The recently completed Australian Open provided plenty of the excitement due a Grand Slam event. It also provided the opportunity for statisticians to demonstrate – using a method first used to rank chess players – the secrets to success.
Using data from 1990 to the present day, male Grand Slam winners were ranked based on where they finished at each Grand Slam tournament until they first won a Grand Slam event.
Deploying Elo ratings (named after the physicist who developed the method to rank chess players) means ranking each match based on measures such as the quality of the competitor. Elo ratings are a stronger predictor of performance than official rankings, because they typically take into account a longer-range of more relative data.
To win a Grand Slam, the data below shows that an Elo rating of 2,000 or greater is required.
Even with that Elo rating, the data then shows the player needs to win 25 or more Grand Slam matches to be a chance to win a Grand Slam tournament, and what is more, needs to win those 25 matches in their first 200 total matches won.
Developed by the Australian Open Game Insight Group, in partnership with Victoria University, these rankings show its tough to win a Grand Slam. It also shows that there are many uses of data being pulled together in a manner that can predict the likelihood of something occurring – and even when it might occur.
