Wednesday, 12 July 2017

2017 Wimbledon Semi Finals Update

We're now in to the semi finals of Wimbledon so let's take a look at how the quarter finals have impacted on the rankings of the players still in.

The current leaders in the Big Gainer of the tournament awards are Sam Querrey (+55) and Magdalena Rybarikova (+57). Sam will win the award if he wins his semi final, or if Roger Federer wins the title, otherwise the award goes to whoever does better out of Tomas Berdych and Marin Cilic. In the Women's race, Magdalena Rybarikova is the favourite and will win it if she makes it to the final, or if Johanna Konta wins the title, otherwise it goes to the tournament winner.

But before the rankings, some stats I calculated for the top 150 players:


Men's Rank± ELO
1-10-24
11-2067
21-5033
51-1005
101-150-34


Women's Rank± ELO
1-10100
11-20-4
21-501
51-100-66
101-15010

Now we're at the semi final stage, I'm also going to add some prediction percentages for the final and the title, based on these ratings.

Men's rankings

#NameELO
1Roger Federer1942
2Marin Cilic1785
3Tomas Berdych1764
4Sam Querrey1660

Men's predictions

#NameELOFinalTitle
1
Sam Querrey
166032.75%7.04%
2
Marin Cilic
178567.25%23.68%
3
Roger Federer
194273.59%55.35%
4
Tomas Berdych
176426.41%13.93%

Women's rankings


#NameELO
1Johanna Konta1771
2Venus Williams1738
3Garbine Muguruza1722
4Magdalena Rybarikova1577

Women's predictions


#NameELOFinalTitle
1
Garbine Muguruza
172269.73%31.47%
2
Magdalena Rybarikova
157730.27%7.97%
3
Venus Williams
173845.27%26.32%
4
Johanna Konta
177154.73%34.24%

2 comments:

  1. I understand the prediction percentage for a single match, but what two numbers are being calculated to predict the outcome of the following match, in this case R2?

    For example, if Muguruza wins her semi-final, is her Elo updated for the prediction of the final? And is her Elo then being compared to the most likely winner of Konta and Williams, or to an average of their Elo, or something else?

    ReplyDelete
    Replies
    1. For the predictions, I assume that Elo stays the same (It doesn't, but the difference is tiny). And then for the final it's compared to both possible opponents, multiplied by how likely that player is going to be in that match.

      Delete