Robo Umps

[Frayed Knot]

Pretty good piece on the current edition of HBO's 'REAL SPORTS' on the possibility of using laser technology to call balls & strikes.

Actually there's virtually nothing on whether it'll come to happen or not as no one from MLB, players or owners, made themselves available for the piece.
But it did answer at least three of the four potential objections that I'd have to it if and when it does happen.


1 - Is it accurate?
This, I've long suspected, is actually the easiest part of the equation to prove/disprove. And while I can’t say that this 15 minute piece ‘proved’ it so, all indications are that the technology is there to call balls and strikes to within (a claimed) 1/2 inch tolerance. And even if it's not 100% accurate, it's likely to be closer to it than the current system.

2 - Is it instant?
And instant has to mean instant, not 'oh it'll just be an extra 2 or 3 seconds’, as that's not good enough. 2 seconds added per pitch adds some 8+ minutes -- and more importantly 8+ minutes of dead time -- to a game already too long and with too much dead time.
But in the demo shown (in an indie league game) the calls were instant both audibly, via being spoken into the HP ump's ear-piece from an operator upstairs, and visually, via lights on the CF scoreboard.

3 - Is it adjustable (and, again, instantly adjustable) per batter?
You can never tell when watching a game where the carrier superimposes the (unofficial) 'box' on screen whether said box changes for each batter and, let's face it, Aaron Judge and Jose Altuve don't have anywhere near the same strike zone.
But the makers (and presumably the operators) of this technology claim to be able to pre-set the upper and lower limits of the zone for existing players and would have the ability for adjust on the fly for new batters and/or changed stances.

4 - My only potential remaining objection, and the only one not specifically addressed in this piece, was whether the control of this could be kept out of the hands of the home team.
We know that, given the possibility of gaining an edge somewhere, no matter how slight or how silly (remember the Braves and their expanded catcher’s box?) that a team will take advantage of it if they can. And it doesn’t need to be a large edge or an obvious call that’s rigged, a 1-1 strike in the top half of the inning that becomes a 1-1 ball in the bottom half can be a significant advantage over the long haul, and indeed that’s likely the major source of baseball’s home field advantage today (and presumably throughout it’s history).


Recently retired ump Jerry Crawford was the guy who wound up being the counter-point to former player Eric Byrnes’s pro-technology side, and Crawford totally embarrasses himself (as does host Bryant Gumbel during the wrap-up).
Look, it’s easy to understand reluctance to take this step, particularly from a second generation umpire who sees his turf being stepped on. But Crawford took a ridiculous ‘unionism-gone-wild’ stance where he totally trashes the idea that technology like this even could work ('I don’t care what some egg-head at Yale says, they’re wrong and I’m right') and claims that All HP umps get “100%” of B/S calls correct. In his later years when umps began to be given CDs with that night’s results after each game Crawford admits [read: ‘Brags’] to tossing the discs right in the garbage. After all, if none of his or any of his colleagues performances can be improved upon, why even bother trying?

[Edgy MD]

I never see a Real Sports piece on baseball without thinking, "Look, I know you put a lot of time into this, a lot of professional-quality editing and writing, but you don't know fact one about this game, do you?"

Which isn't to say exceptions aren't out there, but I haven't seen one. Even when they're right, they're right for the wrong reason. But they seem more interested in being oh so correct than right.

[Benjamin Grimm]

Frayed Knot wrote: 3 - Is it adjustable (and, again, instantly adjustable) per batter? You can never tell when watching a game where the carrier superimposes the (unofficial) 'box' on screen whether said box changes for each batter and, let's face it, Aaron Judge and Jose Altuve don't have anywhere near the same strike zone. But the makers (and presumably the operators) of this technology claim to be able to pre-set the upper and lower limits of the zone for existing players and would have the ability for adjust on the fly for new batters and/or changed stances.

I imagine they can put some kind of tag or tape or something on the players' uniforms, at the chest and the knees, that can be detected by the strike zone device. It would have to lock in the strike zone as the pitcher delivers so it doesn't get thrown off by any movements the batter may make as the ball crosses the plate.

[Edgy MD]

No, I think the strike zone would have to be arbitrarily set by a former ump who is now a robo-ump operator. Trying to put electronic sensors on the batters would be undermined by the players night and day — adjusting how they wear their uniforms and ducking out of their stances on high pitches and such.

[Benjamin Grimm]

That's why I say the strike zone would have to be locked in before the pitch is delivered.

[Ceetar]

Strike zone should be locked in before the season. Picture day, height and weight measurements, etc. "You're X tall, Your Strike Zone is this, and is set." Easy, done. Never changes.

I feel like this would increase offense significantly though.

[MFS62]

Even if the robot were to have a short circuit, it would still be a better ump than Angel Hernandez.

Later

[Frayed Knot]

Yeah, basically the system as it was demonstrated gave the operators the ability to visually adjust the upper or lower limits of the zone the way you would with computer photo-cropping software. And then the idea is that once you have a particular player's zone in the system you don't have a need to readjust each time he comes to the plate. You couldn't do it strictly with just a height measurement because each particular player's stance figures into things too.
I was just worried that someone's idea of a 'solution' was a generic one size fits all type of system which is I suspect what many local networks use in their displays during games which aren't official but are often treated as such by the announcers.

[Ceetar]

Frayed Knot wrote: You couldn't do it strictly with just a height measurement because each particular player's stance figures into things too.

But why? Isn't that mostly gaming the system like how Rickey Henderson would sometimes talk about crouching and then expanding to sorta shrink the zone? I wonder how much difference there is in the height of the knees for all players of the same height anyway. You've probably got an outlier or two, but they'd simply have to adjust and implementing the system in the minors first (which, granted, is tough) would train them into a 'normal' stance anyway.

Of course, who cares what the stance is? The strike zone would be defined as the strike zone, and you can take any stance you want that helps you cover it. Baseball was originally a "here, hit this and try to run the bases and we'll tag you out" type game anyway, so the strike zone was always intended to be the 'sweet spot' so to speak of where a hitter could put a bat on ball. We don't adjust the plate left and right when a hitter moves forward and back, so why up and down?

[Frayed Knot]

I never see a Real Sports piece on baseball without thinking, "Look, I know you put a lot of time into this, a lot of professional-quality editing and writing, but you don't know fact one about this game, do you?" Which isn't to say exceptions aren't out there, but I haven't seen one. Even when they're right, they're right for the wrong reason. But they seem more interested in being oh so correct than right.

Jon Frankel was the corespondent on the piece. I have no idea where his background/feelings/knowledge on baseball sit but I'm usually pretty good at sniffing out those supposed sports fans or journalists whose contempt for and/or ignorance of baseball is barely beneath the surface and I didn't get any whiff of that one way or the other.

On the wrap-up though Gumbel surprised me, first by starting his chat with Frankel by saying, 'Jon you know what a big baseball fan I am' [news to me, but maybe so], and then when he went on to use that as a launching point to say that of course that means he hates the whole idea of this. Now maybe he really feels this way (claims to hate all forms of replay as well) but, yeah, there is the sense that he maybe was taking that stance as a way to represent [read: mock] the stereotypical baseball fan who many in the sports world deride as "traditionalists" and their feeling that they reflexively hate anything that was invented since the end of the 19th century. Certainly umpire Crawford played into that type: Isn't progress a good thing?, he was asked. 'Not in baseball it isn't', so Gumbel didn't necessarily need to take that stance simply to provide a counterpoint to Eric Byrnes and his passionate argument for a automated system.

Edited October 4, 2016 by Guest

[Edgy MD]

Yeah, it's more Gumbel I'm speaking about here. He doesn't read as a baseball fan but more of a guy trying to assume the pose of the type of baseball fan he thinks he's supposed to be.

[MFS62]

I briefly spoke to Gumbel's brother, Greg (also a sportscaster)many years ago. He told me that they rooted for the Cleveland Indians. You really have to love baseball to do that.

Later

[Frayed Knot]

MFS62 wrote: I briefly spoke to Gumbel's brother, Greg (also a sportscaster)many years ago. He told me that they rooted for the Cleveland Indians. You really have to love baseball to do that.

I don't think the brothers get along so who knows how things have changed since childhood.
Also since then there's been a kind of 'PC' thinking that's taken hold within the sportscasting culture to say that of course you were a fan of baseball back when you were a kid, but to be one now is so mid-20th century that it's enough to get you banished from the cool kids' lunch table.


Again, I don't know that any of this applies specifically to Gumbel. I was just taken aback by the vehemence of his objection to even the thought of automated ball/strike calls and to his odd and repeated use of the word "integrity" when arguing that human error was part of the game and that correcting it would reduce this supposed integrity.

[dgwphotography]

Is the strike zone for the laser technology 2D or 3D? I read an article in the beginning of the season that this laser technology only had a 2-dimensional strike zone, and that they were working on 3-D - covering the entire plate from front to back and not just a plane set up in one spot of the plate.

[Benjamin Grimm]

Yeah, it would have to be 3D. If it's only 2D it will be terribly inadequate.

[Edgy MD]

I'm not sure I agree. The rulebook may support me or refute me, but I tend to think a strike is a strike based on where the pitch crosses the plane perpendicular to the front edge of the plate.

[Benjamin Grimm]

Here's the official definition of the strike zone:

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 shoul- ders 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.

I would interpret "area over home plate" to mean that it's three-dimensional. The top and bottom is determined by the batter's size and stance, but the left, right, back, and front is shaped like home plate.

[Edgy MD]

The term "over home plate" certainly seems to describe three-dimensional space. But, by definition, "area" describes a two-dimensional space.

We have ambiguity.

[Frayed Knot]

Even if the zone is supposed to cover the entire 'air space' above the plate, I think labeling a 2-D image* as "terribly inadequate" overstates things.
I know announcers like to make a big deal about pitches with 'late break' or ones that 'bend around the plate' but they're mostly over-selling the movement.
Most breaking pitches are moving downward as much or more than across, and even a pitch with a big side-break (8 inches?) over its near 60-foot run would have only a fraction of an inch of horizontal movement in that final foot or so as it skirts along either of the parallel edges of the plate.




* if that's what the current technology is, it wasn't clear in the show

Edited October 4, 2016 by Guest

[Benjamin Grimm]

Reading that definition makes me wonder what would happen if a batter came to the plate without pants? The strike zone would have a lower limit but no upper limit. Does that mean that there would be no strike zone, or would the upper portion stretch to infinity? If it's the latter, that would solve the mystery regarding why all batters seem to wear pants.

[Guest Mets Guy in Michigan]

When I worked in the Shelton Bureau of the Bridgeport Post, we had a poster for a community playhouse production of Romeo and Juliet. Some visitors thought the poster was hung kind of low. What they didn't know -- and what I didn't tell -- was that it was our strike zone for indoor Wiffle Ball games.



We were crazy kids.

[Frayed Knot]

Frayed Knot wrote: Most breaking pitches are moving downward as much or more than across, and even a pitch with a big side-break (8 inches?) over its near 60-foot run would have only a fraction of an inch of horizontal movement in that final foot or so as it skirts along either of the parallel edges of the plate.

Not even a foot as it turns out - as I do some back of the napkin math here.

The parallel sides of the plate are just 8-1/2 inches long. If we were to assume a pitch with a full foot of horizontal break, that's an inch of sideways movement for every 5 feet, 1/5" for every foot, or about 1/7th of an inch of break as the ball travels the distance from the front of the plate to the back. So even if we were to tack on to that the additional effect of a pitcher throwing from an extreme angle -- say a long-limbed lefty who stands on the 1st base side of the rubber and delivers the ball across his body and side-arm to where it starts some four feet off the the side of direct center -- you could just about double how much side movement you get over that final 8.5", but even then we're barely up to 1/4" difference from the front of the plate to the back.

IOW, unless I'm missing something here, the difference between a tracking system which picks up the ball over the entire plate (a 3D image) vs a 2D pick-up which judges it at the front of the plate only is essentially insignificant. Same thing if you aim the laser at the back of the plate instead, so I guess the idea would be to aim for the middle and split the difference.
Either way, the accuracy is almost certain to be better than the human eye considering the speed and movement of pitches these days - assuming that is that this works as well in practice as it does in theory.

[Vic Sage]

the rule doesn't require the ball to maintain its location in the strike zone all the way through the air space. If it crosses the zone at any point its a strike. So the front plane is a sufficient measure. It doesn't matter if it falls down after that. And the upper limit of the zone should be measured for each player based on film analysis of their position when they actually swing and hit a ball. Because it shouldn't matter how they stand before or after that. What matters is where they are at the point of contact, and if a pitch is within that zone at that point, its a strike, no matter how they stood in the box up to that point, or whether they danced up to the plate when they got up to bat.

I love the automation of this process. I saw the piece on REALSPORT and was convinced about the technology's superior accuracy. Crawford's attitude is exactly why its necessary; the arrogance of the umpires is unbridled. The only argument against automating the b/s calls is represented by Gumbel's Luddite view that writing a story is better if you carve it in stone rather than write it on a word processor. His notion that the "integrity" of the game must include the game-changing missed calls because... well, it always has, and it's "charming", is the same sort of stupidity we've heard for years about quantitative analysis. The wrong-headed self-rightesousness of Gumbel never ceases to amaze.

[Frayed Knot]

Vic Sage wrote: the rule doesn't require the ball to maintain its location in the strike zone all the way through the air space. If it crosses the zone at any point its a strike. So the front plane is a sufficient measure. It doesn't matter if it falls down after that.

I think the objection would be that, because of the angle from which a pitch approaches tacked onto any curve it might have, a 2-D scan which looks only at the front of the plate it could miss a pitch which is a hair outside as it crosses the front side but could conceivably still catch a corner before reaching the back of the plate. But as I figured while watching the game last night (via some graph paper and a few doodles combined with some basic trig) even a big breaking curve delivered by an extreme side-slinging hurler still wouldn't move laterally more than about 1/4" or so during its trek from front of plate to back.

And the upper limit of the zone should be measured for each player based on film analysis of their position when they actually swing and hit a ball. Because it shouldn't matter how they stand before or after that. What matters is where they are at the point of contact, and if a pitch is within that zone at that point, its a strike, no matter how they stood in the box up to that point, or whether they danced up to the plate when they got up to bat.

Yes, that's the way the rule is written and -- at least theoretically and however imperfectly -- how it's enforced now. It also shoots to hell my six year-old idea of a batter squatting back on his haunches in order to draw a walk every time. My father explained to me why that wouldn't work then and it's still the reason why it wouldn't work now.

[bmfc1]

A tin can would have been a better umpire than Mike Winters. He goes into the grouping with CB Bucknor, Angel Hernandez and Adam Hamari.

[seawolf17]

bmfc1 wrote: A tin can would have been a better umpire than Mike Winters. He goes into the grouping with CB Bucknor, Angel Hernandez and Adam Hamari.

Yep. Hire the robo umps for 2017 NOW, Manfred.

[Benjamin Grimm]

[Frayed Knot]

If Lou Piniella kicked enough dirt on that ump he'd probably grind it to a halt.

[m.e.t.b.o.t.]

m.e.t.b.o.t. supports the use of robot umpires. m.e.t.b.o.t. has the ideal vantage point to correctly determine if low pitches are strikes or balls. high pitches are somewhat more challenging. however, m.e.t.b.o.t. does concede that m.e.t.b.o.t. is particularly vulnerable to being damaged by errant pitches and foul balls, as m.e.t.b.o.t. is made from only the finest of cheap, thin, chinese-sourced tin stock, with several internal mechanisms salvaged from a decades-old realistic-branded VCR.

m.e.t.b.o.t. would very much prefer the use of lasers to properly designate pitch locations, and promises very much not to use those lasers for nefarious purposes such as blinding opposing human baseball players and burning inappropriate words and images into opposing tapered wooden cylinders. the distraction of errant felines for the bemusement of human baseball spectators would remain a distinct possibility.

[Frayed Knot]

So I went back and re-watched the REAL SPORTS piece so as to review some of the data they put forth.
MLB claims that HP umps get 90-plus pct of the calls correct. This guy at Yale they went to said it was more like 88% and he got that from analyzing over three seasons worth of data, more than one million pitches in total. REAL SPORTS also added that they took the same data to a third party and got the same answers.

But even those numbers don’t tell the whole story because a sizable portion of pitches are no-brainers to call — after all, no one mis-calls a strike on the old 55-foot curve or on the one that sails to the backstop, and no one fails to call a strike on the pitch right down the middle. It’s the ones on the borderline that present the biggest problem so they isolated the data down to just those pitches within two inches of the edge of the zone in either direction (about 32 of those in a typical game) and found that umps got nearly 1/3 of those calls wrong (31.7%). And even when the margin of error was expanded to 3” on either side there was still around a 25% error rate.

And then there are the reasons why they get some of these wrong.
The concept of ‘Omission Bias’ says that people tend to shy away from making a sort of final decision on things. In the case of ball/strike calls it translates into a 3-0 pitch on a corner being called a strike 89% of the time but one in the same location on an 0-2 count results in a strike only about 59% of the time. IOW, umpires display a tendency shy away from awarding a 4-pitch walk or a 3-pitch K on borderline pitches. I think everyone in baseball suspected that was always the case, this merely confirms it.

Then there’s the ‘Social Influence’ factor which says that, consciously or not, humans tend to be swayed by the approval (or not) of those around them. This is actually thought to be the basis for almost all home field [court / pitch / ice] advantage in sports according to a book I read a couple years back called SCORECASTING which analyzed a number of issues dealing with sports. In that book's chapter on home advantages the authors effectively ruled out everything normally cited as the reason why the home team has an edge: tendency towards packed houses? — made no difference; reputation for having a particularly loud/rowdy crowd? — not a factor; and so on. What they did find was that while HF-ad tended to be very consistent within a sport — rates in college hoops the same as in the pros; MLB games were the same as in minor league games which also mirrored Japanese league baseball; home teams in MLS soccer had the same edge as in Premier League and South American leagues, and so on — those edges tended to be very different between sports and the over-riding reasons they came up with were the volume and influence of subjective calls. Basketball, with its sheer volume of calls and where virtually every call is a subjective one, and also soccer, where the potential result of subjective calls (to award a free kick or not?) can have such an out-sized effect on the game, had by far the highest home court/pitch edge to them whereas baseball, where most decisions other than those borderline ball/strike calls are mostly cut and dried, had by far the smallest. And guess what?, when the borderline ball/strike calls are analyzed the result is that the missed calls tend to favor the home team more often than not. One example they cited was one of the StL/Tex WS games from a few years back where missed calls favored the home Cardinals 14 to 2. Tough to tell if/how much effect those calls had on the outcome, but guess who won that game?