Why we need computers calling balls and strikes

[M2]

Damn, I've got to get me a rocking chair. Didn't realize we fogies were so into technical fixes for obvious problems.

[AtomicDumpling]

Originally Posted by Brutus

I haven't been corrected on anything, AD. I already explained to you that the 99% used by the new system isn't being measured in the same way. The old system claimed 97% of the strike zone called accurate. The new system is measuring 99% accuracy "within an inch" of the location. They are two different things.

The accuracy still hasn't improved "leaps and bounds" like the hyperbole you suggested earlier.

You were indeed corrected. You were quoting very old data and representing it as current data.

You don't think the science and technology behind making measurements is improving? Technology everywhere is improving. Why would pitch tracking be any different?

We can track enemy missiles launched thousands of miles away and pinpoint their position within centimeters even though they are moving at the speed of sound or faster, then fire our own missiles to shoot them out of the sky. Compared to that, measuring the location of a fastball in a ballpark is child's play.

Anyone who says the technology to call balls and strikes electronically is not already available must not use the Internet to follow baseball games. The technology is already in place, it is getting more awesome every year. If MLB decides to use robo-umps it would only take a year or two to perfect and implement a nearly flawless system. It would be a big upgrade over the umpires who make several mistakes in every game.

[M2]

Originally Posted by George Anderson

But if we are using technology to call balls and strikes then we should also be consistent and use technology to review every other play on the field because you and others believe technology is superior to humans. This will take alot of time.

No, that's just you trying to make it into something absurd so you don't have to deal with the reality that the existing technology for calling balls and strikes is already more accurate and consistent than umps (as an aside, it's the technology they use to determine how frequently umps make/miss their calls).

Goofy as umps can be about bang-bang calls and SB tags, we're going to have to live with it because it's the only way to make an instantaneous call. Balls and strikes is a different story.

[Brutus]

Originally Posted by AtomicDumpling

You were indeed corrected. You were quoting very old data and representing it as current data.

You don't think the science and technology behind making measurements is improving? Technology everywhere is improving. Why would pitch tracking be any different?

We can track enemy missiles launched thousands of miles away and pinpoint their position within centimeters even though they are moving at the speed of sound or faster, then fire our own missiles to shoot them out of the sky. Compared to that, measuring the location of a fastball in a ballpark is child's play.

Anyone who says the technology to call balls and strikes electronically is not already available must not use the Internet to follow baseball games. The technology is already in place, it is getting more awesome every year. If MLB decides to use robo-umps it would only take a year or two to perfect and implement a nearly flawless system. It would be a big upgrade over the umpires who make several mistakes in every game.

No I wasn't corrected. You can keep saying it all you want but it doesn't make it true. Again, the Pitch F/X accuracy statements, if you compare them to what QuesTech was measuring, aren't any better. They're just repositioning how it's measured.

QuesTech was claiming 97% actual accuracy. Pitch F/X claims 99% "within an inch."

You're really stretching to act like that extra 2% "within an inch" is improving leaps and bounds. In fact, it's virtually no improvement at all since it's a different measurement, other than eliminating the margin of error to be a little more compact.

That's not the "leaps and bounds" improvement narrative you were trumpeting earlier.

Basically QuesTech was 97 out of 100 pitches correct. Pitch F/X is claiming 99 out of 100 pitches are correct -- to within an inch.

That's not an improvement. That's simply moving the goal post and calling it so.

[AtomicDumpling]

Originally Posted by Brutus

No I wasn't corrected. You can keep saying it all you want but it doesn't make it true. Again, the Pitch F/X accuracy statements, if you compare them to what QuesTech was measuring, aren't any better. They're just repositioning how it's measured.

QuesTech was claiming 97% actual accuracy. Pitch F/X claims 99% "within an inch."

You're really stretching to act like that extra 2% "within an inch" is improving leaps and bounds. In fact, it's virtually no improvement at all since it's a different measurement, other than eliminating the margin of error to be a little more compact.

That's not the "leaps and bounds" improvement narrative you were trumpeting earlier.

Basically QuesTech was 97 out of 100 pitches correct. Pitch F/X is claiming 99 out of 100 pitches are correct -- to within an inch.

That's not an improvement. That's simply moving the goal post and calling it so.

Again, you are obfuscating to avoid the real question. You are using old numbers. You are claiming that 99% is not better than 97% (also within an inch by the way which you failed to mention). When you consider that 95% of pitches are easy calls that are not disputed by any technique, a difference between 97% and 99% is a huge improvement. Essentially that means that of the 5% of calls that are close enough to be disputed, QuesTech missed 60% of them and Pitch F/X misses 20% of them. That means Pitch F/X is a huge leaps and bounds improvement over the obsolete QuesTech system, which is itself much more accurate and consistent than human umpires.

If the Pitch F/X system wasn't a big improvement over QuesTech then why did MLBAM spend the money to make the upgrade?

Are you arguing that we should not install a ball/strike system because technology is not superior to human strike zone judgement and won't be anytime soon? I hope not, because the technology has already been proven more accurate and more consistent compared to MLB umpires and is getting better every year.

Do you think that measurement technology is getting better or not? I think that technology is improving. If you disagree, then can you explain why this technology is not improving while all other technologies are advancing?

I think Pitch F/X gets better each year. It certainly has added many new features and the pitch classifications have improved every year. There is no logical reason to believe that the rapid progress will stop and the system won't get any better in the future.

Someone can argue they like having human umpires instead of a technological system and I can see that point of view. But to argue that we shouldn't move to robo-umps because the technology doesn't exist or is not good enough flies in the face of reason and the obvious evidence to the contrary.

[George Anderson]

Originally Posted by M2

No, that's just you trying to make it into something absurd so you don't have to deal with the reality that the existing technology for calling balls and strikes is already more accurate and consistent than umps (as an aside, it's the technology they use to determine how frequently umps make/miss their calls).

Goofy as umps can be about bang-bang calls and SB tags, we're going to have to live with it because it's the only way to make an instantaneous call. Balls and strikes is a different story.

Oh my humble apology for being so absurd. See my definition of being absurd is thinking MLB will radically changing the game over night by eliminating humans with a technology system that has yet to be tried at any level of professional or amateur baseball. Also disregarding the fact that the difference between humans and technology in regards to calling balls and strikes is minor at best.


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.

Once again, my humble apology.

[Brutus]

Originally Posted by AtomicDumpling

Again, you are obfuscating to avoid the real question. You are using old numbers. You are claiming that 99% is not better than 97% (also within an inch by the way which you failed to mention). When you consider that 95% of pitches are easy calls that are not disputed by any technique, a difference between 97% and 99% is a huge improvement. Essentially that means that of the 5% of calls that are close enough to be disputed, QuesTech missed 60% of them and Pitch F/X misses 20% of them. That means Pitch F/X is a huge leaps and bounds improvement over the obsolete QuesTech system, which is itself much more accurate and consistent than human umpires.

If the Pitch F/X system wasn't a big improvement over QuesTech then why did MLBAM spend the money to make the upgrade?

Are you arguing that we should not install a ball/strike system because technology is not superior to human strike zone judgement and won't be anytime soon? I hope not, because the technology has already been proven more accurate and more consistent compared to MLB umpires and is getting better every year.

Do you think that measurement technology is getting better or not? I think that technology is improving. If you disagree, then can you explain why this technology is not improving while all other technologies are advancing?

I think Pitch F/X gets better each year. It certainly has added many new features and the pitch classifications have improved every year. There is no logical reason to believe that the rapid progress will stop and the system won't get any better in the future.

Someone can argue they like having human umpires instead of a technological system and I can see that point of view. But to argue that we shouldn't move to robo-umps because the technology doesn't exist or is not good enough flies in the face of reason and the obvious evidence to the contrary.

It seems to me that you don't understand exactly what it is QuesTech was and did. QuesTech's sole purpose was to snap 3-D imagery of each pitch on called balls or strikes to send back to Major League Baseball for evaluation of umpires. The operator would take a snapshot on each pitch and record the log, and the data would be saved to a CD-Rom along with video so MLB could evaluate. Thus, QuesTech's claim was 97% accuracy of the zone. That's it. There was no "within an inch" as their sole designation was to simply say 'in' or 'out' of the strike zone. That was their purpose.

Hence, the reason, then, MLBAM switched to Pitch F/X in all the parks had NOTHING to do with accuracy. It was because the Pitch F/X technology supported by Sportvision simply expanded the capability. Instead of simply grading umpires, the system was meant to capture pitch locations, trajectory and velocity and be used in a versatile interface that was compatible with MLBAM's desire to incorporate online and statistical aspects of the sport. That is why they switched. They could incorporate all of these features while still grading umpires.

Here is the most amazing aspect of all of this that people are missing... this technology to work relies on human input. Did anyone stop to wonder how the strike zones are defined in QuesTech or the "Z-system" that Pitch F/X uses? If you guess HUMAN OPERATOR, you would be correct.

That's right... the technology everyone is pining for still relies on a human to manually set the strike zone before each at-bat. While the technology captures the location of the pitch at 99% accuracy "within an inch," whether that location is in the strike zone is dependent on a HUMAN operator calibrating the zone prior to each at-bat. So even if a given pitch is 100% accurate as to its true location, the relative location as it fits within the strike zone is set by an operator prior to the at-bat meaning that if the human makes an error, it will impact the accuracy of the location of the pitch relative to the zone.

I'm not against technology. But the problem is people talking about how it's "proven" clearly don't know how the system operates. The entire accuracy of the system with a strike zone is based on the operator manually setting the parameters correctly prior to each batter stepping in the box. If the system is set wrong by even an inch or two, then you've basically not improved anything other than consistency of the error margins.

[Caveman Techie]

Originally Posted by Brutus

It seems to me that you don't understand exactly what it is QuesTech was and did. QuesTech's sole purpose was to snap 3-D imagery of each pitch on called balls or strikes to send back to Major League Baseball for evaluation of umpires. The operator would take a snapshot on each pitch and record the log, and the data would be saved to a CD-Rom along with video so MLB could evaluate. Thus, QuesTech's claim was 97% accuracy of the zone. That's it. There was no "within an inch" as their sole designation was to simply say 'in' or 'out' of the strike zone. That was their purpose.

Hence, the reason, then, MLBAM switched to Pitch F/X in all the parks had NOTHING to do with accuracy. It was because the Pitch F/X technology supported by Sportvision simply expanded the capability. Instead of simply grading umpires, the system was meant to capture pitch locations, trajectory and velocity and be used in a versatile interface that was compatible with MLBAM's desire to incorporate online and statistical aspects of the sport. That is why they switched. They could incorporate all of these features while still grading umpires.

Here is the most amazing aspect of all of this that people are missing... this technology to work relies on human input. Did anyone stop to wonder how the strike zones are defined in QuesTech or the "Z-system" that Pitch F/X uses? If you guess HUMAN OPERATOR, you would be correct.

That's right... the technology everyone is pining for still relies on a human to manually set the strike zone before each at-bat. While the technology captures the location of the pitch at 99% accuracy "within an inch," whether that location is in the strike zone is dependent on a HUMAN operator calibrating the zone prior to each at-bat. So even if a given pitch is 100% accurate as to its true location, the relative location as it fits within the strike zone is set by an operator prior to the at-bat meaning that if the human makes an error, it will impact the accuracy of the location of the pitch relative to the zone.

I'm not against technology. But the problem is people talking about how it's "proven" clearly don't know how the system operates. The entire accuracy of the system with a strike zone is based on the operator manually setting the parameters correctly prior to each batter stepping in the box. If the system is set wrong by even an inch or two, then you've basically not improved anything other than consistency of the error margins.

All the problems you just listed with the PitchF/X exist today with the umpire, but with PitchF/X once the strikezone is set (BTW I can think of many different ways to fix the human intervention) the system is more accurate and consistent than an umpire could every be.

[RBA]

Originally Posted by dougdirt

No. I live in Cincinnati where we don't get to watch our team play.

Easy fix, sell out the game 3 days in advance.

[Brutus]

Originally Posted by Caveman Techie

All the problems you just listed with the PitchF/X exist today with the umpire, but with PitchF/X once the strikezone is set (BTW I can think of many different ways to fix the human intervention) the system is more accurate and consistent than an umpire could every be.

Not necessarily. If you calibrate the strike zone prior to each at-bat, as a hitter, what if you crouch more or less after that? The strike zone is immediately thrown off. And while hitters' stances are generally consistent, even an inch or two difference in a stance can throw off the zone.

In theory, you're right that you could achieve a more consistent strike zone. But in practice there are a few things that could greatly diminish the accuracy advantage you gain with the system.

[westofyou]

Originally Posted by Brutus

Not necessarily. If you calibrate the strike zone prior to each at-bat, as a hitter, what if you crouch more or less after that? The strike zone is immediately thrown off. And while hitters' stances are generally consistent, even an inch or two difference in a stance can throw off the zone.

In theory, you're right that you could achieve a more consistent strike zone. But in practice there are a few things that could greatly diminish the accuracy advantage you gain with the system.

Pete Rose would have made it a point to break the setup

[Brutus]

Originally Posted by westofyou

Pete Rose would have made it a point to break the setup

Ha! And something tells me he would not have been bashful about it, either.

[dougdirt]

Originally Posted by RBA

Easy fix, sell out the game 3 days in advance.

I don't have the kind of money to buy thousands of tickets every other week at $70 a pop.

[AtomicDumpling]

Originally Posted by Brutus

It seems to me that you don't understand exactly what it is QuesTech was and did. QuesTech's sole purpose was to snap 3-D imagery of each pitch on called balls or strikes to send back to Major League Baseball for evaluation of umpires. The operator would take a snapshot on each pitch and record the log, and the data would be saved to a CD-Rom along with video so MLB could evaluate. Thus, QuesTech's claim was 97% accuracy of the zone. That's it. There was no "within an inch" as their sole designation was to simply say 'in' or 'out' of the strike zone. That was their purpose.

Hence, the reason, then, MLBAM switched to Pitch F/X in all the parks had NOTHING to do with accuracy. It was because the Pitch F/X technology supported by Sportvision simply expanded the capability. Instead of simply grading umpires, the system was meant to capture pitch locations, trajectory and velocity and be used in a versatile interface that was compatible with MLBAM's desire to incorporate online and statistical aspects of the sport. That is why they switched. They could incorporate all of these features while still grading umpires.

Here is the most amazing aspect of all of this that people are missing... this technology to work relies on human input. Did anyone stop to wonder how the strike zones are defined in QuesTech or the "Z-system" that Pitch F/X uses? If you guess HUMAN OPERATOR, you would be correct.

That's right... the technology everyone is pining for still relies on a human to manually set the strike zone before each at-bat. While the technology captures the location of the pitch at 99% accuracy "within an inch," whether that location is in the strike zone is dependent on a HUMAN operator calibrating the zone prior to each at-bat. So even if a given pitch is 100% accurate as to its true location, the relative location as it fits within the strike zone is set by an operator prior to the at-bat meaning that if the human makes an error, it will impact the accuracy of the location of the pitch relative to the zone.

I'm not against technology. But the problem is people talking about how it's "proven" clearly don't know how the system operates. The entire accuracy of the system with a strike zone is based on the operator manually setting the parameters correctly prior to each batter stepping in the box. If the system is set wrong by even an inch or two, then you've basically not improved anything other than consistency of the error margins.

So now you do finally acknowledge that Pitch F/X is vastly superior to the QuesTech system that you were espousing earlier. At least we are making progress there. There is no need to talk about Questech any more since it hasn't been in use for years. MLB upgraded to Pitch F/X, which everyone agrees has been proven to be far more effective and accurate as well as having many more features. The technology for measuring balls and strikes has been getting better, which is exactly the point I was making.

[Brutus]

Originally Posted by AtomicDumpling

So now you do finally acknowledge that Pitch F/X is vastly superior to the QuesTech system that you were espousing earlier. At least we are making progress there. There is no need to talk about Questech any more since it hasn't been in use for years. MLB upgraded to Pitch F/X, which everyone agrees has been proven to be far more effective and accurate as well as having many more features. The technology for measuring balls and strikes has been getting better, which is exactly the point I was making.

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.