Why we need computers calling balls and strikes

[AtomicDumpling]

Originally Posted by Brutus

There's nothing to acknowledge. Again, they were measuring and attempting to do two different things. Sorry but it's clear to me that you did not understand the QuesTech system or how it was operated.

By the way, where is this "proof" that it's more effective and accurate? Where are these definitive, conclusive studies that show it is a more accurate system?

The only reason it's being used is because it does more things. MLB didn't go to it because it was more "accurate" at calling balls and strikes. Again, that narrative is only being used if you don't understand what it was the two systems were built for. They went to it because it could do other things beyond grading umpires.

Right, they decided to go with an expensive new system but it wasn't an upgrade. OK. Makes sense.

You are the one that doesn't seem to understand MLB's system, after all you are the one that was quoting the old system's numbers until I corrected you. Now you want to pretend to be an expert on the technology. You aren't.

Feel free to pretend the QuesTech system is the best system available and keep using old QuesTech stats to show why using technology to call balls and strikes is inadequate. You can use 2002-2007 figures while the rest of use the 2012 Pitch F/X numbers.

It is clear that the technology has improved a great deal over the last decade and every indication is that it will continue to improve. This means the league has the tools they need at their disposal to make the switch to robo-umps if they decide to make the move. It is not accurate to claim that it can't happen because the technology is not up to the task. If MLB wants it to happen a highly accurate system could be implemented very quickly.

[M2]

Originally Posted by George Anderson

I also thought that if we are going to go your route and use technology instead of humans we should just go all the way and use it on the field also. I didn't realize that an umpire missing a ball 2 inches off the plate was more important than a missed play at the plate.

No, you're just arguing to argue at this point. Ball/strike calls can be automated. Plays at the plate not so much. It's not that hard a concept to grasp.

[George Anderson]

Originally Posted by M2

Ball/strike calls can be automated..

They can huh? Please show me the evidence where the system has been installed and has actually been used in a game without fault?

Originally Posted by M2

Plays at the plate not so much. .

Uh yes, there are these things called cameras that have been used for sometime in other sports to overturn calls and get the play right. You however want to ignore this function of technology but instead install a controversial technology system that has yet to be been tried at any level of professional or amateur baseball.

What was that comment about being absurd again?

[gonelong]

I think everyone knows that eventually we will have automated balls/strikes calls. There is no argument to be made that will win the day once the technology can be proven to be more reliable than human umpiring.

It's only a matter of when and how so I'm of the opinion that we should get started sooner rather than later as the technology exists today to at least give it consideration.

GL

[dougdirt]

Originally Posted by gonelong

I think everyone knows that eventually we will have automated balls/strikes calls. There is no argument to be made that will win the day once the technology can be proven to be more reliable than human umpiring.

It's only a matter of when and how so I'm of the opinion that we should get started sooner rather than later as the technology exists today to at least give it consideration.

GL

It has been proven to be more reliable than human umpiring on balls and strikes. Emphatically. Both Questek in the past and now the current Pitch F/X are used to grade the umpires. If it weren't better than the umpires are, MLB wouldn't use it to determine how good/bad the umpires are doing their job.

[marcshoe]

I think it's been evident for decades that baseball needs to move in this direction. There is no reason to not use a better system when one is available.

[Roy Tucker]

So, I had a few spare moments so I thought I'd think this through. It’s a pretty interesting technical challenge. The geek in me likes to figure these things out.

Right now, we have an umpire who is a living breathing and more importantly trained and experienced person who judges these things. We’re either augmenting or replacing this person and all of what goes through his head between when the ball leaves the pitchers hand and smacks into the catcher’s glove.

The limitations of this person is his visual system and his ability to gauge where the ball is in 3D space and if that crosses the strike zone (in his opinion). We need to be able to improve upon that.

It certainly seems that existing video technology can accurately measure where the ball goes in xyz space. They seem to have that nailed pretty good. So we do know where the ball is (or was).

Next, you need to determine where is the strike zone. Thinking it through, it’s a 5 sided box (5 sides, top and bottom). If the ball touches or crosses any part of that geometric shape, it’s a strike. Again, if you have that strike zone determined at pitch time, existing systems can tell if it crossed it. A question in my mind is can it do it in real time like an umpire does. That is, can it call it a ball or strike in about a second. That’s a pretty responsive system and presents its own challenges.

Then we start getting into the harder stuff. We need this system to be able to detect where the strike zone is on a batter by batter, pitch by pitch basis. Each batter is different and can adjust his stance on a pitch by pitch basis. So we need to be able to accomodate that. I think that crosses off any pre-configuration of a strike zone for a batter, e.g. before a season or a game, the batters strike zone is measured and he is stuck with that for that game, series, month, or season. It needs to be done on each pitch in ral time.

MLB Rule 2.00

The STRIKE ZONE is that area over home plate the upper limit of which is a horizontal line at the midpoint between the top of the shoulders and the top of the uniform pants, and the lower level is a line at the hollow beneath the kneecap. The Strike Zone shall be determined from the batter's stance as the batter is prepared to swing at a pitched ball.

So, “over home plate”. That’s the front, sides, and backs of the plate. They don’t move so that part is easy to define. What’s harder is “upper limit of which is a horizontal line at the midpoint between the top of the shoulders and the top of the uniform pants, and the lower level is a line at the hollow beneath the kneecap. The Strike Zone shall be determined from the batter's stance as the batter is prepared to swing at a pitched ball”. How do you do this? Its a living, breathing, and shifting thing.

The two basic ways I can think of this happening is:

1.) A human operator sets it “when the batter is prepared to swing at a pitched ball”. So this would hopefully be a trained umpire up in the booth armed with a video monitor and a mouse, and he clicks the top and bottom of the zone as the batter assumes his stance (or some method like that). This isn’t a perfect way to do it since we again are asking a human to do this (and is it better than a home plate ump?) but if this system ever sees daylight, this would probably be an improvement. How much improvement would need to be determined. His hand may slip, he might prematurely click the mouse and try to correct and not make it in time. his attention may wander, he might cough or sneeze, or some other dumb little thing will happen over a 300+ pitch game.

2.) The computer or video system automatically does it. Hmmm. We do have a lot of video facial recognition, optical character recognition, and other sophisticated video technology so this is feasible and probably solvable. But, I don’t think any existing systems do this now (but I could be wrong, I’m just talking out my butt here). This would be a tricky thing to do and batters may soon figure out ways to trick or game the system. Humans are remarkably adaptive creatures and the problem with systems like this are the unintended consequences. So the software would need to constantly improved to be able to adapt to funny things.

But anyhow, the video/compute computes it’s a ball or strike using some technique.

Next, how do we signal it? Is it just a green or red light on the scoreboard? Is it a little earbud in the home plate umps ear? The baseball field is a relatively hostile environment so we need to take that into consideration. The ump might get nailed in the head and dislodge the earbud and he didn’t know it. He might sweat like a pig and short out his battery pack. The beep may not register in his brain. Do we allow the ump to override the system? Is the beep in his ear just a suggestion or is it mandatory?

We also need to make sure this system is fault tolerant. No inopportune computer hiccups or freezes, the video cameras have to not waver in the wind or be affected by atmospheric conditions or getting bumped by operators or the crowd or the stadium wavering. All of the parts of this system have to have completely reliable communication. No video glitches, no software loops, no nothing. And we need a pretty decently sized staff to operate and maintain these systems in all MLB parks.

That’s just off the top of my head. There are 1000’s more questions I’m sure I could drill down to. In the biz, we call them “proof of concepts”. That is, let’s mock up the system and see it works. So that’s what I’d like to see. Try it in real life.

And, really, the biggest question is whether or not MLB is willing to do something like this. I don’t know what would cause that to happen. Some fan groundswell, some massive failure in a critical game, *something* that would cause Bud and his pals to go “hey, let’s do that”. And also, after this system is developed and implemented, what is the impact on the game on the field and is this what we really want?

[kaldaniels]

Originally Posted by Roy Tucker

So, I had a few spare moments so I thought I'd think this through. It’s a pretty interesting technical challenge. The geek in me likes to figure these things out.

Right now, we have an umpire who is a living breathing and more importantly trained and experienced person who judges these things. We’re either augmenting or replacing this person and all of what goes through his head between when the ball leaves the pitchers hand and smacks into the catcher’s glove.

The limitations of this person is his visual system and his ability to gauge where the ball is in 3D space and if that crosses the strike zone (in his opinion). We need to be able to improve upon that.

It certainly seems that existing video technology can accurately measure where the ball goes in xyz space. They seem to have that nailed pretty good. So we do know where the ball is (or was).

Next, you need to determine where is the strike zone. Thinking it through, it’s a 5 sided box (5 sides, top and bottom). If the ball touches or crosses any part of that geometric shape, it’s a strike. Again, if you have that strike zone determined at pitch time, existing systems can tell if it crossed it. A question in my mind is can it do it in real time like an umpire does. That is, can it call it a ball or strike in about a second. That’s a pretty responsive system and presents its own challenges.

Then we start getting into the harder stuff. We need this system to be able to detect where the strike zone is on a batter by batter, pitch by pitch basis. Each batter is different and can adjust his stance on a pitch by pitch basis. So we need to be able to accomodate that. I think that crosses off any pre-configuration of a strike zone for a batter, e.g. before a season or a game, the batters strike zone is measured and he is stuck with that for that game, series, month, or season. It needs to be done on each pitch in ral time.

MLB Rule 2.00


So, “over home plate”. That’s the front, sides, and backs of the plate. They don’t move so that part is easy to define. What’s harder is “upper limit of which is a horizontal line at the midpoint between the top of the shoulders and the top of the uniform pants, and the lower level is a line at the hollow beneath the kneecap. The Strike Zone shall be determined from the batter's stance as the batter is prepared to swing at a pitched ball”. How do you do this? Its a living, breathing, and shifting thing.

The two basic ways I can think of this happening is:

1.) A human operator sets it “when the batter is prepared to swing at a pitched ball”. So this would hopefully be a trained umpire up in the booth armed with a video monitor and a mouse, and he clicks the top and bottom of the zone as the batter assumes his stance (or some method like that). This isn’t a perfect way to do it since we again are asking a human to do this (and is it better than a home plate ump?) but if this system ever sees daylight, this would probably be an improvement. How much improvement would need to be determined. His hand may slip, he might prematurely click the mouse and try to correct and not make it in time. his attention may wander, he might cough or sneeze, or some other dumb little thing will happen over a 300+ pitch game.

2.) The computer or video system automatically does it. Hmmm. We do have a lot of video facial recognition, optical character recognition, and other sophisticated video technology so this is feasible and probably solvable. But, I don’t think any existing systems do this now (but I could be wrong, I’m just talking out my butt here). This would be a tricky thing to do and batters may soon figure out ways to trick or game the system. Humans are remarkably adaptive creatures and the problem with systems like this are the unintended consequences. So the software would need to constantly improved to be able to adapt to funny things.

But anyhow, the video/compute computes it’s a ball or strike using some technique.

Next, how do we signal it? Is it just a green or red light on the scoreboard? Is it a little earbud in the home plate umps ear? The baseball field is a relatively hostile environment so we need to take that into consideration. The ump might get nailed in the head and dislodge the earbud and he didn’t know it. He might sweat like a pig and short out his battery pack. The beep may not register in his brain. Do we allow the ump to override the system? Is the beep in his ear just a suggestion or is it mandatory?

We also need to make sure this system is fault tolerant. No inopportune computer hiccups or freezes, the video cameras have to not waver in the wind or be affected by atmospheric conditions or getting bumped by operators or the crowd or the stadium wavering. All of the parts of this system have to have completely reliable communication. No video glitches, no software loops, no nothing. And we need a pretty decently sized staff to operate and maintain these systems in all MLB parks.

That’s just off the top of my head. There are 1000’s more questions I’m sure I could drill down to. In the biz, we call them “proof of concepts”. That is, let’s mock up the system and see it works. So that’s what I’d like to see. Try it in real life.

And, really, the biggest question is whether or not MLB is willing to do something like this. I don’t know what would cause that to happen. Some fan groundswell, some massive failure in a critical game, *something* that would cause Bud and his pals to go “hey, let’s do that”. And also, after this system is developed and implemented, what is the impact on the game on the field and is this what we really want?

Nice post Roy.

[dougdirt]

Originally Posted by Roy Tucker

Then we start getting into the harder stuff. We need this system to be able to detect where the strike zone is on a batter by batter, pitch by pitch basis. Each batter is different and can adjust his stance on a pitch by pitch basis. So we need to be able to accomodate that. I think that crosses off any pre-configuration of a strike zone for a batter, e.g. before a season or a game, the batters strike zone is measured and he is stuck with that for that game, series, month, or season. It needs to be done on each pitch in ral time.

It is impossible for the umpire to do this. They are making the judgement on the batters strikezone at the set up of the batter, not when he is swinging. You can't possibly track a 90 MPH fastball through the strikezone while also watching where the batters midpoint line between belt/shoulders. It is impossible and they don't, nor have they ever done it that way. The current Pitch F/X system measures the top of the strikezone on a player basis by at bat, not by each pitch. But, that can be changed to each pitch, easily, if they wanted to do it.

The two basic ways I can think of this happening is:

1.) A human operator sets it “when the batter is prepared to swing at a pitched ball”. So this would hopefully be a trained umpire up in the booth armed with a video monitor and a mouse, and he clicks the top and bottom of the zone as the batter assumes his stance (or some method like that). This isn’t a perfect way to do it since we again are asking a human to do this (and is it better than a home plate ump?) but if this system ever sees daylight, this would probably be an improvement. How much improvement would need to be determined. His hand may slip, he might prematurely click the mouse and try to correct and not make it in time. his attention may wander, he might cough or sneeze, or some other dumb little thing will happen over a 300+ pitch game.

This is what happens already. That is exactly how the top and bottom of the zones are measured.

[westofyou]

But, that can be changed to each pitch, easily, if they wanted to do it.

This is what I hear about every project I've ever worked on.

I'm still struggling to find what the word "easily" really means.

[dougdirt]

Originally Posted by westofyou

This is what I hear about every project I've ever worked on.

I'm still struggling to find what the word "easily" really means.

There is a guy watching the camera who sets it before each at bat based on the batters stance. Why would it be more difficult to click a mouse twice before each pitch when he already does it at the start of every at bat?

[Caveman Techie]

Ok, some of you guys are just being contrary to be contrary. Denying the existence of the technology we have today that is FULLY capable of doing the job.

Saying it's "unproven" and therefore you can't use it deny's the ability to prove the technology! That's a circular argument. I can understand not wanting to change the game, but I don't necessarily see this as "changing" the game as making the rules real! A strike is a strike is a strike, unless the umpire decides it's not, then it doesn't matter that the ball passed through what the rulebook defined as the strikezone. The umpires get it wrong alot, and this technology can help them get it right. Earlier in this thread someone posted a video of the Hawkeye system in Tennis. Has that solution ruined Tennis? Getting the call right is more important than some umpires ego! And honestly if it takes breaking the umpires union to get it done, after watching this season where an umpire deliberately made two horrible calls to "send a message" to a player that he thought showed him up (think it was a Toronto game, it was discussed quite a bit here). The union needs busted.

[edabbs44]

Originally Posted by dougdirt

There is a guy watching the camera who sets it before each at bat based on the batters stance. Why would it be more difficult to click a mouse twice before each pitch when he already does it at the start of every at bat?

Then we'll have to hear about a computer replacing the mouse operator.

[Brutus]

Originally Posted by Caveman Techie

Ok, some of you guys are just being contrary to be contrary. Denying the existence of the technology we have today that is FULLY capable of doing the job.

Saying it's "unproven" and therefore you can't use it deny's the ability to prove the technology! That's a circular argument. I can understand not wanting to change the game, but I don't necessarily see this as "changing" the game as making the rules real! A strike is a strike is a strike, unless the umpire decides it's not, then it doesn't matter that the ball passed through what the rulebook defined as the strikezone. The umpires get it wrong alot, and this technology can help them get it right. Earlier in this thread someone posted a video of the Hawkeye system in Tennis. Has that solution ruined Tennis? Getting the call right is more important than some umpires ego! And honestly if it takes breaking the umpires union to get it done, after watching this season where an umpire deliberately made two horrible calls to "send a message" to a player that he thought showed him up (think it was a Toronto game, it was discussed quite a bit here). The union needs busted.

Likewise, saying something is "proven" that hasn't even been implemented for the use people are suggesting is not very prudent either.

The system was put into place to do a number of things... incorporate visual data online; give better statistical analysis of players; and further, grade umpires. The system was not designed to call balls and strikes directly.

So saying it's proven to do something it wasn't designed to do is a little bit of a stretch.

[Caveman Techie]

Originally Posted by Brutus

Likewise, saying something is "proven" that hasn't even been implemented for the use people are suggesting is not very prudent either.

The system was put into place to do a number of things... incorporate visual data online; give better statistical analysis of players; and further, grade umpires. The system was not designed to call balls and strikes directly.

So saying it's proven to do something it wasn't designed to do is a little bit of a stretch.

The system was put in place to help judge umpires on their balls and strike calls. So it's a good enough technology to judge umpires, but not good enough to determine if the ball was a strike or not? That argument just seems very weak to me. It was designed SPECIFICALLY to call balls and strikes, now we just need to do it.