Data crunch: Serena favourite for Wimbledon

Serena Williams has an 84 percent chance of winning the Wimbledon final against Angelique Kerber, according to innovative probability data developed by Tennis Australia. The data, sourced by Tennis Australia’s Game Insight Group (GIG), determines the statistical probability of a win based on historical performances.

According to that data, Australian Open champion Angelique Kerber has just a 16 percent chance of winning today’s final, three percent higher than her chances in the 2016 Australian Open final.

As an evolution of this data set, GIG has also developed a tool that tracks how a player’s win probability changes throughout a match. Predictions are based on a player’s ability against their opponent, and also on the difficulty of their position in the match.

A player’s ability is measured by a rating system called Elo, which is popular across many other sports.

GIG analyst Stephanie Kovalchik says that in tennis, Elo takes into account the results of a player’s tour matches, as well as the difficulty of their opponents.

“A player’s Elo rating always improves with a win, but it improves the most when they defeat a more highly-rated opponent. Players improve their rating by winning, and even moreso by winning against the best players.”

Kerber began her 2016 Australian Open campaign with an Elo rating of 2092, based on more than 400 career match results. By the time she’d reached the final, her rating had risen by 32 points to 2124 – still 330 points shy of Williams at 2454.

This differential gave the American an 87 percent chance of winning the match, compared to the German’s lowly 13 per cent.

Needless to say, as the Australian Open final showed, tennis isn’t played on a spreadsheet…

How did Kerber do it?

Kerber claiming the first set was crucial to narrowing the probability gap on her opponent. But even with one set on the board, Williams had a 60 percent chance of coming back and winning in three sets.

“Just like we saw in Roger Federer’s quarter final win over Marin Cilic this week, the best players are masters at dealing with setbacks,” Kovalchik said.

“Williams went into the AO final with a 739-123 career win-loss record and the expectation of winning 62 percent of her service points against Kerber.

“With those numbers, it was still more likely that Williams would come back to win in three than that Kerber would consolidate the win.”

A second set comeback by Williams saw her win probability increase to more than 75 per cent when she levelled the match at a set apiece.

Kovalchik says it wasn’t until the deciding set that the match truly became either player’s to win. Williams found herself in trouble early in the second game of the third set when Kerber opened up a triple breakpoint. Consequently, the American’s win probability dropped to 55 per cent – its lowest up to that point in the match. Williams lost the game, but rebounded by breaking back in the very next game.

The next critical turning point of the match happened in the sixth game – a 16-point tug-of-war in which Williams fought off four break points, before eventually losing the game on Kerber’s fifth breakpoint opportunity. At this stage, her chances of coming back after the mishap were just 35 percent.

Like the champion that she is, Williams broke back in the ninth game to restore parity. With the deciding set on serve, every point became critical. The dramatic swings in win probability show how significant each point, and how easily the outcome could have shifted if one or two points had gone differently.

What does this mean for the Wimbledon final? 

Since their last meeting in January, the gap in Williams’ and Kerber’s Elo ratings has narrowed. This translates into an 84-16 percent win probability in favour of the American.

RELATED: Serena v Kerber: What they need to do to win

If Williams takes the first set, her win probability will increase by nine per cent. But if Kerber goes up a set, Williams’ win probability will drop to 56 per cent, all but eliminating her pre-match advantage.

“Failing to get an early lead in the AO final was a big contributor to her loss there and could be an even bigger factor in the Wimbledon final,” Kovalchik said.

“If the match goes three sets, Williams is a 75 percent chance of prevailing.”

By the numbers, Williams is the overwhelming favourite to win the Wimbledon title under most scenarios. Of course, the statistics don’t account for the intangible factors such as confidence, determination or nerves. And there is no doubt Williams would be feeling the pressure to win a record 22^nd Grand Slam title.

“No other athlete has had to endure this kind of expectation,” Kovalchik said.

“Any statistical model would struggle to account for the emotional factors impacting a once-in-a-lifetime player.

“But because the model provides an accurate picture of what we should expect when players are playing to their natural ability, it can still be useful.

“When an outcome goes against what was statistically expected, it tells us that either the underdog got lucky or some other factors were involved.

“After both of their wins in the semis, you’d think they’d both be going in with a similar sense of confidence.”