8. ATP Hard Court Rankings 2013

Preamble

ATP rankings are based on a player’s overall performance across all surfaces (outdoor hard, indoor, clay, grass). A player’s ranking is made up of points awarded by the ATP Tour according to how many rounds a player advances at each tournament.

However, the same rankings and points are rarely used to analyse a player’s performance on just one surface. I have even more rarely seen a table that tries to rank the population of ATP professionals by just one court surface.

This is illogical: most players perform differently on different surfaces. Think only of Pete Sampras: 14 grand slams, 7 on grass, none on clay.

So, with the Masters 1000 tournament in Shanghai having finished, I have compiled data that ranks ATP players based on their performance in 2013 on just one surface: an outdoor hard court. I will do the same for clay, indoor hard, and grass court tournaments in due course.

Tournaments and matches in scope

  • The tournaments in scope for this research are only those at ATP tour level tournaments on outdoor hard courts. Those tournaments include the US and Australian Open grand slams, Masters 1000 tournaments in Indian Wells, Miami, Montreal, Cincinnati and Shanghai, as well as smaller ATP 500 and ATP 250 tournaments on outdoor hard courts (eg Dubai, Brisbane).
  • Tournament and matches not in scope are: Davis Cup matches, second and third tier tournaments at Challenger and Future level, and matches that do not count towards a player’s ranking such as the Hopman Cup or the Kooyong exhibition. A quick note about the Davis Cup: few points are awarded for Davis Cup matches and these are awarded only for matches played at World Group or World Group Play Off level. Accordingly, despite many players playing Davis Cup (especially at lower levels) it is not fair to include them for some players and not others.

Scoring system

The metric that I rank the players by is called ATP Points per Match. This is how I get there.

  • The scoring system I have used to rank players is the ATP ranking points system. This is useful as it means that the number of matches a player wins or loses is not as important as the size or type of matches he wins. For example, a player who wins two small tournaments might win 10 matches and score 500 points in total. A player who wins a grand slam tournament wins only 7 matches but scores 2,000 points in the process. Given the size, importance, and, if nothing else, mandatory attendance of players at a grand slam tournament, a grand slam is clearly worth more points. We remember players – and consider them “great” – by the number of big tournaments they win.
  • This system also rewards players who win tournaments or reach the latter stages of tournaments. The ATP Tour awards points at tournaments according to a geometric progression not arithmetic progression. At its simplest, a player might win 90 points for winning three matches at a tournament, 150 points for winning four, and 250 points for winning five.
  • The vast majority of ATP points awarded are for players in the “main draw” of an event. However, points are also awarded for players coming through the “qualifying draw” at an ATP tour level tournament. I have included these points from qualifying in my table but not the number of wins and losses from qualifying matches: this would distort their win-loss record compared to those who only play at main draw level. To explain with an example: a player who qualifies for a Grand Slam event but gets knocked out in the first round has won three matches in qualifying and lost their first round main draw match. The win-loss record in my data will show only that the player has lost the first round match. However the 35 ATP points scored (25 points for qualifying and 10 points for being knocked out in the first round) are included.
  • I have taken the total ATP points earned by each player in 2013 on an outdoor hard court and divided it by the number of outdoor hard court matches that player played in 2013. This gives me an “ATP Points per match” score (which I will now refer to as the “ATP PPM”). I have ranked the players by their ATP PPM score to arrive at the hard court ranking for 2013.
  • The ATP PPM score is important. It is possible for players who win matches at several small tournaments to earn more points than players who win fewer matches but win those matches at bigger, more important tournaments.  In this case, a ranking based solely on the number of ATP points won is insufficient. If we divide the number of ATP points won by the number of matches played, we arrive at a fairer ranking (ATP PPM).

Notes to data in the table

There are 9 columns in the table below. Some are self-explanatory, others less so. But to avoid doubt, here is an explanation of the 9 columns:

  1. Hard Court Rank 2013. My ranking of ATP players on a hard court based on ATP PPM (column 9)
  2. ATP Rank as at 14 October 2013. Taken from the ATP website. Shown for context.
  3. Player
  4. Played. Number of outdoor hard court matches played in 2013.
  5. Won. Number of matches in column #4 that were won.
  6. Lost.
  7. Win %. Column #5 divided by column #4.
  8. ATP Points. Number of ATP points won on outdoor hard court matches in 2013.
  9. ATP Points per match. Column #8 divided by column #4. The table ranks players by their ATP PPM score in column #9.

Finally, all players had to have played a minimum of 6 main draw matches on an outdoor hard court in 2013 to be included in the table. Click the table for a higher resolution version.

Outdoor hard court ranking 2013

8 Hard court ranking table 2013

Preliminary analysis – the emergence of a new Big Four

There is plenty of analysis that can be done off the back of this data: this will be the subject of subsequent posts. This data will surely raise questions for debate. All of that is to be welcomed, and please feel free to comment on this page.

What I wanted to suggest today though, is that the data points to the emergence of a new Big Four on a hard court. Whereas previously we knew the Big Four to be Nadal, Djokovic, Murray and Federer, so the data above justifies talk of a new Big Four that includes Juan Martin Del Potro instead of Roger Federer.

As the table below shows, of the 11 biggest ATP tournaments on an outdoor hard court in 2013, all were won by one of the “new” Big Four, and 5 of the 11 finals contained two members of the “new” Big Four.

Outdoor hard court tournament winners and runners up

8 Hard court Winner Runner Up 2013

The chart below plots the win percentage on an outdoor hard court of each player in the main table above against their ATP PPM. In plain English, it plots a traditional measure of performance (win percentage) against a measure that rewards players for progressing to the latter stages of tournaments, particularly big tournaments (ATP PPM).  You can see how points are awarded geometrically – the gaps between the players get bigger towards the top end of the profile.

Outdoor hard court 2013: ATP PPM vs Win percentage

8 ATP PPM vs Win% with CtL

Accordingly, the chart also suggests some tiers of players: the new Big Four, a possible tier 2 of hard court players, a possible tier 3 of players; and players who are nearest to joining tier 3 – as set out in the chart below.

8 ATP PPM vs Win% with text with CtL

Next on Cleaning the Lines

In-depth analysis will be available in subsequent posts. For now it should be sufficient to mull over the data. In upcoming posts, I will focus on players who aren’t in the top 10 as well as providing more granular detail  on what it means to have high ATP PPM: as well as “good” performance, how much does it say about a player’s consistency?

4 thoughts on “8. ATP Hard Court Rankings 2013

  1. Thanks! I keep forgetting that the match that was to all intents and purposes the final was played on the Saturday night not the Sunday…. No effect on the numbers, though. That table was just my addled brain.

  2. Interesting analysis. A (smooth) regression curve drawn through the data in the plot of ATPPPM vs “% outdoor hard court matches won” (POHCMW) would seem to be an exponential rise. This follows from the geometric reward component of ATPPPM. What is the shape of the probability distribution of POHCMW? That is, the number of players whose POHCMW lay within each 5% bin centred on the POHCMW’s from 2.5% up to 97.5% in steps of 5%. (there are other ways of defining the bin centres and widths, but this way would be reasonable for your data). In theory the distribution should be a Gaussian curve, also known as the “normal distribution”, also known as the “bell” curve. The mean and standard deviation are the descriptors of a given Gaussian curve.

    If you created a large number ( a few hundred say) of “virtual” players and gave them a Gaussian distribution of POHCMW then you would obtain a set of data for ATPPPM and POHCMW that would fall on or close to a smooth curve, which should show an exponential rise. Even though the players are virtual, the curve would be very similar to the curve you obtained for real players.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s