With the Australian Open soon approaching, Dan Weston returns to analyse the top 20 players in the rankings and gives his thoughts on who is over and under-rated...
"Certainly, the decline of Kerber - second favourite for the Australian Open, no less - should be of huge concern to those who feel she represents a good bet for the upcoming tournament."
Attempting to use metrics to find over/under-rated players
As has been the case for the last few years, the WTA Tour has been extremely open and difficult to predict, with numerous players all capable of beating each other on any given day.
Having said this, it doesn't mean that difficult to predict equates to being impossible to predict and I thought I'd discuss some useful metrics with which to look at players, to try and identify those players in the WTA top 20 who are under and over-rated and whose results have flattered them, or conversely, whose results are worse than their actual ability.
An example of how to apply the data
An example application in the betting market is as follows. Let's say that we can identify a player who has won 70% of their matches over a short sample - say six months - but only deserved to win 50% of them based on data. Perhaps they dramatically over-performed on key points and won more tiebreaks than expectation. In this case we could have the potential for some continued opposition of that player in the relative short-term, because my detailed previous analysis has shown that it is extremely difficult to maintain such a 'clutch' ability over a long period of time.
Using service/return points won to assess player abilities
To start with, though, I want to look at player ability levels. One way we can ascertain who has performed the best, with as little bias as possible, is to assess service and return points won percentages. Adding them together gives us a concrete assessment of each player's abilities, with the following brackets decent guidelines to a player's overall ability:-
110+: Elite level - examples would be Novak Djokovic, Rafa Nadal on clay, Serena Williams at peak.
105-110: Strong upper top 10 level.
102-105: Not top level, but solid top 20 player.
100-102: Should be ranked in the 20-40 bracket.
Below 100: Unlikely to be ranked inside the top 40.
No current WTA player fits into the 110+ category, with Simona Halep (108.5%) having the highest 12-month all-surface figure. Serena Williams (109.2%) in the last six months, wasn't far from this level, and if she can continue in this vein, perhaps her favourite status for the Australian Open is justified.
Garcia among those who look flattered by their ranking
Using this data, we can then assess which players are perhaps falsely ranked, either positively or negatively, at this current time. Players who are flattered by their ranking include Sloane Stephens, Karolina Pliskova, Daria Kasatkina and Caroline Garcia - this quartet all have combined service/return points won percentages below the level a player of their ranking should have.
Conversely, Anastasija Sevastova (104.2% combined and 105.7% in the last six months) but ranked 13th, should have improvement potential, as should Ash Barty (ranked 15th, but 106.3%), Serena Williams (16th, 107.3%), and Madison Keys (17th, 106.0%).
Bertens and Sabalenka on an upward curve
Understanding which players have improved or declined is also useful, to assess which players are on an upward or downward curve, so what I did here was to assess a players six-month combined percentage with their 12-month combined percentage. The following players have shown big levels of improvement in the last six months:-
Kiki Bertens (2.6% improvement from 104.9% to 107.5%)
Aryna Sabalenka (2.6% improvement from 103.3% to 105.9%)
Serena Williams (1.9% improvement from 107.3% to 109.2%)
Madison Keys (1.6% improvement from 106.0% to 107.6%).
Both Bertens and Sabalenka are fancied contenders for the Australian Open - indeed, I have picked out Bertens as an outright selection - and their improvement is highlighted by these numbers. Williams' improvement has already been discussed, while Keys' numbers are even more impressive, given her injury issues at the tail end of 2018. If she can get back to fitness, she will likely be a huge threat on the WTA Tour this season.
Kerber decline should be of great concern
Players who have declined in the last six months include the following:-
Angelique Kerber (5.7% decline from 106.2% to 100.5)
Simona Halep (5.1% decline from 108.5% to 103.4%)
Petra Kvitova (3.7% decline from 104.8% to 101.1%)
Caroline Wozniacki (2.6% decline from 106.6% to 104.0%)
Elise Mertens (2.6% decline from 103.5% to 100.9%)
Elina Svitolina (2.3% decline from 105.9% to 103.6%)
Garbine Muguruza (2.0% decline from 105.0% to 103.0%)
Some of these players - Halep and Wozniacki in particular - have suffered with fitness issues, so their inclusion on the list is perhaps more justified than others, but certainly, the decline of Kerber - second favourite for the Australian Open, no less - should be of huge concern to those who feel she represents a good bet for the upcoming tournament. Petra Kvitova, despite a strong week in Sydney this week, also has shown a concerning drop in level.
Ability to be 'clutch' over-rated as a concept
Being able to interpret break point over/under performance is also extremely useful. As I mentioned previously, it is very difficult to be 'clutch' over a long-term basis, and such data enables us to see which players are likely to mean-revert (either positively or negatively) in the not too distant future.
We can do this by comparing a player's break point save percentage to their service points won percentage (on serve), and their break point conversion percentage to their return points won percentage (on return). The average WTA player, over the long-run, will save around 2.3% fewer break points on serve than service points won, and convert 2.3% more on return.
Players who have over-performed on break points in the last 12 months include the following:-
Caroline Garcia (+5.6%)
Naomi Osaka (+5.5%)
Anett Kontaveit (+4.9%)
Julia Goerges (+2.2%)
My previous assertion that Garcia is flattered by her ranking - and probable market status - is again backed-up by her over-performance on key points here as well, and US Open champion Osaka also looks rather flattered by her key point over-performance as well. The players above are likely to be slightly over-rated by the market, at the time of writing, and it is difficult to anticipate them being great value in the near future.
On the flip side, the following players have under-performed on key points significantly in the same time period, including:-
Madison Keys (-9.1%)
Ash Barty (-5.2%)
Simona Halep (-5.1%)
Elina Svitolina (-3.3%)
Serena Williams (-2.4%)
It is quite possible that this quintet of players are a little under-rated by the market. It could well explain Svitolina's drop-off in data, and certainly would explain Halep's - although injury has contributed to that as well.
Keys looks one of the best players on tour, when her overall service/return points won data is looked at in conjunction with a likely mean-reversion from underperformance on key points, and backs up my assertion that when fit, she has considerable upside. Ash Barty, to a slightly lesser extent, also looks very capable of further improvement.
Tiebreak records also worth evaluating
A final metric worth touching on quickly is tiebreak records. Some people perceive tiebreaks to be a lottery, but the truth is that long-term tiebreak success is based largely on serve/return points won expectations. The better players, even in such a short format, will win the most tiebreaks.
Summarising, using all the discussed data I can bracket WTA top 20 players as likely to be over-rated, or under-rated, and they are listed as follows:-
Likely to be over-rated (unlikely to represent considerable value in the near future):-
Naomi Osaka - solid but unspectacular combined points won percentage for a player of her ranking, huge over performance on break points.
Caroline Garcia - mediocre combined points won percentage, huge overperformance on break points.
Anett Kontaveit - same as Garcia, but won 11/16 tiebreaks as well.
Angelique Kerber - big drop-off in the last six months, and won a lot of tight sets via tiebreaks.
Julia Goerges - over-performed on key points, and won 16 of 23 tiebreaks in the last 12 months, a record which is unlikely to continue.
Likely to be under-rated (could represent potential for value in the near future):-
Ash Barty - has top five all-surface data, despite underperforming on key points.
Elina Svitolina - one of our Australian Open picks, she has underperformed on key points and only won 30% of her tiebreaks in the last 12 months
Madison Keys - when fit should make a real run to the top of the rankings, for reasons already discussed.
Hopefully this article has given you some food for thought in advance of the Australian Open, and in the near future, I'll be discussing similar angles for the ATP Tour as well.
Follow Dan on Twitter @TennisRatings