Statistical uncertainty in tennis: How Federer lost the Wimbledon 2019 final despite winning more points than Djokovic
Whether it's a heated debate with our colleagues regarding the economic growth under the present regime or a mild discussion about the weather pattern of our city, we love quoting relevant statistics to back our arguments and establish our point. If we go with a popular cliche, stats don't lie. But while that is true, sometimes, instead of revealing secrets, they insulate us from the reality check.
The Wimbledon 2019 men's singles final is a good example that sheds light on the perplexing nature of statistics. Roger Federer was miles ahead of Novak Djokovic in nearly every aspect of the game but still ended up on losing side.
Federer had more aces, fewer double faults, a higher share of net points won and a better conversion rate of break points. Federer even won more receiving points than Djokovic, who is arguably the best returner in the game.
In layman terms, Federer won more points in the match but failed to win the match.
The Wimbledon final is a classic example of what is known as Simpson's Paradox in statistical parlance. It is basically a statistical trend that emerges when the segmentation of a dataset is reversed upon the aggregation of the dataset.
Contrary to popular belief, all points in tennis are neither identical nor equally important. The peculiarity of the scoring system makes some points more significant than the others in terms of altering the flow of the game. Andre Agassi once said about the oddity of the scoring pattern, "It was invented to cause frustration for those who choose to play. Because it makes no sense."
Federer was statistically better, but Djokovic won the mental game. While Federer was playing at an elite level, his opponent was far from his best. What Djokovic did better was to win points at big occasions without losing his equanimity.
Here's a look at the highlights of the match:
Djokovic was outscored by Federer in two of the three sets he won. Nonetheless, he won all three of them on tiebreakers which supposedly favors the comparatively inconsistent player.
A new study by Columbia Univerity's Department of Statistics suggests that the winning chances of the less consistent player can improve by as much as 27.45% when a tie-break is played. And if the player is a big server, then it's the icing on the cake as he/she can simply hold on to their serve and let it get into the tie-break which favours the less consistent player.
In a high voltage match like this, the break point is the most crucial moment that every player waits for as this is the point where you are trying to move one step ahead and gain a psychological advantage over the opponent. Federer outscored Djokovic in this domain too.
However, break point conversion does not include the player's efficiency at the most critical points - a match won in tie-breaks.
Federer made 11 unforced errors in three tie-breaks as compared to Djokovic who produced none. This is the most pertinent statistic of the match. It shows how clinical Djokovic was when the match was on the line while Federer was unable to control his nerves in tense situations.
Federer won 51.7% of the points played, which is enough to win a match in most cases. However, this figure treats every point the same - which is not the case in tennis.
Winning points at the deciding stage of the match are not similar to the points won at regular stages. Djokovic saved two match points in the 5th set to keep his hopes alive. He saved his second match point by hitting a cross-court passing shot with very little margin for error.
The irregularity of the point distribution system in tennis makes the game more of a mental battle than a statistical one. The mental strength of a player is tested after every point they win or lose. Even when Djokovic was down by two match points, he looked very calm and undaunted by the ambiance inside the arena, which was heavily biased and partisan towards Federer.
The Serb was able to insulate himself from all the drama and emotions and played at the top of his game under immense pressure. His clutch ability to perform when it mattered made all the difference.
This is not the first time such a statistical absurdity has been observed at Wimbledon. In the final match of 2009, Federer defeated Andy Roddick despite winning one game fewer than him (and getting broken a lot more times).
Similarly, in the Isner–Mahut match at the 2010 Wimbledon Championships, which holds the record for the longest match ever played in history, John Isner won 24 fewer points than Nicolas Mahut and still won the match.
The distinctiveness of the scoring pattern acts as a lottery for the underdogs. Tennis gave us one of the wildest underdog stories ever when Goran Ivanisevic, a wildcard entrant, went on to win the Wimbledon Championship in 2001. This is what makes the game more interesting, and why the players don't complain about it.