11/21/24
July 30, 2019
Robot umps now!: MLB to adopt in five years or less? (original WSJ paywalled article)
posted by billsaysthis to baseball at 11:18 AM - 10 comments
3) How long will it take for teams to be able to hack this system to benefit them during games? I mean, come on, this is a human-created system, so humans should be able to make changes to it. It's going to happen, you know it.
This is from memory so may be incorrect.
Because the top and the bottom of each batter's strike zone is different, information about the batter must be input into the system. While probably fractions of an inch, I could see players trying to gain the system by slouching, or wearing tall shoes, or something when getting measured.
There also has to be a person who selects the person at bat. Always a chance when a pinch hitter comes in that the wrong batter gets selected.
posted by prof at 01:52 PM on July 30, 2019
Not required. If an umpire is calling either loose or tight strike zones during a game he is consistent and hitters adjust. Pitching patterns are also adjusted based on the calls and are a big part of the game.
posted by cixelsyd at 04:12 PM on July 30, 2019
Would a future Weaver or Piniella send out an IT guy to argue the bad calls with a tablet?
Would there be any allowance built into the system for teams whose lineups are savages in the batter's box?
posted by beaverboard at 05:01 PM on July 30, 2019
@prof I would actually think the system uses realtime images to judge the strike zone borders, not previously-input batter info. Otherwise you don't account for the stance.
posted by billsaysthis at 05:24 PM on July 30, 2019
Oh my.
Technology is what I create to earn a living. Sports are an enjoyable diversion. Please keep it that way.
posted by cixelsyd at 11:42 PM on July 30, 2019
I'm sure that any SpoFites who have read some of my past commentaries on the fine art of umpiring have been waiting for this. You know I can't resist, so here goes.
The strike zone is a personal thing with every umpire. Some call the high strike, and some call a bit below the knees. Some give a little bit on the outside, some a little inside. The important thing is that all strive to be consistent from pitch to pitch. So what causes the inconsistencies in an umpire's zone? The most common cause is that the umpire is not setting up at the same place on every pitch. If on the first pitch of a game you were to put marks where the umpire's feet were, the ideal would be to have your feet exactly on those marks for each subsequent pitch. The position for a left handed batter is different than that for right handers, but the result would be horizontal symmetry around a vertical line through the point of the plate. I remember some time ago a discussion we had about a video in which two consecutive pitches in what appeared to be the exact same spot were called a ball in one case and a strike in the other. Looking carefully at the video you could see that the umpire's head position was just an inch or two different on the pitches, thus the difference in the calls. So on very close pitches that are called inconsistently, position is important.
What else goes into calling balls and strikes? Preparation is one thing. In the majors and high minors a plate umpire will have a "book" on the pitchers he is likely to see in a game. He will know what they feature, what their release looks like, and what the rotation on the ball will do. Even this is not enough to avoid being fooled on the occasional pitch. My rough analysis from watching baseball on TV over the years is that the most missed calls come on breaking balls. If you are not expecting one, it is awfully easy to miss it. Another problem for umpires is catchers that move around a lot. Most umpires set up on "the block", that is you are in a position where it looks like you are almost ready to sit on a block of wood, and you set up in the slot between the catcher and the batter. You might have a good view of the plate, but if the catcher moves to field the ball, he might block your view. Usually this will result in a strike being taken away from the pitcher even though the pitch was in the zone.
One thing that was said in the article (not the original WSJ piece) was that the umpire has too many other things to do that interfere with his calling balls and strikes. Whoever wrote this has no understanding of umpire mechanics. Once the pitcher is committed to delivering the pitch, the plate umpire has absolutely nothing to do other than to call a ball or strike, swing or no swing. Think about it for a moment. With the ball about to leave the pitcher's hand or in the air on its way to the catcher what else can happen? Runners may be going, but no play can be made on them until someone has the ball; fielders may be adjusting their positions, but as long as they are not attempting to distract the hitter, they may move. Yes, the plate umpire has a lot to do during the game, but at the time of the pitch his only job is to call balls and strikes.
The device used to determine ball or strike in the Atlantic League experiment was a Doppler radar. This type of radar will ignore stationary or slow moving targets and show only targets moving above a set threshold velocity. I spent nearly 4 years in Egypt writing a maintenance manual for a mobile pulse Doppler air defense radar that we had sold to the Egyptian Army Air Defense Command. The accuracy of this one was on the order of a few meters at several kilometers range. More than good enough if you are trying to put a missile with its own terminal guidance system close enough to knock down an aircraft, but hardly good enough to measure the difference between a fraction of an inch outside or "on the black" for a strike. I have heard of (and seen) some Doppler sets that work at frequencies well up into the 18 to 22 GHz range, and these could easily have much more precision than our old stuff, but I am not familiar enough with them to discuss. I would like to see the technical details of antenna placement. scan type and rate, frequency range, and so on. The antenna would need to be placed where its "view" of the ball is unobstructed, and preferably on the line between mound and plate. High and behind the plate would be the best location. A better choice of tracking device might be a laser tracker. These can give location data with accuracy on the order of millimeters. The problem with these is that they must have the object to be tracked identified and in their field of view returning the laser energy to the tracker. This can be difficult to accomplish in the few milliseconds between the time the pitcher pulls the ball out of his glove and it leaves his hand.
Let's get back to the strike zone for a bit. I'm sure it is one of the problems with automated ball - strike detection that the vertical limits vary with each batter. You can't put in a one-size-fits-all approach. How that problem was dealt with in the Atlantic League experiment would make for some interesting reading. I also know that each umpire goes into a game with a good idea of where his zone will be. This will last until he makes a call on the first tight one. At that point he will have to figure out whether that will be the limit of his zone or whether he should have given a little more to the pitcher or a little more to the batter, or whether it was just a mistake. The big thing to remember for any umpire at any level is to be consistent. If you do blow a call, admit it. Let the catcher or batter know what you missed and the fact that the same pitch will be called differently next time. This happens in the majors more frequently than one might think, and if the umpires and players are honest about it, there are not many problems. Try not to blow a call on a "decision" pitch (with 3 balls or 2 strikes on a hitter), for this will put someone on base or in the dugout that might not deserve to be there.
So how many pitches does an umpire miss during a game? The MLB umpires will admit to missing 4 or 5 in a game, and that's about right. They do not consider the really close ones to be missed calls, nor should they. Many years ago, when I was fairly new to umpiring, I was calling a JV game between two of the smaller schools in NH. The coach of the visiting team was obviously new, and I think he got the job because he admitted he had been to a baseball game once. The coach of the home team had been around for a while, and he eventually became the highly respected and successful coach of the varsity. Between innings, about 3 or 4 innings into the game, the visiting coach, while on his way to the 3rd base coach's box, stopped and commented that my strike zone wasn't too good. I asked hi how many I had missed. He said that I had missed 4 or 5. I loudly exclaimed, "That's wonderful; it's way better than I usually do." The look on his face was priceless, and the home coach standing in the 3rd base dugout was nearly crippled from suppressing his laughter. Keeping a game moving without constant griping and arguing is sometimes best done by maintaining a sense of humor and admitting your mistakes.
posted by Howard_T at 12:12 AM on July 31, 2019
/all the high-fives to Howard
posted by NoMich at 07:40 AM on July 31, 2019
Awesome, Howard.
As a coach and athlete I have a real appreciation for officials who understand their importance to the game .. even the ones I sometimes get into.
There should be greater focus on the importance of officials: more training, more incentive to become one, more desire to be the best. I'll accept an occasional bad call from a good official to a robocall any day. And technology is not always without it's flaws for various reasons .. many billions of dollars in grounded aircraft back my angle on that comment.
posted by cixelsyd at 10:15 PM on July 31, 2019
That was one hell of an explanation Howard, thanks for the effort! And remember a few weeks back when I kicked dirt on your shoes? I'm sorry about that. Keep up the good work.
posted by tahoemoj at 12:53 PM on August 01, 2019
Bummer that I'm not able to read the WSJ article that this link refers to, because I have questions.
1) Watching baseball for the last few years has shown me that home plate umps usually get the ball/strike call correct. When they yell out a call that does not match what we saw on TV (you know, in the little vertical rectangular box that they superimpose on the screen that indicates the strike zone), it's usually just barely off. So, what's the reality? Has someone actually quantified this, and if so, how bad is for the humans? Like, what's the percentage of incorrect calls?
2) From the first article:
The sooner Major League Baseball embraces the "robot ump", the better it will be for everyone, including the umpires. Home-plate umpires have lots of other tasks to do"and now they won't catch grief from players or their own bosses for the mistakes they make doing an essentially impossible job.
I think the home plate ump got grief from a player and/or manager during the first game that this technology was rolled out. The batter or manager disagreed with the computer's call on that pitch and started yelling at the ump to overturn the call. Umpires will never be able to escape the wrath of petulant players/managers.
3) How long will it take for teams to be able to hack this system to benefit them during games? I mean, come on, this is a human-created system, so humans should be able to make changes to it. It's going to happen, you know it.