The apocalypse is nigh: robo umps are coming to AAA and not just to eat the flesh of your deadest prospects. They also plan to adjudicate balls and strikes by assigning an actual rectangle to the space over home plate, and you may or may not have a couple opportunities to challenge calls that appear to come from the defective “Robo Angel Hernandez 3” shipment.
Robo umps are a lot like MLB prospects: once they reach AAA you know they are but a step away from the big leagues. Is the jump from A to AAA harder for a robot than it is for a living and breathing athlete? We don’t really know because available data is lacking. But if you want to know where to put your smart money, odds are 11010001 to 0101001 that robo umps will perform about the same at any level owing to the fact that they are machines.
With an automated strike zone now just a step away from the big leagues, debates can resume raging as to whether MLB robo umps for balls and strikes would be a good or bad development.
Replay already has polarized fans, who range pretty much from 0 to 100 on the “hate it” to “love it” continuum. It’s worth noting that I fall far on the left end, having disliked the idea of replay since before it was implemented and believing that it has done more harm than good since being introduced.
So clearly I am not a “get it right at all costs” type of fan. I think if you are going to have replay at all, it should only be to challenge calls that are egregiously bad — which should be easily confirmed in a few seconds looking at most any angle. I also think challenges should either have to be issued immediately (probably by the player, or else instantly from the bench based on the real time naked eye view), or else requested by the umpire himself when he questions the accuracy of his own call and agrees it should be checked.
No, I don’t care if a team gets gypped because the ball beat the runner to 2B but the tag came 1/100th of a second after the first cleat nipped the corner of the bag, just as I don’t think a runner is out just because there is a split second off the bag as he pops up from his slide. Human error is an inherent, and often even interesting, part of the game and I don’t need to wait 90 seconds to see if my team’s walk-off single will hold up in robo-court.
So then why would I support an automated strike zone when the data suggests umpires are roughly 93.5% accurate calling balls and strikes?
1. Accuracy is not actually high enough
93.5% sounds pretty good in a vacuum. You wouldn’t mind playing a lottery in which you had a 93.5% of scratching a winning ticket. Then again you probably would mind drinking a glass of milk that was 93.5% free of bacteria.
While umpires do get the great majority of ball/strike calls right, this excellent in depth article from 2018 notes that those 6.5% of pitches missed translate to 34,294 pitches a year, which reflects 14 per game, 1.6 per inning.
Imagine if umps were missing about 1.6 calls on the bases each inning, or 14 balls a game were called fair or foul incorrectly. In a game where over 200 pitches are thrown, that 6.5% error rate shows up too often.
More importantly, remember that most pitches are easy to call. Some hit the center of the strike zone and many are in the dirt or a foot outside. When you reduce the pitches in a game to the borderline ones, and consider that nearly all the missed calls come from that subset, it’s not as if umps are getting about 93.5% of the close ones right.
Maybe umps are still nailing the close pitches 80%-85% of the time, but a 15%-20% error rate on the calls that most matter becomes a real problem. So the status quo isn’t really working.
2. Umpires are out of position
This is a hill I will die on even if every fan and reader resists and I stand alone in saying it: you can’t call balls and strikes most effectively from behind the catcher. From that vantage point, you are overly dependent on using the catcher’s glove to gauge where a pitch is, which is why umps are so vulnerable to being deked by good framing.
Keep in mind that from behind the catcher, an umpire quite literally loses sight of the ball at the end of its flight because the catcher’s body and glove are in the way. It is precisely when the ball enters that invisible rectangle that defines the strike zone that the umpire’s view of the pitch is obstructed.
As a result of seeing the catcher’s glove more than the ball, at the critical moment a pitch becomes a ball or a strike, it is difficult for an umpire to avoid being fooled by framing, confused by a catcher who has to reach for a pitch that misses its intended location, or misled by a pitch the catcher fails to receive cleanly. It is also hard to assess height and width at the same time — especially width when the catcher’s glove is many times wider than the ball.
Compare that to the view from behind the mound. This is the view fans enjoy from the CF camera when watching on TV. I am thoroughly convinced that an umpire standing on the grass between the mound and 2B, even if off-center a smidgen to clear the pitcher (the eye quickly learns to calibrate for this angle), would call balls and strikes with an accuracy that cut the “error rate” at least in half.
That’s because from the “CF angle” you get a clear view of the ball itself as it crosses the plate. Low pitches pass the “eyeball test” as to how much space is between the catcher’s glove and the plate at the instant the ball gets to the the glove (before it can be framed). It’s just very possible to accurately compare the ball to the invisible rectangle, when it crosses the plate, from that vantage point.
Cutting the error rate at least in half would mean a 96.75% success rate or higher, and now you’re in business. But positioned where they are I don’t think umpires have a fair chance and so it’s not reasonable to ask them to improve as much as they need to. So unless MLB is prepared for the radical move of either repositioning the “home plate umpire” or having the 2B umpire call balls and strikes, umpires are going to be stuck being fooled too easily and missing too many calls.
3. A foolish consistency is a good thing
Pitchers and hitters say it all the time: “I don’t care what the strike zone is, just be consistent about it.” When the same pitch is a ball one time and then a strike another time, it’s tantamount to saying, “Yeah the last ball that landed on the chalk was fair, but this time a ball on the chalk is foul.”
When one umpire is known for having “a low strike zone,” another is considered a “hitter’s umpire” while his partner is more reputed to “hunt for strikes,” and every umpire in a crew has a different idea of where the outside corner ends, how is this good for the game?
Close calls at 1B are important, but they don’t come up every single pitch. Balls and strikes do, so if there is a complete lack of consistency as to what is a ball and what is a strike, this is not enhancing the game and keeping the focus primarily on the players.
If balls and strikes calling needs one thing, it’s consistency and to a fault that’s what robo umps can offer. (Get ready for jaws to drop when a pitch with terrific late movement nearly bounces and is called a strike by the robo ump.) Robo umps don’t “interpret” a strike zone; they note when it is and isn’t found by the ball.
So maybe just in the one area that impacts every single pitch, we need consistency over “the human element” — an element that will still show up plenty thanks to all the bang-bang plays at 1B, tag plays on the bases, and the almighty balk.
My bottom line: so long as the human umps can report the news in real time, which might be the hardest part of the system to perfect, I think robo umps will be better for the game than the middle aged man with 20/30 vision trying to guess what the ball did just as his left his view and became a cleverly moving catcher’s glove.
Bring on the apocalypse!