know what: they're exactly right! The Red Sox's Win Expectancy for that game was also as low as 0.4%.
I still don't know what a snatch-'em-back win is, but spiritually it seems like it should be less frequent than a 1-in-100 shot, so sorry, Marlins, your July 9 effort doesn't cut it. But if the WE drops into the 0.x% range, I'll pay attention. By my own, arbitrary method of counting, only the Red Sox collapse counts as a snatch-'em-back win for the Orioles. Oh, that makes me feel much better.
A couple of other notable comebacks, WE-style:
- 2008's Red Sox playoff comeback against the eventual AL champ Rays saw the Red Sox coming back from a 0.7% WE. (This one had the additional improbability of JD Drew showing some emotion on the field: a subtle bat-flip on his two-run homer in the 8th.)
- On opening day, 1998, the Sox stunned the Mariners on a Mo Vaughn grand slam in the ninth, thereby reversing a 0.4% WE.
- On Aug 5, 2001, the second biggest comeback my sleep-deprived brain can find occurs when the Mariners, in the midst of a 116-win season, blow a 14-2, 7th inning lead to the Indians, who win 15-14 in 11. The Indians' lowest win probability: 0.1%.
- August 21, 1990, the biggest comeback I can find: the Dodgers lead the Phillies 11-1 going into the 8th and 11-3 going into the 9th, both of which situations give them a 100% WE. Well, that must be rounded up to 100%, since the Dodgers lost 12-11.
Posts
Excellent research on this one, Ted --- and very interesting subject matter. I didn't run the numbers (though I suspect they wouldn't rank that high), but the biggest "snatch 'em back win" for me has to be the 2004 ALCS.
ReplyDeleteOn a sidenote, I find it amusing that you and I do most of our writing around the witching hour --- something to do with the amount of free time that children allow, perhaps?
I checked games 4-7 of the 2004 ALCS, and the games weren't this improbable - the lowest WE was 12.2% (top of the 8th of game 5). Maybe that's also a testament to Tito's mantra to win tonight's game and worry about tomorrow, tomorrow: the task of taking the series was WAY more daunting, but any individual game was clearly winnable.
ReplyDeleteAnd yes, about 1:00 a.m. is when I have free time!