Nobody’s Perfect: The Case for Artificially Intelligent Officiating

May 12, 2015; Houston, TX, USA; Los Angeles Clippers head coach Doc Rivers argues with official Sean Write (65) as then Clippers play the Houston Rockets in the second half in game five of the second round of the NBA Playoffs at Toyota Center. Rockets won 124 to 103. Mandatory Credit: Thomas B. Shea-USA TODAY Sports

Like most people, you only have time to read yesterday’s box scores first thing in the morning. Right before Chad walks over to your cube rambling on about “man did you see that call last night? I can’t believe they didn’t T him up!” Dumbfounded you shake your head yes, and blurt out “terrible call!”. You can’t have Chad questioning your manhood by admitting you binge watched Gilmore Girls on Netflix again instead of catching the game.

But what are you missing by only checking the box scores?

If you only read the box scores from this year’s Cavs-Bulls NBA playoffs series, for example, you would have seen that the Cavs won the series four games to two. The final three games going the Cavs way, with scores of 86-84, 106-101, and 94-73. The box scores let you know that the series ended after Game 6 in Chicago, in a 2nd half route. But was the series decided in Game 6?

The Bulls went down in typical Tom Thibodeau defense first fashion. Despite being down double digits, Thibs went with his no offense all-stars. Kirk Hinrich manned a crucial guard spot over 3-point specialist Tony Snell. Defensive and missed layup specialist, Joakim Noah, played over Pau Gasol. And Mike “I can’t believe I’m playing 30 minutes a night” Dunleavy played over 2014-15 All-Rookie First Team player, Nikola Mirotic.

What that box score missed was not that Tom Thibodeau may have coached his last game for Chicago. But that the series itself finished much earlier. In the waning seconds of Game 4, Derrick Rose drove into the lane and made a difficult layup to tie the game 84-84. Following that play, Cavs coach David Blatt channeled his inner Chris Webber, and ran onto the court to call timeout. Just like Webber’s Wolverines, Cleveland was out of timeouts. Unlike Webber, the officials missed the call. Instead of the Cavs being down 3-1, the series was now tied at 2-2. Saving the Cavs season and David Blatt’s NBA coaching career.

Game 5 was not much better for the Bulls or for Bulls fans. The box score showed Chicago losing another close game 106-101. The game turned after a crucial missed call sent Taj Gibson to the locker room and put the Bulls in an insurmountable hole. Ejected after kicking Cavaliers guard, Matthew Dellavedova. Replays showed Gibson retaliating on a Rick Flair approved leg-lock from the Cavs guard. The call was so egregious the NBA reviewed the play and assigned a technical to Dellavedova the next day for his antics. The refs got it wrong even after they went to the monitor to review the call. Read that last sentence again, and ask yourself how often that rings true.

Other recent examples of questionable officiating:
– April 6, 2015 – The entire ending of the 2015 Duke-Wisconsin National Championship College Basketball Game.
– January 11, 2015 – Dez Bryant’s “dropped ball” in the 2015 Cowboys-Packers NFC Playoffs.
– January 4, 2015 – Blatant hold on Suh in the 2015 Cowboys-Lions NFC Playoffs. Suh is one of the dirtiest players in the NFL, and we’re not losing any sleep over this one.
– September 24, 2012 – The NFL replacement ref debacle. Otherwise known as the time the NFL found a group of dudes that stayed at a Holiday Inn Express the night before. Led to someone spending hours creating this enormous Wikipedia article about a questionable call in the Seahawks-Packers game.
– June 2, 2010 – Armando Gallaraga’s botched perfect game. A call so bad that the umpire and pitcher would go on to write a book together, called “Nobody’s Perfect.”

“To err is human, to forgive divine” – Alexander Pope

Human beings are prone to making mistakes. We all make mistakes every single day. Some of us make mistakes all day long (see my ex-girlfriend). Joyce and Gallaraga’s co-authored “Nobody’s Perfect” is a description of the human element summarized in two ironically perfect words.

In the wake of Gallaraga’s controversial lost perfect game, ESPN’s “Outside the Lines” conducted a two week study on how often umpires made the correct call (excluding balls and strikes). They discovered an astonishing 20% of all close calls were incorrect.

A 2014 article by Beyond the Box Score, identified incorrect calls on 15% of all pitches thrown. 38 umpires made the wrong call over 15% of the time, compared to a mere 11 umpires wrong less than 15% of the time. The fact is umpires are biased and inconsistent on balls and strikes, with even more detail on the phenomenon in this FiveThirtyEight article. This is especially frustrating when you take into consideration that pitch tracking technology has been around for years.

In 2014, Bloomberg took a look at some NBA data. Over a two week period, the study found that refs messed up 14% of the time in crucial close late game situations.

The NFL does a great job of protecting it’s image and information on incorrect calls is scarce. A 2008 study showed the league reported that only 2% of all calls were mistakes. With that said, there are still bad games like the 2010 Super Bowl, where crucial calls went against the Seattle Seahawks.

If you want to see how deep the rabbit hole goes, take the red pill and strap on in to the Tim Donaghy experience. Corruption and officiating go together like peas and carrots.

Those facts can be pretty depressing. But we are alive in an era where we may see answers to many of our current problems, including one of the less important issues: The officiating problem.

There are companies out there escorting us into the next generation. Companies like Lily are pioneering self-flying cameras. An evolution of these cameras may in theory make it possible to see any angle necessary to make the right call. We have companies like NextVR pushing the envelope on Virtual Reality. Bad calls are more noticeable when you have the best seats in the house. And we have STATS LLC, a company specializing in sports technology and data collection. They are acquiring and partnering with all kinds of cool companies. And one of these companies has the potential to create Artificial Intelligence officiating. Technological evolution may be slow when it comes to officiating, but it is happening. In the last few years alone, we’ve seen a massive increase in use of replay technology in all major professional sports.

If we have the data from the camera angles, what’s stopping us from using that data to create a level playing field? An algorithm that identifies the correct call seems reasonable. It’s not a matter of if this will happen, but how long.
A logical progression might look something like this:

Step 1 –  officials move away from arenas, into centralized locations. Viewing live feeds of every angle. Eliminating “homecourt” bias where fearful officials make close calls in favor of the home team.

Step 2 – the creation of an algorithm that determines the call based on probability analysis. Close calls determined by a human judge.

Step 3 – creation of an Artificial Intelligence whose only programming is to make the correct call. The AI will analyze the data against all available information and determine the correct call. Margin of error is negligible.

Law of Accelerating Returns states that rate of change in a wide variety of evolutionary systems tends to increase exponentially. If this holds true, we may not be that far off. We already have wearables that can determine foot strike and prevent injuries. Soon we may have nanosensors imbedded in jerseys that detect force necessary for something like a basketball foul.

If you’re interested in reading more about Artificial Intelligence, checkout an easy to understand breakdown in this Wait But Why article. If you’re not interested in learning more about Artificial Intelligence, well you should be. There’s a reason why guys like Elon Musk, Bill Gates and Stephen Hawking are concerned. For the purpose of this article we are talking about a one dimensional Artificial Intelligence (Artificial Narrow Intelligence), specifically one created with the task of calling a game correctly. The end of the Joey Crawford race, not Skynet ushering in the end of the human race.