With the dawn of the 2014 season came also the dawn of comprehensive instant replay in Major League Baseball. Despite some trials, tribulations, and the sometimes clumsy application of needlessly complex replay rules, more plays are ultimately being called correctly this year than in the past.
Getting the calls right. That’s indisputably a good thing, isn’t it? With all due respect to those who cast blown calls in a positive light by evoking the term “human element,” I tend to think that any game is made better by eliminating variables external to the folks playing it. Instant replay takes a step in this direction, and I welcome it with open arms.
A chasm nevertheless remains. It remains in the form of miscalled balls and strikes. The fundamental unit of a baseball game – along with its box score – is a pitch. And in those cases where an umpire is wholly responsible for establishing the truth and consequence of a pitched ball – that is, when a batter chooses not to swing at it – the umpire has already gotten the call wrong a staggering 10,753 times through May 12th. This is especially troubling to me since the technology already exists to reduce that number to zero – that’s how we know what the number is in the first place.
However, rather than delivering a sermon in favor of RoboUmps, I thought I’d take a look at how this imprecise umpiring has affected Colorado Rockies pitchers in particular thus far in 2014. As it turns out, my journey through the depths of ESPN Stats and Info both confirmed my worst fears and also undermined my rage. Let’s see how.
Through Pitch f/x, a pitch-tracking system installed in every major league park, we know the precise location of every pitch thrown in every game since 2006. I looked at the data for each Rockies pitcher who has thrown at least 10 innings and 100 pitches this year. I also pulled the data for all pitches thrown by all players league-wide to serve as a baseline. We’re talking about a massive amount of statistical information – information that can sliced and diced in countless ways. Here are my cuts:
- Pitches in and out of the strike zone. These are “true” strikes and balls, not what umpires think are balls and strikes because of Tom Glavine’s hypnosis.
- Pitches swung at and pitches taken. When it comes to pitches that are swung at, it’s essentially the hitter who calls balls and strikes. So, Vladimir Guerrero called essentially all of his own pitches.
- For pitches taken, was the umpire’s call wrong or right? A good call is one in which an in-the-zone pitch is called a strike, and an out-of-the-zone pitch is called a ball. Everything else is a bad call. There are certain immutable laws in the universe, and the strike zone is one of them, no matter what Angel Hernandez thinks.
- Percentage of all pitches thrown that were called incorrectly. For in-the-zone pitches called balls, this is a measure of undeserved harm done to the pitcher. For out-of-the-zone pitches called strikes, this is measure of undeserved help to the pitcher. Yes, I hereby deem these calls “undeserved;” no debates on the sanctity of pitch-framing allowed here. This is a Jonathan Lucroy-free zone.
- Among pitches taken by the hitter (not swung at), percentage called incorrectly. This has the same numerator as the split above, but a different denominator. This one does a better job showing the umpires’ error rate on just those pitches that were the umpires’ sole responsibly to judge. This is as good an opportunity as any to note that I don’t blame umpires for any of this. I’m actually surprised they get as many of these calls right as the do. These MLB-level pitches move fast and crooked and I’m pretty sure that if I were to call balls and strikes for someone like Jose Fernandez, my error rate would be near 100% due to the fact that I’d been face down in the dirt crying and shaking in fear.
Keep in mind that while I maintain that all bad calls are evil out of principal, the umpires’ mistakes can both “help” a pitcher and “hurt” him. It helps, of course, when an umpire calls a pitch taken by the batter out of the zone a strike. Got it? Good. Here’s the data dump – provided in two tables: bad calls that “hurt” and bad calls that “help.” Each table is sorted by the rate of bad calls. If you’re a pitcher, you want to be at the bottom of the “hurt” list and at the top of the “help” list.
|True Strikes Called Balls (The Hurt)|
|Name||Total Pitches||Pitches Taken In Zone||Taken in Zone – Called Strikes (Good Calls)||Taken in Zone – Called Balls (Bad Calls)||% of All Pitches in Zone Called Incorrectly (Hurt Rate to Pitchers)||Among Taken Pitches, % Called Incorrectly (Umpire Error Rate)|
|Boone Logan||83||33||24||9||10.84%||27.27%||This man is building a RoboUmp Right now|
|Tyler Chatwood||156||56||45||11||7.05%||19.64%||Deep breaths…karma is real, right?|
|Jorge De La Rosa||352||117||93||24||6.82%||20.51%|
|All Pitches||82198||29403||24238||5165||6.28%||17.57%||Typical Umpire Screwing|
|Juan Nicasio||359||126||108||18||5.01%||14.29%||More than a little irritating|
|Jhoulys Chacin||79||29||28||1||1.27%||3.45%||Sigh… Whatever|
|True Balls Called Strikes (The Help)|
|Name||Total Pitches||Pitchens Taken Out of Zone||Taken Out of Zone – Called Balls (Good Calls)||Taken Out of Zone – Called Strikes (Bad Calls)||% of All Pitches Out of Zone Called Incorrectly (Help Rate to Pitchers)||Among Taken Pitches, % Called Incorrectly (Umpire Error Rate)|
|Matt Belisle||101||86||72||14||13.86%||16.28%||Thou shall not covet thy teammate’s umpire|
|Jordan Lyles||411||295||263||32||7.79%||10.85%||Earns this luck by saving kittens in free time.|
|All Pitches||85727||61933||56345||5588||6.52%||9.00%||Typical charity (not tax-deductible)|
|Boone Logan||94||62||57||5||5.32%||8.07%||Umm… thanks?|
|Jorge De La Rosa||415||300||282||18||4.34%||6.00%|
|LaTroy Hawkins||85||56||54||2||2.35%||3.57%||Well, at least his strikes are earned.|
When I saw the numbers for the first time, I couldn’t decide if I thought they were small enough to shrug off, or big enough to consider an outright sporting tragedy. I had both gut reactions almost simultaneously. On the one hand, we’re talking about an average of only about 6% of all pitches not meeting the fate they deserve. On the other hand, that’s equivalent to a dozen or so pitches a game. How many of those might have made a real difference in the outcome? And an error rate suggesting that the umpires miss almost one in every five pitches taken in the strike zone?! Alright, now I’m good and angry again.
Note also that the league-wide “help” rate is a touch higher than the “hurt” rate. The net effect of bad calls has actually benefited pitchers. I like to think of this as the quantifiable manifestation of a league-wide “Glavine Effect.” Most of us believe intuitively that this extra bit of strike zone exists for certain pitchers and/or in certain situations. The actual data seems consistent with this.
These are general observations. What about the Rockies in particular? Well, they have a fair number of pitchers both above and below each league-wide mid-point, with a few more above the line on the “hurt” scale. However, when total pitch volume is considered, the Rockies appear to be getting both “hurt” and “helped” a bit more than average. Let’s rearrange the tables to get a better sense of net effects. This one is sorted by most “helped” to most “hurt” after combining both effects).
|The Hurt||The Help||The Net Effect|
|Name||Total Pitches||Total Pitches Taken||Taken in Zone – Called Balls (Bad Calls)||% of All Pitches in Zone Called Incorrectly (Hurt Rate to Pitchers)||Taken Out of Zone – Called Strikes (Bad Calls)||% of All Pitches Out of Zone Called Incorrectly (Help Rate to Pitchers)||Total Blown Calls||Total Blown Calls (%)||Net Help (Pitches)|
|Jorge De La Rosa||767||417||24||6.82%||18||4.34%||42||5.48%||-6|
Maybe this is why Jorge De La Rosa’s been so upset this year, because he’s got the largest karma deficit on the team so far? And is this why Brett Anderson broke again, weakened due to the emotional roller coaster he’s been on (highest overall blown call rate)?
Probably not. Jorge’s likely too busy tinkering with his Wilin Rosario voodoo doll to dwell on being six pitches in the hole, and he probably thinks it’s more like 106 pitches anyway. I’m guessing most pitchers do. And if you read Brett Anderson’s twitter feed, you know his mood stays pretty dang chipper 25 hours a day and 8 days a week.
It was after I calculated these net effects that my anger began to subside. Or at least evolve into something more like resignation. As much as I want ball/strike perfection, and as much as I want it now, there is certainly credence to the idea that, given large enough sample sizes, these sorts of things tend to even out over time. I’d still prefer a smoother ride to equilibrium, fewer Ump Shows, and less benefit conferred to pitchers who happen to know Jedi mind tricks. But when the dust settles on a long season, all of this ball/strike mismanagement likely regresses to the mean. Math wins, I guess. As for my little study at the season’s Quarter Pole? Behold Math’s statistical coup de grâce:
|Total Blown Calls||Net Help (Pitches)|
|All Rockies in Study||351||1|
One pitch. One pitch to the good. I couldn’t possibility be more annoyed – even if the net effect was literally nothing.
I should have just gone with the sermon.