Artificial Intelligence & Baseball
As early as 2002, Billy Beane and the Oakland A’s of Moneyball turned over power from scouts to computers as they began to let analytics determine their front-office decisions. The analytics they relied on were relatively basic, such as on-base percentage, slugging percentage, and walks/hits per inning pitched.
In the past few years, the Houston Astros have led baseball’s second technological revolution. The 2017 Astros had one of the most dominant teams of recent memory, and much of that can be described by the legal technology that the team utilized.
The Astros assembled a “Nerd-cave” in Minute Maid Park that consisted of Ivy-League math gurus, computer scientists, and even a NASA rocket scientist.
The “nerds” worked tirelessly.
Researching, running the numbers, doing whatever they could behind a computer they produced the best product on the field. What they found helped them take one of the worst teams in baseball history (back-to-back-to-back hundred loss seasons) to become the 2017 champs.
Houston’s team was built upon analytics that combined just about everything, from simple baseball statistics to complex algorithms.
For example, Astros mathematicians discovered through countless models that the spin rate of a pitcher's off-speed pitch, a statistic overlooked by just about everybody, was highly related to the pitcher’s success. So, of course, they targeted the pitchers who had the highest spin rate on their curveballs, yet rarely threw them.
Today, hundreds of in-game decisions are made using analytics that take into account each batter and pitcher: Should we steal? Where should we position our outfielders? Do we even need anybody covering the left side of the infield against this hitter? Is it worth bringing in a lefty to face their next left-handed hitter, who has hit significantly better against righties in his career?
As much as analytics can help a manager and front-office make their decisions, analytics, however, contain a major flaw. They only assist humans in making decisions using historical data. They can tell you how a batter has fared against pitchers with certain pitches, spin-rates, arm-slots, velocities, and release points, but what if a hitter hasn’t faced a certain blend of all those factors?
Or if they have, say the sample size is too small? Analytics fail to give us an answer.
However, what if instead of running models based on the past, we had simulations that could combine historical data points with current and possible future conditions, automatically updating its inputs to predict how certain scenarios will play out? That is where we turn from analytics to artificial intelligence.
Instead of running models based on the past, AI gives us the opportunity to look into the future. With AI, we can simultaneously run hundreds of thousands of simulations based on everything from possible weather patterns, to spot in the lineup, all the way to factors such as how many hours of sleep the hitter had the night before.
An AI-powered scouting tool, Draftpoint, is currently helping teams sift through tens of thousands of baseball prospects all over the world.
Draftpoint “provides the next level in scouting and with complex algorithms can do things that humans can’t—pull out key words, phrases and similarities in scouting and provide comps to similar players,” confirmed Baseball America editor B.J. Schecter. Shecter continued, “The ability to sift through thousands of scouting reports and combining it with decades of BA scouting reports will give teams an invaluable tool.”
The Astros, attempting to find the right balance between numbers and humans, have an AI-powered system that gives grades to scouts based on the correctness of their evaluations. The system, of course powered by a computer, then predicts which scouts have an eye to see things that computers miss.
Artificial Intelligence won’t be used in baseball solely by front-offices, trying to increase the output of their team, AI also extends into the realms of player health, fan experience, and ticket sales.
AI is already changing the healthcare industry in a number of ways, which are being implemented by sports teams to maintain the highest physical well-being of players on and off the field. Athletes frequently undergo tests that utilize AI to analyze a number of health paraments and movement to detect early signs of injuries.
The quality of one’s sleep, for example, is a major indicator of the physical success they will have in the coming day, and AI is used to analyze sleep patterns. With good sleep, performance increases along with the functionality of your immune system. Artificial Intelligence has been proven to better detect sleep apnea. For right-handed pitcher Josh James, his diagnosis and treatment of sleep apnea took him from a solid minor league pitcher to one of MLB’s most exciting 100 mph arms.
Satisfi Labs, a major MLB partner since 2016, recently announced plans to work with MLB teams, using AI to sell tickets. Using artificial intelligence, teams can analyze more precisely the factors that enhance the fan experience. Satisfy Labs inputs conversations that fans have with guest-services staff to conduct their research.
Before Satisfi Labs began their research, MLB teams, for the most part, didn’t appreciate that different people approach ticket purchases in different ways. Some fans only care about buying the cheapest tickets available, some will only purchase tickets if the most expensive seats are available, many care about the opponent, and some may only care about the price of beer.
The AI at Satisfi Labs is providing teams with the information they need to approach ticket sales with a profit maximizing focus directed towards fan experience.
When it comes to streaming games from our couches, AI has room to revolutionize the way we watch. It can automatically provide subtitles in a number of different languages, read lips of players on the field and in the dugout, automatically choose the right camera angle to display on screen, and identify the right times to display commercials and promotions, a win-win for both the viewers watching and the companies advertising.
Written by Ethan Samuels
Edited by Michael Ding & Alexander Fleiss