-
Novak Djokovic (1.728/11) is likely winner but too short
-
Carlos Alcaraz (5.04/1) can win his half, but doesn't represent value
-
Daniil Medvedev (34.033/1) can be trusted on the surface
Cognitive dissonance
I'm unsettled. Conflicted. Cognitively dissonated. Two conflicting beliefs have taken hold in my think-box, and I can't shift either one of them. Here's why...
Every once-in-a-while, I do some analysis on my beloved tennis ratings to see if they still work. As a statistical zealot, I know that simply tracking profit-and-loss is not enough. Often, in the short-term at least, poor stats can still bring profit. Even a blind squirrel finds the occasional nut. Something more thorough is needed.
How good are my stats?
This week, then, I've taken the last 13 years of results in ATP events from my database and compared them to what my Elo ratings predicted would happen. To simplify the process, I did this by grouping predicted results into bandings - all those where the ratings predicted a roughly 1% chance of victory, then 2%, 3%, and so on. I then charted the results on to a scatter diagram - predicted results versus actual results.
If my ratings had been perfect over that period, the lines-of-best-fit would have drawn themselves on top of each other. The bigger the distance between the lines, the worse my ratings would have been shown to be.
In this analysis, the Elo rating performs well. It slightly undervalues favourites and overvalues outsiders. I've known that for a while, haven't been able to figure out why, so make some crude adjustments when deciding on bets.
So, belief number one lodged in my noggin: my ratings work.
The surface problem
Belief number two, unfortunately, always shouts louder at this time of year, as the tennis season transitions to grass: playing surface matters.
This is something I feel I know, instinctively.
Perhaps it's because, ever since I was kid watching Boris Becker win Wimbledon, the world has told me, again and again, that certain playing styles suit certain surfaces better, and that, with temperatures rising and grass courts being mown, this is the time of year for the greenster specialists.
The problem is, when I do the same analysis described above, but filtered by playing surface, those lines-of-best-fit look like two drunks walking across a bouncy castle.
The analysis leads to an unequivocal conclusion: playing-surface doesn't seem to matter nearly as much as we may think.
Wimbledon 2023
Which brings us to this year's Wimbledon.
I am desperate to tell you that the likes of Matteo Berrettini (160.0159/1), Andy Murray (80.079/1), Alexander Bublik (100.099/1) and Tallon Griekspoor (400.0399/1) are worth an outside bet. Honestly, desperate. Griekspoor's draw, I'm telling myself, is fortuitous. He could go further than expected. Then, who knows?
And this is how long-held beliefs work. Even when you know, for certain, that they are hogwash, they prove impossible to shift.
So that's not what I am going to tell you.
Novak Djokovic (1.728/11) is the most likely winner. Those odds are far shorter than my ratings say they should be (2.77/4), but then as we've discussed before, his scheduling consistently sees him outperforming his raw rating when playing the grand slams.
Carlos Alcaraz (5.04/1), in a different half of the draw to Djokovic, is the most likely runner-up. His odds are about right (I have him at 5.59/2). Supporters will be buoyed by his performance on the grass at Queen's Club (there I go again), but he doesn't represent any value.
Daniil Medvedev (34.033/1), however, does. He's the antithesis of the archetypal grass-court maestro, never having reached the quarter finals of Wimbledon. But, as we've seen, it's time to reject that kind of folk wisdom and trust the data that works.
Medvedev is the third-best player in the world, the draw has been kind enough, and he's a four-time grand-slam finalist and previous US Open winner. He's the value bet here.