Header Ads

Laying the Attractive WTA Players

For many years, I have had a theory that the ‘attractive’ players in the WTA are overvalued in the betting markets. If this were true, then there might be a valid strategy in laying these players. However, without testing this hypothesis, it had its merits, but it would be impossible to justify putting real money on.

The first thing to do is to compile a list of the top 10 most attractive players. Obviously, there is a large subjective element in determining which players are deemed attractive – different people find different women attractive. To put together a list, I ran a poll on DW on Sport asking people to select the three players that they felt were the ‘most attractive on the WTA Tour’. There were three standout players – Ana Ivanovic, Maria Kirilenko and Eugenie Bouchard. The next group contained Alize Lim, Sorana Cirstea and Mandy Minella, all of whom polled in double figures. The rest of our ‘Top 10’ were made up of Daniela Hantuchova, Heidi El Tabakh, Julia Goerges and Dominika Cibulkova.
Alize Lim is a lesser-known player on the WTA, but should be be laying
her purely based on her looks?

I will test this theory by focussing on when the players are the favourite in the betting market. There is no particular reason for this, but the majority of money is usually for the favourite. The favourite-longshot bias would tend to suggest that backing the favourite is the better strategy, which is backed up by the statistics. If you had staked £10 on every player in the top 100 for the last 50 matches where they have been favourite, you would have made a total of £640 from stakes of £50,000 at 1.3%. While this is hardly a great return, it is profitable nonetheless.
Ana Ivanovic has long been one of the most photogenic players on the WTA Tour

The first thing to do is to set a benchmark to compare our ‘attractive’ players against. As we have a list of 10 players, we shall randomly take samples of 10 players from the top 100 in the world. We can record the amount that we would have won or lost from £10 stakes per match, or £5,000 per sample. Having taken 100 random samples, the mean amount that we would have won from each sample of 10 players comes out at £57.68 or 1.2%. The standard deviation comes out at £51.17.

Now, we look at the sample of the ten players that we have. If we had staked £10 on each of the last 50 matches that each of them had played as favourite, we would have lost a total of £153 at 5.0%. Clearly, this reflects pretty badly compared to the overall average, however, at neither the 95% nor the 99% level, a hypothesis test does not prove that this is not just random variation. The table below shows the values for each of our players:

Return As Favourite (Last 50 Matches)
Ana Ivanovic
Maria Kirilenko
Eugenie Bouchard
Sorana Cirstea
Julia Goerges
Dominika Cibulkova
Daniela Hantuchova
Mandy Minella
Alize Lim
Heidi El Tabakh

As we can see, there is plenty of variation among the players in the sample. Julia Goerges is comfortably the worst with a very poor -15.8% return as favourite, while Daniela Hantuchova, Sorana Cirstea and Alize Lim also show returns of below -8%. However, Eugenie Bouchard performs very well, returning a positive double figures, while Kirilenko and Ivanovic both show returns of over 4%.
Julia Goerges shows a particularly bad -15.8% return as a favourite

One question to look at is why our theory might be true. Casual punters will often be tempted to back the players that they are most familiar with. Generally, this is likely to be the top players – the likes of Serena Williams, Maria Sharapova, Victoria Azarenka are well-known to even casual tennis fans. Even the likes of Sam Stosur, Petra Kvitova, Li Na and Agnieszka Radwanska will be known to many.

However, the name recognition of players is usually related to their media profile and mainstream presence. Anna Kournikova, despite only just breaking into the top 10 in the WTA rankings, is arguably one of the most famous tennis players of all-time among the general public. This is not due to her tennis, but entirely due to her looks and the marketing and modelling presence that she had off the court. In the same way, the current attractive WTA players, albeit it to a far lesser extent than Kournikova, will often have name recognition that is significantly greater than it would be compared to another player of a similar standard, but without the looks.
Anna Kournikova was known more for her looks than her tennis

Conversely, we could argue that there are enough smart punters involved in the market that they will base their pricing purely on tennis-related statistics and reasoning. While casual punters might force the price out of line by supporting the attractive players, these smart punters will take the higher price and push it back into line.

Looking at our figures, it is not entirely clear whether our theory is correct. While the mean of our sample is low – negative in fact – but statistically, we cannot prove that it is not due to random variation. There are also plenty of other variables that affect the price – young players breaking through are often undervalued by the market, hence the impressive returns for Genie Bouchard, while some players simply struggle as favourites. Certain players may be overrated by the market for other reasons.
Genie Bouchard is receiving huge publicity both for
her tennis and her looks

The theory may well be true. The statistic is very close to the level where we would reject the null hypothesis of no difference in mean at the 95% level – a couple of matches won or lost either way might have nudged this over that level. However, we cannot say for certain that laying the attractive players is a guaranteed profitable strategy. We might be able to refine the strategy by taking more variables into account, but for now, it remains relatively inconclusive. Indeed, based on the hypothesis testing, there is a reasonable argument to be made that there is no link, which would back up the idea that the smart money outweighs the casual money in these markets. 

*The data used was correct as of 15th February when this study started

No comments:

Powered by Blogger.