As discussed earlier this week, the Clutch Index has a compelling correlation with ATP ranking: both conceptually and empirically. The Clutch Index ranks players on their ability to win tight sets and tight matches (see Methodology at the bottom of this page): better players win tight matches. The Clutch Index for 2008-2013 featured all players to have won a Grand Slam in the period, as well as 7 of the world’s current Top 15. See the previous post.
And as for the period 2008-2013, so for 2013 alone. The chart below plots the top 50 ATP ranked players (after the US Open; x axis) against their Clutch Index for 2013 (y axis). Accordingly, by ranking, Djokovic is the red marker furthest left, Nadal next to him, then Murray and so on. The marker furthest right is Horacio Zeballos, ranked 50^{th}.
Clutch Index 2013 vs ATP Ranking (post-2013 US Open)
Again, the striking correlation between ATP Ranking and the Clutch Index: a high ranking is correlated with an ability to win tight matches.
However, other interesting analysis occurs when you reorder the same data above not by ATP Ranking but by Clutch ratio. First of all, here’s the chart reordered:
Clutch Index 2013 ordered by Clutch ratio
The table below is the shaded area above, which forms a natural top tier of the Clutch Index. The top tier comprises 13 players. Their data are set out below.
Clutch Index 2013; Top Tier
Clutch Index 2013 Rank |
Player |
Clutch ratio |
ATP Ranking (post US Open 2013) |
1 |
Tursunov |
0.89 |
32 |
2 |
Isner |
0.68 |
15 |
3 |
Youzhny |
0.63 |
20 |
4 |
Nadal |
0.59 |
2 |
5 |
Cilic |
0.58 |
24 |
6 |
Murray |
0.58 |
3 |
7 |
Del Potro |
0.56 |
7 |
8 |
Nieminen |
0.55 |
42 |
9 |
Djokovic |
0.52 |
1 |
10 |
Kohlschreiber |
0.48 |
25 |
11 |
Simon |
0.46 |
16 |
12 |
Ferrer |
0.43 |
4 |
13 |
Tsonga |
0.43 |
8 |
Dmitry Tursunov leads the Clutch Index for 2013 by a staggering 21 basis points. His score is a good third higher than Nadal’s six year score of 0.66, shown last week. Out of 22 “Clutch matches” at ATP tour level, Tursunov has won 6 matches 7-5 or 7-6 in the deciding set and lost zero. His ranking has, not surprisingly, headed north: from 127 at year end 2012 to 32 after the US Open 2013.
Marcos Baghdatis currently props up the Clutch Index for Top 50 players. It is unfortunate for Baghdatis that Tursunov and Baghdatis would play 5 days ago in St Petersburg after The Clutch Index 2013 above had been calculated. Tursunov won 7-6 7-6 in sublime support of The Clutch Index. Irresistible.
Impossible to prove and unlikely to be the full story, the Clutch Index may explain a rise or fall in confidence of players. Exhibit A: after 12 first round losses in 2013, Baghdatis’s confidence is unlikely to be high. A fall in confidence may be linked to an inability to close out tight sets.
However, changes from year to year in a player’s Clutch ratio do not consistently show a corresponding rise or fall in a player’s ATP ranking. Yes, Baghdatis’s ranking has fallen from 35 to 48, and Tursunov’s has risen. However, this is not consistently seen across the data. This is also in contrast to the long-term picture where players at the top of the Clutch Index are consistently drawn from the Top 15 in ATP rankings (6 of the top 8 – absent Berdych, Federer – are represented in the table above).
The ability to win tight matches accordingly may be inherent for the most part. And certainly at the top of the game, to paraphrase singer-songwriter Sammy Cahn, “You’ve either got or you haven’t got clutch”.
Clutch; Automatic Clutch; Transmission Problems
So what are the take aways from this research?
- As before, the ability to win tight matches in 2013 correlates strongly with higher ATP ranking (0.51 for those interested in correlation coefficients). Tennis needs to find a way to incorporate this into mainstream analysis: it is too signifcant to be ignored
- The Clutch Index provides correlation over the long-term of a player’s ranking, but changes in a player’s Clutch Index in the short-term are not consistently reflected in a corresponding change in a player’s ATP ranking.
Accordingly, other important correlations with a player’s ATP ranking – and in particular the need for different skills to break into different strata of the rankings – are the subject of next week’s Cleaning the Lines.
Annex 1 – Methodology for The Clutch Index
I have taken all ATP Tour matches and men’s Grand Slam matches played between 2008 and the US Open 2013 (all the matches have to have been completed). Something around 16,500 matches. Points have been awarded to players as follows:
- +1pt – win a set 7-6 or 7-5
- -1pt – lose a set 7-6 or 7-5
- +2pts – win the final set of a match (3 set or 5 set) 7-6, 7-5 or 8-6, 9-7 etc
- -2pts – lose the final set of a match (3 set or 5 set) 7-6, 7-5 or 8-6, 9-7 etc
Example: in R16 (round 4) of the US Open 2013, Gasquet beat Raonic 6-7 7-6 2-6 7-6 7-5. In this match, Gasquet scores 3 points (-1 for losing the first set TB, +1 for each of the second and fourth set TBs, and +2 for winning the final set 7-5). Raonic scores -3 points (+1 for the first set, -1 for each of sets 2 and 4, and -2 for losing the last set 7-5).
Terminology
- Clutch matches – a match involving at least one set with a scoreline of 7-6 (or 6-7) 7-5 (or 5-7), 8-6, 9-7 etc. Players for the table below were eligible if they had played in 50 such matches since 2008.
- Clutch score – the total of a player’s score totalling up his “negative” sets (eg sets lost 7-6 7-5) and his “positive” sets (sets won 7-6 or 7-5)
- Clutch ratio – the key measurement. This is the Clutch score divided by the number of Clutch matches. It is important not to over- or underweight Clutch scores based solely on the number of Clutch matches.