The subject of instant replay is once again in the hearts and minds and throats and fingers of our nation’s baseball commentators, spurred on by the existence of actual news on the subject:

Major League Baseball owners approved the trial of two advanced instant-replay systems to be used in games beginning next week, an industry source confirmed to

Yahoo Sports first reported that MLB will use two different systems — one using radar and the other using cameras — to judge balls down each foul line. MLB implemented the technology in Yankee Stadium and Citi Field and will commence the trials next week when the Mets play host to the Rockies and Astros, and the following week, when the Yankees play host to the Blue Jays and Orioles.


For now, the systems will not influence the games. According to Yahoo, MLB wants to establish that the system is accurate and worthwhile before using it in other Major League ballparks.

Now, for those of you who are new to the Internet, opinions on this subject tend (and I think it’s actually a pretty strong tendency) to fall into one of two camps:

1. The human element of umpiring is a vital part of baseball, and removing it would irreparably damage the game.


Obviously, there is little middle ground between those two points of view. Given the site I write for and the fact that I have been mistaken for a robot myself from time to time, I would suspect that most readers would put me firmly in the second camp. For some time now, though, I’ve never really been certain where I fell, if I fell anywhere at all.

This is my attempt, then, to form a third way between the human element and the robots. Given my body of work up until this point, I leave it as an exercise to the reader to determine if this is me trying to find ground for compromise on a contentious issue or just trying to find a way to disagree with the largest number of people.

What I think every side already agrees on is that something is required to adjudicate the rules of baseball at the professional level. The stakes are too high to trust players to self-police honestly, and the game is played at too high a level and has too complex a system of rules to trust players to be able to self-police accurately even if that weren’t so. There is no camp that believes that some solution isn’t required. What we should establish are the criteria we use to determine what solution to use. Allow me, if you will, to propose the following criteria. Importantly, I am not numbering these, because I don’t want to imply that the order has any significance.

  • Inconspicuous: To the extent that it’s possible, we want whatever we use to determine calls not to interfere with the action on the field. That means we don’t want the presence of umpires to slow the game down, to block our view of the action or draw attention to the umpires themselves rather than the players on the field.
  • Unobtrusive: This may seem redundant after the first criterion, but it’s slightly different in meaning (at least as I’m using it here). What this means is that we want the umpires to have as little actual impact on events as possible. We want umpires to make the right calls as often as possible, and we want the calls they make to reflect what the players did rather than the umpire’s opinion of what the players did.

In order to do this, three things need to happen: something needs to be able to determine what happened on the field, determine what the rules say about that thing, and communicate that decision to the players. I should point out that there is nothing intrinsic to this process that requires it to be done by a human being, and no particularly human element that makes umpiring work. We have umpiring because it’s necessary; if it were unnecessary for the game to be played, we wouldn’t add it.

There are aspects of live, human umpiring that do bring enjoyment to baseball fans, I think—a distinctive call of “strike” or an enthusiastic hand motion when ringing up a batter can be interesting, to be sure. But I think we enjoy these things because of their association with baseball; we do not enjoy baseball because of its association with umpiring. Prizing the human element of umpiring, no matter the effect it has on the game itself, is like prizing a photograph of someone you love above the person themselves. So if we are to decide that human umps are better than the robots, it has to be because they’re better at achieving our objectives, not because we desire human umps for their own sakes.

Let’s just step through the process again:

  1. See the play,
  2. Determine how the rules apply, and
  3. Communicate that ruling to the field.

All of these elements take time. And before you respond that it takes time for players and managers to protest bad calls, too, let’s consider fair/foul calls. There are a few ways you could make these relatively speedy: you could have a replay ump up in the booth who is calling the play live, or you could use some sort of ball-tracking technology to determine fair or foul similar to how “in” or “out” calls are made in tennis. (I will note for a moment that tracking such things on a baseball field is at least an order of magnitude more difficult than doing so in tennis, and leave it at that.) But that gets you through two steps of our sequence; you still need to let the players know about the decision in time to determine their actions, unless you want every fielder and baserunner to play out every somewhat close foul ball until the booth can conclusively determine it was in fact a foul (which would add a fair amount of time to a ballgame).

Or consider a fielder diving for a ball in the outfield. Nobody argues that replay is more likely to determine correctly whether or not a catch was made. Intuitively, this would seem to make it more likely that the players, not the umpires, determined the outcome of the play. However, the other fielders and the baserunners have to act according to the call as it was made in the moment, not as the call will eventually be made. And if the call is eventually reversed, the only recourse is for the umpires to decide what ought to have happened on the play if the call had been made correctly to begin with. I think having an umpire rule incorrectly about what a fielder did is less troubling than having umpires try to determine what they think a player would have or should have done, had things gone differently.

None of this is to say that replay should not be used where it is possible to enhance the game. It’s to emphasize that getting the call correct is not the only criteria we need to consider, and that the time it takes to improve a relatively small number of plays may have other, undesirable effects on the game.

I was at Wrigley Field for Star Wars Day last season (don’t act surprised), and the Cubs were trailing for much of the game (don’t act surprised about this either). It was cold. It was raining. To this point, I had spent most of the game trying to distract myself from how badly the Cubs were doing by paying attention to a bunch of Star Wars characters and various meat and cheese products (that, right there, is a pretty good summary of my life to date). The bleacher section was totally exposed to the rain, so finally I retreated to the concessions stand to get out from under the rain (fine, fine, there was probably also bratwurst involved).

Then Carlos Pena jacked a ball out of the park, and the Cubs had a lead. Everyone was on their feet, cheering, complete strangers were giving each other hugs… and then an umpire trudged into the outfield and vanished through the ivy. And we waited.

And waited.

And waited some more.

And the umpire came out, and runners who had presumably scored instead took their bases, and there was no lead anymore.

Was it the correct call? Yes. Absolutely. Not arguing that. And Astros fans were probably as happy to see the call corrected as Cubs fans were saddened. But the larger point I’m trying to make here is that instead of cheering for moments, with enough replay, we are left cheering for reviews instead. And unless you can make those reviews unnoticeable, you are ever so slightly draining the magic away from the moments themselves.

Is that enough of a reason to shy away from instant replay? I don’t know. I think it’s enough of a reason to be cautious and patient in implementing it, though.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
I'm just thinking of NFL replay. People still cheer whenever a a catch ortouchdown is ruled. Replay doesn't take away from the cheering, sometimes it will even create an extra cheer when you know it's official. If in baseball they use replay with 2 outs or bases empty then umps can make the more impact full ruling knowing they can overturn it like how refs in football are calling more touchdowns because they all get reviewed.
Can we put #1 option on a t-shirt?
Number two, you mean? I only have Jay's permission to use his likeness for this article (no seriously, I checked), not any merchandising.
Is this not just an engineering problem though? If we could get the robots to process fast enough, then the calls could be instant.
In order to do a complex calculation on a computer, I used to have to fill out a punch card, walk it over to the mainframe building and then wait for the results. If I wanted to communicate the result back to the person who asked me the question, I had to walk it back over to them.

Now I just get the question via email from some remote location, pull my iPad out of my backpack and do the calculation, then transmit the answer wirelessly across the globe in a couple of nanoseconds.

If we had said "let's never do calculations this way because it takes too long" then I wouldn't have my iPad.

The point is, sometimes it takes time, but things improve over time, and until you get started, you don't know how much it will eventually improve.
If I were to summarize the two views I presented, it would be "robot umps never" and "robot umps now." To summarize my view similarly, it would be "robot umps maybe someday."
Balls and strikes are a much bigger deal. Laser beams to either provide visual help for the umpire or to automate calls. Would eliminate 95%+ of the frustration with umpiring.
Augmented reality glasses that show the accurate strike zone so all they have to do is watch whether the ball breaks that plain.
The positioning of Jaffe's thumb in that photo makes him look like a distant relative of Antonio Alfonseca.

Also echo what Tommy Fastball said. Ball/strike calls are far and away the heaviest influence that umpiring has on a ballgame, and I for one will welcome our strike-calling robot overlords.
It is so simple. A fifth ump in the press box with all the camera angles. Chief ump on the ground with an earpiece.

Review: plays on the bases, ground rules, home runs, fair/foul.

Corrections would take a matter of seconds.

David Blocher
Agreed. And I still think the fifth umpire should be a part of the crew and rotate through the four bases and the booth every 5 days. This way the guy upstairs is a "part of the team, helping get it right" rather than showing up the men in blue on the field. And that, (including the extra jobs created) is how you sell it to the Umpires' Association.

One of the big straw men that get trotted out by opponents of replay is the time it would take. Done right, it would only need a few seconds, and that would be more than offset by having managers come out and argue less often. (Of course, we lose some entertainment value that way...)
Agreed, except that even multiple camera angles can take time and can still be ambiguous in what they reveal, and potentially negate the purpose of the on-field umpires if we're trying to get every call right with the replays. Reviewing every close call would, I think, change the feel of the game.

I think we need to distinguish between an error of judgment, i.e. a close call that turns out to be wrong, and an error of execution, when a call is so unambiguously wrong that every single person watching knows it before the first replay. (The most obvious example, but only one of several every season, would be Galarraga's stolen perfect game.)

Errors of judgment I can live with, since judgment is the essence of an umpire's job, and the nature of judgment is subjectivity. Even if an umpire gets a call wrong, he's still doing his job as asked, and it's not clear that there exists an alternative for reducing the subjectivity of close calls without changing the feel or pace of the game.

It's the errors of execution I can't abide, because that's when an umpire demonstrates not judgment, but incompetence -- he is not, in fact, doing his job correctly. It happens; they're just people, and I'm sure even NASA engineers have an off-day sometimes. We can respect an umpire's authority to judge (the "human element") without having to tolerate when he simply fails to do so. We shouldn't have to accept as true something that everyone knows is false.

So my amendment to the above suggestion would be to let the booth ump overturn only the most egregious bad calls, the ones where there's no chance whatsoever that the on-field ump got it right. The overrule has to be made within 15 seconds of the play, or it stands as is; this will limit it to the errors of execution, and keep the game moving. And If we're at all concerned about the booth ump getting it wrong, there could be two, and unanimity would be needed for an overrule.

Then we can let the umpires do what they do best -- make judgments -- and baseball will continue to feel like baseball. But when an umpire fails to execute becuase his brain was on the fritz at that moment, we have a means of correcting what is, within the universe of baseball, gross injustice -- for his own sake, and those of the players and fans.

I don't know why this would be hard or controversial to implement.
This is an excellent comment. I agree completely. Well argued.
No. This is lousy. It is the same standard that the NFL uses and it sucks. Almost everyone can see what the correct call ought to have been on the replay, yet the officials ignore it because they lack "incontrovertible" evidence.

Let the umpires/referees use the replays to get a second look at the close plays so that they can make the best call possible with whatever means are available. Players and fans deserve no less than the best attempt to get the call correct and to hell with any artificial constraints on that process, particularly if they those constraints, like those in the NFL, are put there largely to soothe the egos of the game officials.

I can't be the only one who thinks the current home run review process was made that time-consuming and cumbersome only to "prove" that replay can't work in baseball. There is ZERO reason for the umps to have to retreat under the stands for five minutes to review the home run. A 5th ump in the box could make the correct call in 15 seconds.
How about this, to assuage the umpires union, hire another new, full, four man crew. Put them in the regular umpire ration, with a new stop. MLB HQ. All four umps are watching all the games, and are able to chime in when needed.

Instead of having Joe West waddle across to the diamond to say that he had a better view, he doesn't. The four man crew in New York gets the replay from 7 angles and the super-slo-mo, and makes the right call.
I suspect that, like many things, this can't really be framed as an either/or question. I would like to see the human element augmented by technology to improve the accuracy of the things we can quickly reverse without monkeying with the flow of the game.
I just want balls/strikes called electronically now.
Crowdsourcing is supposed to be an effective tool. So, let's just turn it into American Idol and let anyone watching at home or in the stadium text to 55555 with "Safe" or "Out". Based on the mid-inning promotions, the technology is definitely there. Besides, it can become another source of revenue.
Good point about review calls where the players must decide how to run the bases based on the call made on the field.

You should replace "arguing", as it appears twice in this article, with "disputing". Arguing is actually the opposite of disputing.
Whatever. Maybe not in five years, but probably in ten, no more than 15, balls and strikes will be called by computer. So will foul balls and home runs. Calls at first base not long after. Umpiring responsibilities will become largely limited to the application of the rules and not on physical events. It is as inevitable as the adoption of fielding gloves. Get used to it.
A few years ago, the Mets playing the Nationals in DC. Runner on first, none out. The batter hits a line drive into short right field. The right fielder comes in on the ball and makes a diving catch.

The umpire, however, incorrectly rules that the ball was trapped. Good for the batting team (Mets), right? Not so.

The baserunner, who went part of the way to second and had an unobstructed view of the play, sees the catch made and goes back to first. The batter running to first sees the umpire's safe signal, thinks he has a hit and rounds first. He passes the retreating baserunner and is called out just before the runner takes of for second to eliminate the force. Upon arriving at second, the runner finds a fielder waiting with ball in glove and is tagged out. The Nationals thus get a double play because of a bad call that went against them!

My question is, how would a replay official have handled this situtation? The Mets would have been in the odd position of challenging a call that went in their favor.