After a walkover from Kohlschreiber in the R64, Stefanos Tsitsipas advanced to the third round with one fewer match played this week than any of his competitors. As Tsitsipas takes the court today for a spot in the 4th round, we look at whether skipping a round of play could have an impact on his result at the first slam of the season.
If you are like me, you are watching the Australian Open men’s draw with a keen eye on the rising stars of the tour. Daniil Medvedev and Stefanos Tsitsipas are two at the top of the list who have the potential to cause real trouble for the Big 3.
And if you have been tracking that story line, you had to wonder whether Tsitsipas being gifted a R64 victory by the withdraw of Philip Kohlschreiber was going to make any difference for his run at Melbourne Park. With an 81% expected win probability against Kohlschreiber, it wasn’t expected to be a difficult match. So, the odds would suggest that the walkover didn’t change Tsitsipas’s chances to get to the 3rd round but did get him there with at least a few hours less of match play. Will the reduced load be a help? Or could less time on court actually be a hindrance?
To try to answer these questions, it would help if we could identify other Grand Slam players who got a walkover early in the event and compare their results to similar players who had to play every round. Have early walkovers been frequent enough at Grand Slams to put together a reasonable sample?
The chart below shows how frequent walkovers have been at Majors in the past 20 years. We can see that these are generally rare, with most Majors having 0 walkovers in any round. However, about 20% (that is 1 in 5) Majors have had one or more walkovers in the R64 during this period. That gives us approximately 30 cases of a player who advanced due to a walkover in the first week of a Grand Slam.
As an aside, you may be wondering how any first round could have a walkover. Normally, a lucky loser should get a spot when a player withdraws in a first round. But in the case of the 2007 French Open first round, when Mikhail Youzhny got the walkover over Jan Hernych, the withdrawal must have happened too late to allow for another player to fill in. In general, first round walkovers are extremely rare.
If any effect of a walkover comes down to less match play, we could also consider expanding the sample by including retirements that happened early in a match. The combined event of walkovers and first-set retirements happen at least once in 1 of 4 Grand Slam first and second rounds. This effectively triples the number of players we can look at to get some idea about the possible effects of reduced play for a player’s results at a Major.
You may be wondering which events had 4 or more retirements in some years? These were the 2014 French Open, where 4 men retired in the first set of their first round matches. At the 2013 Wimbledon, it was a similar story for the R64, with two walkovers and two first set retirements. Fun fact! Kenny De Schepper benefited from those incomplete matches at both events.
Now that we have a recent sample of players who experienced a walkover or a walkover/first-set retirement at a Grand Slam, how can we say whether they benefited or not? One strategy is to say whether they did better overall than expected, which we can do by comparing their actual wins at the event against the sum of their predicted chance in each round. If a player had more wins than expected, this difference will be positive; while if they exited early, it will be negative. It’s basically a residual for event match wins.
Like residuals in a regression, we would expect the average of the residuals to be zero when taken over all of the players in a slam (since our predictions are certainly unbiased!). But this subset of players who get a walkover or have an opponent retire early on them are probably not an average tour player, in fact the majority of them have a rating over 2000 at the start of the event. So we need to compare them to players of similarly high ratings.
Restricting the comparison to 2000+ rated players, here is what we find about the spread of actual versus predicted wins for players who got a walkover in the past two decades at a Grand Slam. Despite the small sample, we can see an average boost in the actual minus expected wins, which corresponds to +0.2 more matches won than expected with a walkover. Two players that did especially well in this group were Stan Wawrinka who had +2.4 more wins than expected in his 2014 title win at the Australian Open that included a 3rd round walkover from Vasek Pospisil; and Grigor Dimitrov who had +2.2 more wins than expected in reaching the semifinal of the 2019 US Open with a R64 walkover from Borna Coric.
The edge is much the same when we expand the group and include players who had incomplete matches due to their opponent retiring in the first set. In the larger sample the average gain is closer to +0.3 extra matches.
Expectations were high for Tsitsipas before the second round, and if history tells us anything, his walkover ought to make them even higher now.