Three stewards took over 20 minutes to decide whether to disqualify Maximum Security in the 2019 Kentucky Derby. Artificial intelligence (AI) could have made a decision instantanteously and more scientifically. Following the lead of sports like professional baseball and world-class gymnastics, racetracks and state racing authorities should explore how to use AI to render faster and better DQ rulings. This article suggests a methodology.
Artificial intelligence conjures up an image of computers thinking like humans and even having emotions. This advanced form of artificial intelligence, in which a machine has cognitive abilities, is commonly referred to as General Artificial Intelligence and is likely at least decades away from coming to fruition, if ever.
Today’s version of artificial intelligence, or simply AI, denotes machine learning. Software learns from being fed vasts amounts of data and is eventually able to make distinctions with minimal or no human intervention. For example, software could be taught to recognize various shades of the color blue or to identify trading signals or patterns in the stock market that produce profits.
Artificial Intelligence is being used in many endeavors, including sports. To illustrate, Major League Baseball is testing AI’ s ability to call balls and strikes more correctly than human umpires. The Atlantic League (comprised of eight independent professional baseball teams) is experimenting with AI this summer as described in the Wall Street Journal:
“…last week marked the introduction of…an automated strike zone, shifting responsibility for calling balls and strikes from a person to an emotionless piece of technology free of the biases and inconsistencies of mere humans. And if the test goes well, the days of big-league players imploring umps to schedule an eye exam could soon come to an end.
[Long Island] Ducks manager Wally Backman predicted that MLB will adopt the system within five years.
‘It’s going to happen,’ he said. ‘There have been a few pitches that are questionable, but not as many as if it was a human. The machine is definitely going to be more right than they are.’
Every Atlantic League stadium…now features a TrackMan device perched high above the plate. It uses 3-D Doppler radar to register balls and strikes and relays its ‘decision’ through a secure Wi-Fi network to the umpire, equipped with an iPhone in his pocket connected to a wired earbud. That umpire, positioned behind the plate as normal, hears a man’s voice saying ‘ball’ or ‘strike’ and then signals the verdict.”
Similarly, another recent Journal article titled “The Robots Are Coming (to Gymnastics)” stated: “The international gymnastics federation has voted to expand the use of sophisticated new technology in major upcoming competitions, a move that could ultimately shift control over the sport’s judging from humans to robots.”
Here’s is a brief summary of how artificial intelligence conceivably could be developed for use by horse-racing stewards.
The first step would be to employ machine learning to perfect a “best practices” AI model that stewards would have at their disposal to make decisions about whether to uphold or reject claims of interference in races and would ensure more consistent rulings across racetracks. By inputting data (videos) from thousands of past races in which a claim of foul was lodged or in which stewards initiated an inquiry on their own, AI software might be able to learn to become a highly accurate surrogate steward free of any personal biases or emotional leanings. If this proves infeasible, an AI “best practices” model could be built upon knowledge solicited from the most experienced and competent stewards.
The next step would be to tailor a Doppler-radar system for tracking the paths horses take during races. Data derived from such a system would be plugged into the AI model to determine whether each and every race has been run fairly, or whether interference requires a disqualification. According to the National Weather Service: “By their design, Doppler radar systems can provide information regarding the movement of targets as well as their position.” Thus Doppler Radar has the potential to provide stewards with precise data pertaining to the paths horses took during a race.
For instance, in the 2019 Kentucky Derby, the availability of an AI model fed Doppler radar data from the race could have immediately told the Churchill Downs stewards whether the two claims of foul against Maximum Security were warranted. They would have had the benefit of scientific data about Maximum Security’s route vis-a-vis other horses and the collective wisdom of stewards embodied in an AI model indicating whether a DQ was appropriate.
Regardless of whether stewards are bound by International Federation of Horseracing Category 1 or Category 2 rules for disqualifying horses, artificial intelligence would assist them greatly in reaching sound, timely, and defensible calls, even though they could still have the option to disagree with the decision recommended by the AI software.
Copyright © 2019 Horse Racing Business
Recent Comments