Tuesday, April 14, 2015

Strasburg and Struggles with Errors


In the midst of some disappointing play by the Nationals over the last week of real baseball came Stephen Strasburg’s first start of the season. It was another disappointing game for the Nationals in a string of disappointments, but especially for disappointing Strasburg, who pitched 5.1 innings and gave up 6 runs (3 earned) in the loss. Strasburg was really undone by the error committed behind him by Ian Desmond. As the Mets runners crossed the plate, FP Santangelo mentioned what all Nationals fans were surely thinking, namely that Strasburg struggles with composing himself after an error is committed behind him. Surely, this is right. I can definitely think of times where this was the case. Then again, maybe we are victims of selection bias. It’s much more likely for us to remember the cases where an error behind Strasburg led to runs for the opposing team. But if Strasburg squashed the rally quickly, we probably moved our focus onto whomever was due up in the next inning and forgot all about the error and Strasburg’s recovery.

I decided to find out for myself whether an error behind Strasburg really causes him to lose his composure or if we are fooling ourselves with this narrative. To investigate, I used the following methodology: With Baseball Reference’s Play Index, I found all instances where an error was committed while Strasburg was pitching during the years 2012-2014. That gave us a total of 22 data points. The way I understand this narrative, the possible effect of the error on Strasburg is limited to only that half inning, i.e. once the inning ends, Strasburg has a chance to retreat to the dugout and collect himself and the impact of the error is eliminated. So for each of these 22 errors, I pulled the data for the rest of the innings to see what happened. Enough of the boring mechanics, let’s get to the results. The chart below shows the games where an error occurred, the opponent, the inning of the error, and Strasburg’s final line for the game.

Date
Opp
Inn
IP
Runs
K
BB
b7
7
1
9
0
t2
6
3
4
1
t2
5
3
8
1
b3
5
7
3
1
t1
6
0
9
1
t4




b3
5.1
6
4
5
t3
6
2
7
1
b1
6
4
6
2
t1
7
3
7
1
t5
5
4
7
2
b5
8
2
4
3
t6
7
0
8
4
t5
8
1
12
0
t1
7
5
9
1
b6
8.2
4
8
2
t4
4.1
6
6
3
t5
6.2
1
12
1
t2
7
2
4
1
t2
6
2
9
1
t5
7
4
11
0
t3
6
3
8
0


Innings pitched following the error: 13.2 (i.e. 13 and 2/3 of an inning)
Runs allowed: 16
Strikeouts: 19
Walks: 7
Homeruns: 1
Batting Average Allowed: .241
On Base Percentage Allowed: .328
Runs/9 innings: 10.54

It certainly does look like Strasburg has some issues following errors. It’s not a one-to-one comp to ERA, but the runs/9 average of 10.54 certainly isn’t good. Strasburg’s career strikeout to walk ratio is over 4, but following an error that drops to 2.7 strikeouts per walk. His career batting average allowed is .224, only slightly lower than the .241 we found he allowed following an error. Pretty much any stat I could come up with looks worse in this analysis.

So case closed? I’m not sold yet. Taking Strasburg’s stats over all 21 starts (one start had two errors occur), you get a Runs/9 rate of 4.23, a much more reasonable number when you consider that includes the unearned runs that resulted from those errors. His strikeout to walk ratio is even up over 5 in those games, so you could potentially make the argument he actually picks up his game in other innings.

I’m not ready to conclude with certainty that Strasburg does in fact struggle following errors because I have a lot of issues with my own analysis. First off, this is a pretty small sample size: only 22 instances over 3 years and 13 innings in total. There is a lot of noise in these numbers. Look hard enough and I’ll bet you can find another 13 inning stretch of innings in Strasburg’s career that will result in similar stats (keep in mind that our sample size is only about the length of two starts). Second, we don’t have anything to compare this to. I created my own analysis for Strasburg’s performance following an error. Since I have a day job that actually pays the bills, I stopped at that one comparison. In a perfect world, we could see how Strasburg has performed and compare that to his peers. Strasburg’s 10.54 R/9 is not pretty, but maybe that’s actually low for the league? Without more analysis, we can’t know for sure. By definition, an error occurring means there is at least one opposing player on the base paths. Pitchers on the whole pitch slightly worse with baserunners on and the opposing team is more likely to score with a runner on base than with no one on base at all, so an error that allows a baserunner on is inherently putting any pitcher in a tough spot.

While this analysis is limited, it certainly doesn’t refute the notion that Strasburg struggles following errors. There may not be a fire in the numbers yet, but I get a whiff of smoke. In 13 of the 22 data points, Strasburg did escape the inning with no damage done, good enough for over a 50% success rate. But over those 13 innings pitched following an error, Strasburg allowed 16 runs and his walk rates and strikeout rates both got worse, maybe a bit of an all-or-nothing effect taking place. If we have in fact identified this as a problem for Strasburg, I’m not sure how you resolve it. Being a professional pitcher is not a stress free job, so Strasburg has had to deal with pressure his whole career and men on base due to errors is nothing new. Is there something about an error that just hits so deep within Strasburg’s psyche that he can’t escape it? Has his shrink known about this for years but has been unable to tell anyone due to client confidentiality laws? Maybe there are some questions that we will never be able to solve. If Ian Desmond would just stop booting groundballs then Strasburg wouldn’t have to deal with these types of situations at all.

No comments:

Post a Comment