Monday, June 3, 2013

Two-Out Runs

I've added another element to the Mistake Index, one that deserves its own explanatory post. It's the number of runs teams score (or give up) with two outs in an inning. All data is through games played on Saturday, June 1st.

I was watching the Cubs-Diamondbacks game on Sunday and watching Edwin Jackson pitch another infuriating game--four wild pitches (the record for a game is 6, and only five players have thrown 5 or 6 wild pitches in a game), plenty of questionable defense by  the Cubs, and despite this, tied 4-4 after five innings. In the top of the 6th, Jackson got the first two outs and then went single, single, single, intentional walk, single, at which point he was replaced. Three runs scored with two outs--I can already get inside Dale Sveum's head and read his thoughts after those two outs: "He'll get us through six and then the bullpen can take over, except I'll have to use Carlos Marmol and I don't really want to." That close to getting what he wanted, and (steak dinner) BOOM, the game was essentially over because of those runs scored with two outs.

This charts show runs scored in 2013 by number of outs in the inning:

0 outs21497193991604499910469063815191718514154076.258.313.420.732.294103
1 out20922185782724480393680537260532613317274135.259.322.404.727.300102
2 outs20296182532703442687689496258139511018734313.242.319.382.700.29295
Provided by View Original Table
Generated 6/2/2013.
It's easily found at Baseball-Reference in the splits data. It's illuminating to see what happens when teams reach two outs, since two opposing forces are at work:
1. Chances are there are more base runners than with 0 or 1 out.
2. Hitters really do perform worse with two outs--based on some research I'm doing on clutch hitting, that 17-point drop in batting average from one out to two isn't just statistically significant, it's STATISTICALLY SIGNIFICANT (even though true statistical significance doesn't have magnitude--it's binary, either significant or insignificant).

So, most runners on base, but pitchers have a decided edge--could very well be a stalement. The numbers suggest otherwise, since 38.4% of runs are scored with two outs, around five points higher than if runs were evenly distributed by outs. Add to this my humble acknowledgement that my first attempt at quantifying two-outs runs was, um, how shall I put this, er, WRONG, I became very curious--my question:
Are some teams better at preventing two-out runs than others? And if so, does that translate into success?
The easy answer, of course, is NO, there's no way that this ALONE can explain anything, but when I throw it in the hopper with the other variables I tabulate in the Mistake Index, it adds more grist to the mill (like those agricultural metaphors? I'm from Iowa, after all). After writing the CORRECT equation, this chart shows how many runs teams have scored with two outs AND how many they've given up:

I don't want to lose sight of the most important item, which is that no matter when runs are scored, a team will obviously have greater success if it scores more runs than it gives up. This kind of analysis can bury that very basic fact, but consider the two questions I posed:

1. Are some teams better at scoring with two runs? We have to look at both percentage AND quantity--fully 50% of the White Sox runs have been scored with two outs this year, and they're also 29th in the league in runs scored (thank God for Miami, right, Sox fans?). Texas is good at scoring runs and doing it with two outs, but they play half their games in the second-best park for offense in the game. Oakland does NOT do well scoring runs with two outs, but they overcome. It's a mixed bag, to be sure.

2. Are some teams better at preventing opponents from scoring with two outs? I'll keep writing it in every post I do on this subject, but the beauty of what I quantify in the Mistake Index isn't just what teams do, but their opponents as well--it's the DIFFERENCE in mistakes that is a fairly substantial line between success and failure. The Reds give up quite a few runs with two outs, but not many overall. 

As I mentioned earlier, I'm not suggesting this is the end-all for baseball success, and it is a mixed bag--the greatest opportunities to score with base runners and the fewest opportunities to bat, since the next out will be the last in the inning. I don't know if it means anything, but since it bugged me to watch the Cubs give up those runs with two outs, and I have the ability to quantify it, well, why the heck not?

And here's one for the ages--it may not be the worst example, but it's gotta be knocking at the front door. On July 30th, 2010, the Cubs were at the Rockies and down 5-2 going into the bottom of the 8th inning. Sean Marshall was put in to pitch and gave up hits to Clint Barmes and Melvin Mora but struck the next two batters out. With runners on 1st and 3rd with two outs, Tom Tango states the chance a run will score at 28.8%, with the expected runs scored to be .53, essentially a coin flip--sometimes a team will score a run, sometimes not. On this day, the Rockies scored, and scored, and scored...12 runs with two outs in the inning. THAT'S a humiliating kick in the crotch, to quote The Police.

No comments :

Post a Comment