Rafael Nadal is today the most successful tennis player in the history of sport. His victory at the Australian Open allowed him to add his twenty-first Grand Slam and surpass Roger Federer and Novak Djokovic as the player with the most top-level tournaments ever. World No. 1 Djokovic is likely to overtake him at some point. But right now Nadal is the pinnacle of tennis history.
A feat against all odds.
The data. The previous phrase, so hackneyed in sports journalism, is in this case literal. During a moment of the match the AI in charge of predicting the winner projected onto the track of Melbourne a little encouraging data for Nadal’s interests: he had a 4% chance of being victorious compared to 96% for Medvedev. A few hours and a memorable comeback later, the Spaniard lifted his second Australian Open.
📸Al empezar el tercer set.. esto decía el #WinPredictor
96% Medvedev
4% Nadal#Radioestadio • #AusOpen pic.twitter.com/60Sq5nqK1V— Radioestadio (@Radioestadio) January 30, 2022
The context. It should be noted that Nadal was at no time the favorite for the match. Both the bookmakers and the algorithms responsible for predicting the result gave Medvedev a 64% chance of winning the tournament, compared to 34% for Nadal. His victory in the first two sets raised his chances to 96%. Only on five previous occasions has a tennis player come back from two sets down in a Grand Slam final. The feat was unlikely.
But it happened. Even more so when, in the middle of the third set dispute, Medvedev ran into with three balls breaking (0-40). If he had hit just one of them, he would have placed himself two games above Nadal (2-4), sinking his chances even more. Nadal, in a match that is now tennis history, successfully saved those three balls and recomposed himself to win the third set. From then on Medvedev would be systematically imposed.
Why predict? The 4% that the AI gave to Nadal at the beginning of the third set has caused certain sarcasm in the social networks. Actually, he wasn’t wrong. Medvedev entered the Australian Open as world number 2 and top seed. had won the previous major played on a hard surface and had only lost five sets on his way to the final. Computing all the variables and discounting other imponderables such as age or Nadal’s recent injury, 4% seemed a reasonable prediction. But statistics is not everything.
How does it work. All sports have relied for some time now on artificial intelligence to develop predictive models. For tennis there are many and they can be consulted online. Based on previous results, similar match situations, performance on the court type and other parameters, the AI, the algorithms, issue a prediction. These are advanced statistics and are usually very useful for a) preparers b) journalists interested in data and c) bookmakers.
Naturally, such predictions are snapshots of the moment. If something has happened many times in the past, it’s likely to happen again, right? Often, the sport is decided by imponderables where the statistics shipwreck. The mental toughness of a player like Nadal or the relative inexperience of Medvedev are not measurable data. Greatness, if we get poetic, either. Thus it expires at 4%.