clock menu more-arrow no yes mobile

Filed under:

Andy Sonnanstine's Thesis About Random Walks in the Ballpark

In the end, it's one start.

I remind you and myself of this by stating it upfront. If this were Wade Davis or Jeremy Hellickson, Matt Garza or James Shields, this would be the proverbial adhesive stuck to the mental box score with a staple attached just in case. It's not though.

Four walks, by Andy Sonnanstine, in one game? That makes no sense. I've checked the stat line thrice over at this point because I find it hard to believe. From 2006-2008 Sonnanstine never walked four in one game. That stretch spans three levels of baseball, 93 starts in total and Sonnanstine never, ever walked four in one game. Tonight marked his third major league start of the season in which he's walked four.

It's one game, but it's really not.

You watch Sonnanstine pitch without the counts and without the batted ball results - strip the game like they do in football - and it's incomprehensible how he walked this many. In the first inning 12 of his 20 pitches were strikes against five batters faced; 13 of his 22 pitches in the second were strikes to even batters faced; 8/13 in the third versus three; 17/24 in the fourth to eight; and 4 of his 5 against two in the fifth. 60% all told.

With James Shields and David Price I've discussed the random variation related with walks and such. Some would say random lapses in control. How do you explain these random splotches occurring to a guy renowned for his ability to avoid them? Why is it happening, why now, and how can it be fixed?

I don't really know. I can splice pitchfx data from now until dawn and odds are I couldn't find anything damning. If Josh Kalk presumably can't find a problem - and it's his wont and job for crying out loud - then I'm no match.

Assuming Sonnanstine's true talent walk rate was ~5% of the total batters, then the odds of him walking 4 within 30 is less than 2%. You can see where eyeballs get protruded when it happens so many times in such a short period of time.

Either you assume his control has worsened - a distinct possibility given his increase in balls - or that he accidentally ran over a black cat during the off-season.  It gets hard for me to label it random variation at this point. Either he's hurt or mentally he thinks he has to be perfect every time he throws a pitch, leading to some physical gaffes.

How do you fix that? Can you fix that? Again I don't know. If I was enlisted by Oliver Perez's foundation - the Center for Nervous Pitcher Breakdown Prevention - to write a pamphlet on prevention would read as this:


Joe Maddon is apparently not a fan of my theoretical literature and decided to pitch out and force Sonnanstine's hand against Dustin Pedroia. The at-bat resulted in a walk and maybe it occurs anyways, still pitching out twice in one plate appearance is a terrible idea. Maddon has confidence in Sonnanstine to throw a strike but really he should have more confidence in Sonnanstine's ability to hold a runner. Dating back over 69 career starts in the bigs, Sonnanstine has allowed 3 steals and saw 13 caught stealing. Either he, his catcher, or both know how to keep the guy on the base. Yet Maddon didn't trust him with Ellsbury.

Enough about the other things though, let's look at the physical properties of his pitches.

He used a rainbow of release points to toss 38 cutters (only one swinging strike), 29 sliders (two whiffs), 9 curves (1 swing and miss), a change-up, five four-seamers, and a two-seamer. That doesn't add up to 91 pitches and for whatever reason Gameday missed seven, but four hacks out of 84 pitches is 4.8%. Not good, not good at all.

It's one start back with issues.

I hope it stays that way.