Quantifying Command – Part Three: Stealing Strikes

Series Introduction:

With the modern wave of information, every facet of a pitcher’s game can be quantified now better than ever. ‘Command’ – or the ability to locate a pitch – has remained elusive in being properly encapsulated. The goal of this series is to better analyze players’ command solely through the use of statcast data while providing insight on evaluating metrics along the way.

Part Three: Stealing Strikes

In 2015, Dallas Keuchel put on one of the best displays of command across a season in recent memory. With a fastball sitting at 90 miles per hour, he won the Cy Young award with an ERA of 2.48. He had a good walk rate (~80th percentile), but a below average edge percentage (~40th percentile). If those metrics do not recognize his command ability, what does?

In the early 2010s, one of the biggest revolutions in baseball analytics was in pitch framing. In essence, this is a catcher’s ability to present the ball in a way that makes it appear to have been in the strike zone. By doing so, elite pitch framers were able to buy their pitchers strikes that the umpire would have otherwise called balls.

By hitting their intended targets, elite command artists aid their catcher’s efforts to steal strikes. In 2015, Keuchel’s elite command bought him 168 called strikes out of the zone – the highest total in the statcast era (2015-present). No pitcher is infallible; Keuchel did lose 61 strikes that were in the zone, whether due to a missed location or an umpire having an unfavorable strikezone. 

Nevertheless, Keuchel benefitted from a net gain of 107 strikes that should have been balls. Additionally, these miscalls have pervasive effects on the rest of the pitcher’s game. With the strike zone effectively expanded, batters are forced to chase pitches farther off the plate, inducing ill-advised swings and weak contact.

For example, take this recent at bat between Sergio Romo and Josh Bell. On the first pitch – due to a good location, elite pitch framing, and/or questionable umpiring – a changeup clearly out of the zone is called a strike. Knowing an advantage had been established in that location, Romo proceeded to keep tossing chaneups to that low and away spot. Forced to respect the expanded zone, Bell chased two changeups he normally may not have. Shifting the count from 1-0 to 0-1 with a generous call had value in itself, but the true value was in the ripple effects. One call shifted the balance of the encounter, essentially ruining the at bat for Josh Bell.

Clearly, stealing strikes is valuable. Which players have championed doing so over the statcast era?

Along with topping the strikes stolen list, Dallas Keuchel set the highmark with net strikes stolen (strikes stolen – strikes lost) in 2015. Carlos Rodon represented the opposite extreme in his 2016 campaign where he produced at a league average rate despite elite stuff. Adam Ottavino was an interesting name on the Lowest Net Strikes Stolen list considering he is a relief pitcher. For him to have made this list, he needed to be losing strikes far quicker than the starting pitchers. To account for this, Net Strikes Stolen needs to be turned into a rate statistic. 

As expected, Ottavino’s 2017 season shoots to the top of the list with a bunch of other relievers. Being a rate statistic, it is subject to small sample size volatility; hence my 500 pitch minimum and the lack of starters. There is another interesting trend that exists within these seasons… 

Nine of the ten best strike stealing seasons were primarily paired with above average pitch framers (via Baseball Savant) while all the ten worst strike stealing seasons were paired with below average pitch framers. This makes sense as the pitcher-catcher battery is together responsible for stealing strikes. So how would the pitcher’s ability be isolated?

Similar to how framing metrics isolate for factors like the pitcher and umpire, the same would be done to find how many strikes the pitcher was adding with their command. So with Chris Bassitt and Jesse Chavez sharing the same primary catcher in 2015, it is safe to say they are on opposite ends of the spectrum due to command, not framing.

An astute reader will also note certain years produce more seasons at the extremes than others. 2015 and 2016 dominated the leaderboards in lost strikes while 2018 and 2019 seasons were lacking on most lists. The rate of missed calls has simply been declining over the statcast era.

Umpires have been getting better as they can now be objectively analyzed on success rates. There are fewer stolen strikes and even fewer lost strikes as umpires grow increasingly accurate with their calls. For proper year to year comparisons, the changing reality of the strike zone needs to be considered as well.

Creating this comprehensive web of adjustments to quantify a pitcher’s command was done on Baseball Prospectus in 2017. Harry Pavlidis, Jonathan Judge, and Jeff Long would estimate the probability of each pitch being called a strike, then credit or debit the pitcher based on whether they got the call or not. The resulting command metric, Called Strikes Above Average (CSAA), gave the called strike percentage the pitcher was adding with each pitch. Effectively, it is Net Strikes Per 100 but properly adjusted for other factors.

The problem is that getting granular requires a good sample size. It needs pitchers to be caught by different catchers, umpires to face different batteries, etc. If a battery is mutually inclusive – the catcher catches no one else and the pitcher is caught by no one else – isolating the effects of each cannot be done with trackman data alone. With this in mind, bringing a CSAA variant to the college game is complicated. Sample sizes are smaller and keeping track of all the variables across the conference could prove difficult.

Sticking to Net Strikes is a simpler approach. This type of analysis is discrete as opposed to continuous. From this perspective, a ball was either in the strike zone or it was not; the call was either right or wrong. As long as the ball tracking technology is calibrated and a strike zone can be estimated, Net Strikes Stolen can be tracked.

Stealing strikes may take two to tango, but it can be a symptom and powerful benefit of elite command. Whether it is a comprehensive CSAA-type metric or a Strikes Stolen counting statistic, there needs to be more information available to capture this effect at the lower levels of the game. It just might help find the next Dallas Keuchel.

One thought on “Quantifying Command – Part Three: Stealing Strikes

Add yours

  1. Nice analysis. So it appears that technology is impacting the game in many ways, including improving pitch calling ability of umps

Leave a Reply

Powered by WordPress.com.

Up ↑

%d bloggers like this: