If You Want To Do Instant Replay Right
This article is very unlike anything I have written on this blog before. It has nothing whatsoever to do with the market, the economy, or finance. It involves something that is far more important: football. (It therefore is likely to be my most-popular blog post ever.)
A great frustration for any fan of American football is the process known as “instant replay,” where an official’s live ruling is questioned and potentially reversed by means of careful study of video taken of the play from a bunch of different angles. The process slows down the game, adds a layer of odd strategy and introduces some quirky rules (such as the bizarre rule, obviously introduced by sadists, that a scoring play is always reviewed automatically, unless it’s challenged by a coach in which case it cannot be reviewed no matter how bad the call was).
But the real problem with instant replay boils down to this: it was introduced as a way to “correct” heinous calls made on the field, and it is marketed as such. The idea is to correct Type II errors – correcting the on-the-field failure to make the right ruling – but without introducing Type I errors – incorrectly changing an on-the-field ruling that was in fact correct. Consequently, the rule states that the ruling on the field stands unless there is “indisputable visual evidence” that the ruling was made in error.
But in practice, this standard doesn’t work. Moreover, it cannot work. “Indisputable” is an impossibly high standard that should ensure that calls are very rarely overturned; however, if they are very rarely overturned then there is little point to allowing so many challenges. And, in practice, something far different from “indisputable evidence” is actually implemented.
Consider the following play, which occurred at the end of a furious comeback by the Dallas Cowboys against the New York Giants in week 8 of the current NFL season. The Cowboys, at one point down by a score of 23-0, scored an apparent touchdown to take the lead with 10 seconds left when receiver Dez Bryant made a catch in the end zone. It was ruled a touchdown on the field, but on further review the catch was waved off and ruled an incomplete pass, after inspection of the video concluded that there was “indisputable evidence” that Bryant’s pinky finger had brushed the grass out-of-bounds before his elbow landed, making the play dead and the pass incomplete. Here are two images of the catch. Keep in mind that in order to overturn the call, you must be able to state without fear of dispute that the white-gloved pinky grazed the white-painted grass out-of-bounds.
Now, it is very likely that Bryant was out-of-bounds on this play, but the standard is “indisputable.” Given the rule, that was an incorrect call.
Yet it is very likely that Bryant was, in fact, out of bounds. In a room of ten (neutral) observers, probably seven or eight would say he was out while two or three would say he was not, or that they couldn’t tell. That’s not close to indisputable, but it is probable. The legal standard would be “preponderance of the evidence” rather than “beyond a reasonable doubt,” and if that was the standard the Cowboys still lose the game. In a cosmic sense, that’s probably a just result although it’s the wrong football ruling given the current rules.
This happens, and is going to continue to happen, as long as the replay rules are enforced in this way. But there is a better way to do it.
The way instant replay is currently implemented involves an official or officials “up in the booth” as well as the referee on the field. It is the referee who makes the call, after looking at the reply, about whether his own call (or the call of someone on his team) was correct or not. This is absurd on two levels: first, that one person makes the call, and second, that the official makes a re-judgment of his own rulings.
If I were designing an instant replay system, here is how I would do it:
- I would hire, say, 50 officials, trained in the rules as referees, to sit in a room at NFL HQ. (Technologically, there is no reason they couldn’t be located around the country rather than in one room, but this helps the image).
- When a play comes under review, each official looks carefully at the play and without consulting the other officials presses one of three buttons: “let play stand,” “reverse play,” or “indeterminate.”
- A tally is taken, and if the vote exceeds a predetermined threshold then the play is reversed.
The advantage of this system is that we can set the dial on the “pre-determined threshold” wherever we want. If we truly want “indisputable” evidence, then it needs to be unanimous to overturn. But more reasonable is that we set a supermajority level like 75% or 80% to overturn. This recognizes the fact that almost nothing is “indisputable,” and allows the league to set the evidentiary standard wherever it wants to. The referee could even report the result: “with only 55% of officials disagreeing with the call, the play stands as ruled on the field.”
In fact, this opens up more neat options. We could have different evidentiary standards for scoring plays compared to turnover plays, for first-quarter plays compared to fourth-quarter plays, for playoff games compared to regular season games. It wouldn’t cost the league very much at all, because the “upstairs replay official” at each game could be eliminated. The technology certainly exists to do this. And it could be done more quickly than the current method allows.
That is how to do instant replay correctly. I think it’s indisputable!