So, about automating the strike zone …
There’s very little doubt in my mind that Major League Baseball’s umpires are the best in the world. Granted, there are minor-league umpires better than some major-league umpires; union rules keep relatively incompetent umpires from being weeded out in good order, so it’s far from a strict meritocracy. But these guys are really, really good at their jobs.
That said, their jobs are really, really difficult. Ever try to judge a 90-mile-an-hour pitch from the batter’s box? It’s really hard, especially if you don’t have superhuman eyesight. Granted, it’s a little easier from behind the plate. But few umpires have superhuman eyesight, plus sometimes the pitches are coming in at a hundred miles an hour rather than 90.
I’m just saying it’s not easy, back there. We should hardly be surprised when umpires miss by an inch or two, sometimes even three or four. And in fact, they are missing from time to time.
Recently, FiveThirtyEight published a tremendously interesting article by Etan Green about umpires’ tendencies. Granted, the results have been reported in other places. But if you like three-dimensional graphs, this is the article for you! Anyway, those well-confirmed results include this salient fact: Umpires, whether subconsciously or otherwise, don’t like to determine outcomes. We know this because given exactly the same pitch, location-wise, they’re less likely to call a third-strike than a second strike, or a fourth ball than a third ball.
These mistakes are frequent — pitchers tend to pitch to the borders of the official strike zone. And they are consequential — they happen in the most pivotal calls. When a 50/50 call becomes a 60/40 call, as it does with three balls, umpires are mistakenly calling strikes on 10 percent of borderline pitches. When a 50/50 call becomes a 30/70 call, as it does with two strikes, umpires are mistakenly calling balls on 20 percent of borderline pitches.
Major League Baseball has embraced technologies that are meant to make calls on the field more consistent. The league has long used pitch-tracking technology to encourage home-plate umpires to behave more like machines, evidently without complete success. This past offseason, the MLB extended replay review to cover essentially all umpire decisions — except ball and strike calls. Now as before, no justice will be served when a pitcher throws a strike and the umpire drops the ball.
Please keep that last little bit in the back of your mind. Please remember, too, that umpires are missing calls for other reasons. It’s just difficult to see the edges of the strike zone, especially considering how hard these guys throw … and of course they’re throwing harder all the time. Of course the umpires can and will adjust, just as hitters can and will adjust. But we also know that hitters are missing pitches more than ever. Doesn’t it follow that umpires will miss more than ever, too?
Earlier this week on this page, Rob Arthur wrote about the big changes in the strike zone over the last six years. Just last week, there was a related article at The Atlantic’s Website by Derek Thompson with an intriguing/inflammatory headline: “The simple technology that accidentally ruined baseball”…
Can’t anybody hit, these days?
Baseball has become too boring is one of those popular but unfalsifiable statements about sports, like Curling is stupid and Americans should watch more soccer. But if you like guys crossing the plate and balls falling beyond the outfield walls, baseball’s growing tedium is a straightforward observation. Both home runs-per-game and runs-per-game are down about 20 percent from their early-2000s highs. Strikeouts are up about a fifth. In the last decade, the number of players slugging higher than .500 has collapsed from 45 to 15.
Defense wins championships, as coaches like to say, but offense wins sweeps. Fans aren’t tuning in to the Era of the Pitcher. Across a range of networks including Fox’s Saturday game, TBS, and ESPN, ratings have fallen by between 25 and 35 percent since baseball’s collective hitting slump started in the middle of the last decade.
Harumph! Alas, it seems our friends at The Atlantic somehow missed Maury Brown’s story about the outstanding local TV ratings of late. There’s just no real evidence to suggest that “fans aren’t tuning in.” I’m not saying baseball in the Era of the Pitcher is the most attractive baseball. But it’s hardly driven the fans away.
Fortunately, one can quickly zip through the faulty economic analysis and arrive at some damn-fine baseball analysis, with more cool graphs and stuff. And Thompson’s point is the same as Rob Arthur’s: Since MLB started tracking pitches systematically and using this data to judge umpires, the effective strike zone has changed. It’s actually sort of amusing … You might remember how much everybody used to bitch about umpires not calling the high strikes. But that was probably an optical illusion, mostly. Because they were calling the high strikes. They just weren’t calling the low strikes. Why they weren’t, I don’t really know. Except it’s probably just harder to see those pitches, given an umpire’s eye level. Anyway, they’re calling the low strikes now. According to Thompson, “Since 2008, the strike zone has grown by about 30 square inches around a hitter’s knees… The entire increase in strikeouts is happening on pitches between 18 and 24 inches off the ground.”
But they didn’t stop calling the high strikes.
Yes, it’s easy to attribute the decline in power (and scoring) to drug testing, and nearly everyone (except maybe Joe Sheehan) believes drug testing has made a difference. But a) if the drugs were so important, shouldn’t pitchers’ be showing the effects of that, and b) should we assume that’s just completely coincidental that hitting started trending downward when the strike zone started changing?
If there’s one thing we should have learned from history, it’s that the strike zone can, and usually does, mean a great deal. As Arthur points out, all this is a tremendous example of the Law of Unintended Consequences. Everyone wanted the umpires to call the strike zone uniformly, and as it’s written in the book. As Joe Morgan used to say, “It’s not the umpire’s strike zone. It’s baseball’s strike zone.”
Or at least it’s supposed to be.
The umpires have improved. I think that’s clear. I believe they’re more consistent, and I know they’re calling a zone more in line with the rulebook. Which isn’t to say they’re perfect. They’re not close to perfect. They’re missing all those two-strike strikes and three-ball balls, which is psychological. They’re also missing a larger number of pitches on the side of the strike zone – especially with left-handed hitters, for some reason – and on the top, too. The effective strike zone still isn’t quite in line with the rulebook strike zone. All of which is really frustrating, especially when a bad call beats you.
Of course the solution seems obvious: Automate the strike zone. Here’s Joe Sheehan in a recent, typically hard-hitting newsletter:
It’s time to stop pretending that human eyes are the best available technology for the job. We have systems already in place that do a better job of calling a consistent, rulebook strike zone than human umpires could ever possibly do. Let them do the job. They don’t have to be perfect; they just have to correct for the biases that currently exist, and they have to not allow for swinging hundreds of runs a year based not on whether the pitch crossed the strike zone, but on what the catcher did after that happened.
This is no small thing. An inch in the strike zone, Bill James wrote, is worth 30 feet in the outfield. The de facto strike zone is, along with higher velocity, the key factor in today’s stagnant, strikeout-laden version of baseball. Wide strikes and low strikes are changing the game, unbalancing it, with very little end in sight.
The framing data is the way out of this mess. It demonstrates clearly that umpires are not calling balls and strikes according to the rules of the game, but rather based on the crutch of catcher actions. This isn’t out of laziness, out of a character flaw, out of a desire to bend the rules, but a concession to what has been true for decades: that human eyes cannot possibly track a baseball and render a decision on its position pursuant to the letter of Rule 2.00. Until recently, there wasn’t much that could be done. Now, with PitchF/X in place, indicting umpires every single day, we have both the data to make the case and the technology to do something better. An automated strike zone will be more fair to all the players, while putting an end to a condition in which virtually invisible movements are as valuable as the acrobatics of a Gold Glove shortstop.
Commissioner Manfred, here’s your first task: put automated ball-and-strike calling in place in time for the 2016 season.
I don’t believe the partisans of automation have thought through all the issues involved. I’ve thought of a few of them, and I’m not smart enough to solve them. Which isn’t to say they’re not solvable. I just think the notion that MLB could wave a wand and, voila, there’s a well-oiled system for calling a perfect strike zone without materially affecting the flow of the game is … reductive? Simplistic? Naïve? Wildly optimistic?
At least one of those, I think. But hey, everything is awesome and people are doing incredible things with transistors these days so let’s assume we can automate the strike zone in time for Opening Day in 2016.
Shouldn’t we figure out what we want the strike zone to look like first?
Because I can almost guarantee that we don’t want the effective strike zone to look like the rulebook strike zone. You think we’ve got too many strikeouts now? Joe does; he bitches about the #StrikeoutScourge even more than I do. For which I’m thankful. Without Joe, I would feel pretty lonely out here. With Joe, we just need someone in Chicago to join in, and then we’ll have 42nd Parallel covered.
But the current, effective strike zone is a) bigger than it used to be, but b) smaller than the rulebook strike zone. Not a lot smaller, since umpires are calling strikes on pitches outside the zone, especially with the left-handed hitters. But a bit smaller, especially since they’re still missing pitches on the corners. And we don’t need to run some sort of experiment to find out what happens when you make the strike zone bigger. We’ve been running that experiment for the last eight years, and we know what happens: more strikeouts, fewer walks, less scoring.
Then again, maybe the rulebook strike zone is exactly what we need. Baseball’s going to reach a tipping point, just as there was a tipping point in 1968, when Carl Yastrzemski led the American League with a .301 batting average and Bob Gibson finished with a 1.12 ERA. The graphs had been trending downward for years, but it took The Year of the Pitcher for Baseball to actually do something about it.
Baseball shouldn’t need another Year of the Pitcher to do something. I actually believe the people running the business side of things know exactly what’s happening. Which would have been enough in 1968. It’s not now. Now the players are involved, and the players don’t really care what’s going on. You think owners, with their five-year depreciation schemes, are short-sighted? Ha. The players are short-sighted and self-absorbed and everything else that keeps them from caring what happens to baseball in 2020.
Not to paint with too broad a brush. I’m sure they’re not all like that. But the union exists to protect current major leaguers’ jobs and salaries. Which is a tremendous argument for the status quo. Which is why nothing’s going to change without a tipping point that leads the RSN’s to say, “Hold on a minute there, pardners. We’re having a tough time selling these three-and-a-half-hour, 3-1 games. You guys really gotta get together and do something, or else the money-hose gets turned off."
I think Baseball should do something, and the sooner the better. If that means we get those games in 2016 rather than 2018, because the new automated strike zone is bigger than Petco Park, so be it.
That seems pretty optimistic, though. My first choice would be to automate the strike zone – again, assuming that it’s actually a practical thing – but only with a new strike zone, subject to adjustment in future seasons as data warrants. My second choice would be to leave things alone, and just wait for that inevitable tipping point. And my third choice is just to automate the strike zone without any real thought about those good ol unintended consequences.
I’ve seen that movie before. All it does is get me and Joe all worked up.