Artificial Intelligence in Sports

Artificial Intelligence in Sports

Why technology needs to be the first name on your team sheet As humans, we innately love sports. It is thrilling, dramatic and speaks to our inborn combative nature. It divides cities, yet unites countries. It provides an entire vocabulary that transcends borders. Games and matches can be won or lost by the merest of fractions. Every sporting event is a script whose ending is not yet written.

Why technology needs to be the first name on your team sheet

As humans, we innately love sports. It is thrilling, dramatic and speaks to our inborn combative nature. It divides cities, yet unites countries. It provides an entire vocabulary that transcends borders. Games and matches can be won or lost by the merest of fractions. Every sporting event is a script whose ending is not yet written.

However, like it or not, sports are also a multi-billion- dollar industry. They employ millions. Lives and livelihoods can rest on the outcome of contests. Whether it is the NFL, the World Series, the Soccer World Cup or the Olympics…they are all big business!

So, it’s no surprise that as the influence of sport increases, it should combine with another enormous and influential sector: Technology.

In this article, we look at a couple of examples of how tech is influencing the world of sports, and allowing coaches to get those all-too-important fractional improvements from their players and teams.

Why is Tech important in sports?

The basic answer is that humans are no longer fully capable of covering the sheer amount of information generated by the sporting world. There are a number of other reasons too:

  1. Cognitive bias. This phrase explains how a human can misinterpret a situation, or a set of data because of pre-held beliefs. It describes how subjectivity can cloud an opinion that should be firmly based on objective fact. It can lead to systematic errors and wrong decisions. Computers do not suffer from cognitive bias; they make ‘blank sheet’ assumptions based solely on the facts.
  1. Hidden facts. Humans are complicated; they do not always present all the facts about themselves…often they don’t know all the facts! Consider then sports recruitment, where interviews play a key part. How can an interview be error-proofed if it is reliant on human interaction which will by definition be unreliable?
  1. The decision-maker. Big decisions in sports are usually made by one individual – the manager or coach. These decisions will be distillations of enormous amounts of data from the medical team, coaches, talent scouts, the media…even the general public. That the decisions come down to one person, who is likely incapable of sorting and using this data effectively is a huge weakness. They are also unlikely to be able to spot patterns within the data.

Case Study One: MoneyBall

Perhaps the best known application of data science in sports is also one of the first. So well-known in fact that it spawned a successful Hollywood film.

The principle of MoneyBall is based on a new approach to data that was used in 2002 by the then general manager of the Oakland Athletics baseball team, Billy Beane.

The A’s, as they are known, faced an enormous challenge: How were they expected to compete against the giants of baseball and their respective spending power, when their own budget was so small?

Through a rigorous process of statistical analysis, the A’s management team were able to determine that certain unused indicators were actually more useful than other more conventional indicators. This meant that individual players were being overlooked, even though there was the distinct possibility that they would actually make the grade. And often the overlooked players were lower profile, commanding lower salaries.

This doesn’t of course mandate the use of technology. This job could be done manually. But, by applying technology, the range of data to be analysed increases exponentially, and therefore the chances of finding a hidden gem increases.

Furthermore, with the advent of machine learning, a predictive model can be applied. This will help the recruitment team to continually refine and improve their searches, further reducing the risk of bad appointments.

Case Study Two: Grand Slam Tennis

Tennis has embraced technology, specifically AI, over the last decade as a way of supplementing coaching in order to improve player performance. It does this by breaking down an overall performance into individual moments, then investigating these moments in immense detail. A tech solution trumps a human solution here due to:

  1. The huge amounts of data involved
  2. The ability to quickly train to build accurate algorithms
  3. The ability to detect areas of significance by ‘layering’ data. ‘Layering’ means finding patterns in data which would be undetected by humans

The introduction of AI into tennis can be traced back to 2015, where data from a grand slam match was fed into a computer. It determined that 70% of all points involved a player hitting the ball only two times. And yet, coaching had previously concentrated on the consistency needed in longer rallies. This coaching had actually been concentrating on only a small area of significance within the sport.

And, as more data gets fed in, the predictions become more accurate.

As of 2022, more than 1,000 players and coaches are now using AI technology.

One additional spin off has led to an improvement in the watching experience too. The same AI can enable the audience to select highlights much more quickly than before, enabling laser-sighted content to be available to audiences in a fraction of the time.

Case Study Three: NFL

We know that a sports person’s physical ability is only one aspect of their propensity to succeed; just as important is their behaviour.  Deep Data Insight worked with the NFL to produce a platform called Fitinsights; this was created to address the problem of behavioural issues with players.

Fitinsights evaluates the ability of sportspeople to navigate risk, optimize strengths and develop untapped potential. The platform utilizes behavioral and social data to predict the risk associated with defined risk categories of team players.

The risk behaviors associated with five areas were used: aggression violence, delinquent behavior, emotional distress, sexual behavior, and drug alcohol behavior. They were analyzed by extracting data from Twitter and Instagram profiles of the players. This data was then passed through five separate risk models, returning a score between 0 and 1 for each of the five categories.

The resulting reports made it easier for psychologists and recruiters to find information about the players that would not have been obvious to humans…or would have taken massive amounts of time to unearth.

What is PERC3PT, and how does it use AI in sports?

PERC3PT is an AI powered psychometric product designed by Deep Data Insight. PERC3PT will reduce the costs and risks associated with team management and recruitment. It works by applying AI to social media feeds, producing a series of reports that can also monitor change over time.

Stage 1 – Candidate details (name/ age/ location etc) are fed into PERC3PT, which trawls social media for data.

Stage 2 -PERC3PT uses Natural Language Processing to prepare the data.

Stage 3 – PERC3PT uses AI to interrogate this data and analyze for risk.

Stage 4 – PERC3PT draws conclusions using Big Data and presents these as a ‘Big 5’ personality trait report, Sentiment report, Risk report and Deep Psychological Profile.

For more information about PERC3PT, get in touch https://www.deepdatainsight.com/contact/

Share this post