No Escape from Alcatraz

And Why Our Laws Are Obsolete in the age of AI

Machines Run Sports, What’s Next?

What a US Open. I’m still having withdrawal symptoms. A set of new faces emerged this year, including Carlos Alcaraz, who I’ve nicknamed Alcatraz. It’s hard to escape his clutches. It’s the first time I’ve seen John McEnroe in love. He couldn’t stop raving about Alcatraz during the tournament. “He has it all,” said Johnny Mac. He sure does.

It is noticeable  that the linesmen are gone, replaced by machines. Which never make a mistake. There were plenty of times when it looked like the ball was clearly out, but the “close call” replay always showed the truth.

It took me back to the 1980s and all the times McEnroe ranted on the court during his prime. I wonder how many of those calls were wrong, both ways. We will never know. But it makes me think about all the honest mistakes professionals make in all areas of our lives, especially health, and we never know the truth.

Ironically, the humans running the show have essentially become robots. A human still calls the ball “out” after the machine tells him what to say. I’m sure the machine will say it as well in the future. The role of the umpire is similarly robotic now, reduced to maintaining order: tell the crowd to calm down if the decibel level exceeds a threshold, and warn spectators to keep quite during points.

Our Justice & Insurance Systems Are Next

Our AI vision systems have become really good. They will create all kinds of “disruptions,” a term used by management scholar Clay Christensen to describe what happens when the capability of a technology exceeds some threshold, enabling it to do something that wasn’t previously possible. The Internet has been one of the biggest disruptors of our time.

In sports, the disruption of eliminating human judgment is driven by the need for accuracy, period. We don’t need to look far for high-stakes games where a referee’s call decided the outcome. I recall a Euro soccer final years ago featuring France vs Croatia, where Croatia was pressing the French team which was on the back foot, with an occasional counter attack. On one of these counters, the French striker Antoine Griezmann fell near the Croatian box. The referee awarded Griezmann a free kick and France scored, which completely changed the rest of the match. The outcome would probably have been completely different if  Griezmann had received a yellow card for diving instead of the free kick.

Virtually every important sport now relies on machines to make the calls. The cost of error is too high to rely on what Daniel Kahneman calls “noisy” humans.

But noisy humans still make the important calls in our lives, such as in the justice system, where accuracy is arguably even more important. As Kahneman observes:

“Even if you take the crime for which the average sentence is seven years in prison, that difference between two randomly selected judges is going to be four years. So that's the average difference. If you pick two judges at random, and 50% of the time, the difference between two judges will be bigger than four years. This is really shocking. That's noise.” (Brave New World, Episode 21 with Daniel Kahneman)

It remains to be seen what part of human judgment machines will replace or augment in the justice system. But it will happen, and will require an effective system for appeals.

Insurance judgments seem to be crying out for machines. Many of our insurance laws are based on outdated assumptions about what is objectively verifiable. I had a personal experience several in 2018 that is illustrative.

One evening, I was driving uptown, waiting at a red light in Union Square. Suddenly, I heard what sounded like an explosion where my head snapped back and then hit the steering wheel, all in an instant. Fearing an explosion, I leapt out of the door, onto the divider on Park Avenue. A TLC (Taxi and Limo Commission) Escalade had slammed into my rear so hard that my car bumped into the taxi in front of me. For the sleuth reader, that’s why my head had snapped back and forth. The driver walked over to me and said “sorry man, I fell asleep at the wheel. It was clearly my fault.”

He didn’t look sleepy. I figured he must have been texting, but that’s against the law. What could I say?

New York’s finest took about 4 hours to show up to record the incident. When I called 911, I had made the mistake of telling them that I was okay and unharmed. I was later informed that such delays can happen with non-critical cases around shift changes. Evidently cops at the end of their shift are reluctant to take on new cases, and the new shift can take a while to process them.

The cops took some summary notes. I was exhausted after the 5-hour ordeal and just wanted to go home. When I did look at the notes, they were highly incomplete and vague. I went to the police station the next day and met the officers who had recorded the incident. They told me to just go away. I couldn’t  help note the sides of the cars as I walked out of the precinct: Courtesy, Respect, Professionalism. I must have experienced the exception.

My insurance company never got reimbursed from the Escalade’s insurance company. My premium went up. I spent several days of my time and several thousand dollars getting my car fixed.

What’s wrong with this situation?

The reason is the “no fault” insurance laws in many states, which assume no fault of either party so that cases don’t get tied up in court. The law is well intentioned, but it is based on the assumption that the truth is not knowable. Rather, each party must make its case. It’s a slow and expensive process. No fault laws were enacted to speed up the justice system, which would otherwise get entangled in endless “he said, he said” cycles. The case is finally settled on the basis of imperfect evidence and noisy human judgment.

I’m sure a bunch of street cameras in Union Square recorded my incident. However, that information is highly “illiquid” at the moment, and accessible only when major crime is involved, like a terror attack. And there are privacy issues involved. But what if my car is equipped with a camera and sensors that record the incident? Would such evidence be admissible in deciding a case? It should be. All of this is possible with the current state of the art technology. It’s a matter of time before such laws are amended.

I have little doubt that in a few decades, justice will be based more on data, instead of outdated assumptions under which many of our laws operate at the moment. That will be a much better and fairer world.

Meanwhile Back at Alcatraz

Let’s get back to the US Open final? What makes Alcatraz so impregnable?

Interestingly, there are some general lessons on winning  that we can take away from him. They apply to all of us in our daily lives, from individuals to business leaders.

The first lesson is, if the math has been done correctly, trust it unless it is obvious that it doesn’t apply anymore (this can be a difficult determination, that I have discussed elsewhere, but that’s another story). Alcatraz had no hesitation in going back to his shots, even if they kept failing. He hit an amazing drop shot winner in the final set, after a string of failed drop shots. Clearly, he believed in “mean reversion,” meaning that he should expect his success rate to vary around some average, and the variance was “normal.” He didn’t let failure mess with his head. Rather, as McEnroe noted in his usual pithy style, “he has the ability to get out of difficult situations and get positive quickly.”

But his killer weapon was his real-time decision-making, where he employed what people in AI call “the least commitment principle.” The principle says that in decision-making situations, make as few assumptions as possible, ideally zero. Don’t assume that the opponent will move in a certain direction, but wait for the evidence before acting. Alcatraz was able to delay his decisions much more than his opponents. When rushing to the net to retrieve a drop shot, for example, he would keep his eyes peeled on the ball, as well as the opponent’s move through peripheral vision. Almost always, he was able to act based on the latest information. That’s a huge advantage over an opponent.

But what put him at such an informational advantage in the first place? Why couldn’t his opponents do the same, and nullify his advantage? The answer here is fitness and tenacity. Alcatraz is 19 and super fit. And he makes Pepé Le Pew look lame when it comes to tenacity. His speed and fitness position him to be able to observe his opponent, who is typically scrambling to put himself in position.

I agree with Johnny Mac. We’ve got a new phenom and we ain’t seen nuttin’ yet. There’s a lot to learn from Alcatraz.