ericmvan
Veteran
Supposed to be working on something more important
Posts: 8,881
|
Post by ericmvan on Sept 22, 2021 5:01:27 GMT -5
The only thing I don't love about eventual robo umps is the disappearance of catcher framing mattering whatsoever. There's something wrong with calling a strike on a pitch that misses the catcher's glove by 3 feet and has to stab back for it making it look absolutely nothing close to a strike, whether it technically was or wasn't. There is a pretty straightforward way to include quasi-pitch framing and intended location, by tracking catcher's glove movement and using it as a tie-breaker. The best tech will have a margin of error. A pitch that is 40% a strike and 60% a ball when you include the margin of error is a strike if it hits the catcher's glove and a ball if he has to lunge for it.
You tweak the algorithm that adds catcher glove movement to calculated pitch position position until it feels correct. I'd begin by saying that anything 67% with the catcher set up in the neighborhood ignores the catcher. As the catcher moves away, you need to be more certain ... a lunge from one side of the plate to the other requires 100% certainty. No glove movement at all until the ball is caught gets you a strike at 33%. At 50%, you just go by whether the catcher had below average movement or above.
In the same way, a pitch on a corner of the zone gets rounded off, because if it's 60% a strike at the knees and 60% on the black, that's a 36% strike and is a ball ... unless the catcher was set up there.
The crossups where the catcher lunges hugely but the pitch is easily in the zone are still strikes. The batter doesn't see the lunge.
|
|
|
Post by soxinsf on Sept 22, 2021 17:21:20 GMT -5
The only thing I don't love about eventual robo umps is the disappearance of catcher framing mattering whatsoever. There's something wrong with calling a strike on a pitch that misses the catcher's glove by 3 feet and has to stab back for it making it look absolutely nothing close to a strike, whether it technically was or wasn't. There is a pretty straightforward way to include quasi-pitch framing and intended location, by tracking catcher's glove movement and using it as a tie-breaker. The best tech will have a margin of error. A pitch that is 40% a strike and 60% a ball when you include the margin of error is a strike if it hits the catcher's glove and a ball if he has to lunge for it.
You tweak the algorithm that adds catcher glove movement to calculated pitch position position until it feels correct. I'd begin by saying that anything 67% with the catcher set up in the neighborhood ignores the catcher. As the catcher moves away, you need to be more certain ... a lunge from one side of the plate to the other requires 100% certainty. No glove movement at all until the ball is caught gets you a strike at 33%. At 50%, you just go by whether the catcher had below average movement or above.
In the same way, a pitch on a corner of the zone gets rounded off, because if it's 60% a strike at the knees and 60% on the black, that's a 36% strike and is a ball ... unless the catcher was set up there.
The crossups where the catcher lunges hugely but the pitch is easily in the zone are still strikes. The batter doesn't see the lunge.
I probably have missed your meaning, but I do not see how a pitch that is 60% on the black is anything but a strike. Same for a pitch that is 60% at the knees. It is like tennis. 90% out is 100 per cent in. But I may be missing something like you are talking about real world umps who allow themselves to be misled by catchers movements. However, I am doubly confused by the notion that your hypothetical 60/60 strike is really a 36% strike. Those situations should be additive, not subtractive. One could view that pitch as an 84% strike since both locations are strikes, i. e., 40% not a strike times 40% not yields 16% not a strike. Of course, that is why the world recognizes these kinds of discussions as lies, damned lies and statistics.
|
|
|
Post by jimed14 on Sept 22, 2021 17:42:37 GMT -5
The other obstacle is that the strike zone is not a 2-D square. The strike zone is a complicated 3-D shape since home plate is also not square. There are some actual strikes that will look much further off the plate with the 2-D square, especially on something like a Houck slider that goes around the plate and nicks it at the back. Plus I would not be absolutely certain that no mistakes will ever be made and that especially worry that they will tweak it continually to manipulate the game the way they want it.
So I imagine that fans and players will still complain regardless.
|
|
|
Post by foreverred9 on Sept 22, 2021 17:53:54 GMT -5
You wonder if and when robot umps come into play, will pitchers and teams throw all their focus into exploiting certain situations?
It reminds me of playing whiffle ball in the back yard with friends and having the chair as the strike zone. We would do our best to throw those ridiculous pitches than snuck in at the last minute and got the strike. The back-door sliders off the leg, the worm-burner that stayed low the whole way, the one that came in eye-high and dropped back into the chair.
That's probably what MLB needs to test out the most, because we know pitchers will try to throw those impossible strikes. The sinker that nicks the bottom front, the slider that drops in on the back corner, etc. They'll want to get rid of all the kinks possible before the implementation because the Luddites will be out in full force once it goes live.
|
|
|
Post by patford on Sept 22, 2021 18:16:03 GMT -5
The other obstacle is that the strike zone is not a 2-D square. The strike zone is a complicated 3-D shape since home plate is also not square. There are some actual strikes that will look much further off the plate with the 2-D square, especially on something like a Houck slider that goes around the plate and nicks it at the back. Plus I would not be absolutely certain that no mistakes will ever be made and that especially worry that they will tweak it continually to manipulate the game the way they want it. So I imagine that fans and players will still complain regardless. It's not a matter of mistakes as much as it's a matter of consistency. That is why I would favor even the crudest of strike zones like the superimposed rectangle used for television. When someone trows a dart the score is based on where the dart lands not on someone's interpretation of where the dart landed.
|
|