Meles
Bionic Poster
Technical discussion follows the images. All of the hard court lists (All Time Greats, 2015, and 2016) are ordered by a custom dominance ratio formula (see discussion below images for more.) These stats are all together in one view and feature detail to a 10th of a decimal point which is critical and something the ATP site does not do (they don't even order the stats leaders correctly and use their crudely rounded numbers with some bizarre order; very sloppy work.)
In ordering these players difficulties arise when you try to use something like points won. This year Isner is a very strong 5th by points won and if you use a traditional dominance formula he moves up to third (see discussion below for details on the formulas). Isner is not the third best hard court player this year, so after much perusal of the three lists and their accomplishments the key ingredients for hard court success have been unearthed:
1. Not losing points on serve (1st or 2nd) is key, so the higher the serve points won the better and no advantage is accrued by winning these on first or 2nd serve points though 1st serve is best because the player can get hot on first serve and win tournaments (reverse also true.)
2. For return, first return is critical
3. 2nd return is a factor if the player is extremely bad or extremely good. A favorite match stat that is often a key to success is 2nd serve points won. Here the reverse, 2nd return points won really shows the quality of the player in the rallies and general level. Servebots generally don't do well on 2nd return.
For the All Time greats, Sampras is always an issue because of his performance outside of majors, so no model can fully capture Pete, but several are shown in the tables. These are career numbers so its little surprise that the consistent Murray pulls in at 3rd versus the inconsistent Agassi. We see how great Roddick was over his career, but his return game really held him back (Roddick won more points on serve than Sampras including a very high 2nd serve points won number.) The stats start in 1991 so Edberg and Becker's numbers don't reflect their earlier years. Feel free to lookup the hard court stats of your favorite players in their peak years and compare.
For 2015, the big story is Federer who had tremendous stats overall, and some of the best first serve and first return numbers of his career. As discussed above these are important ingredients to success and so its little surprise that Federer is close to Djokovic with a huge gulf with the rest of the players. Nadal had a fine Fall, but basically the gap is huge between the top two and the rest of the tour.
For 2016, sadly Fed does not have enough stats to make the detailed numbers, but his numbers through Australia on hard courts were superlative. The big stories are the improving players:
1. Djokovic - return numbers really high in 2016 which has compensated for a lesser first serve game, but his 2nd serve numbers even better in 2016.
2. Murray - the big mover in 2016 and his stats have been best recently so he's probably stronger than the numbers indicate. Murray's 2nd serve and return numbers are up dramatically and this may reflect his general play in rallies, but the 2nd serve number is much, much higher than Murray's career average.
3. Raonic - look out. Milosh has credible return numbers this year that are better than Roddick's career return numbers and of course Raonic has a serve game in line with Roddick's career average.
4. Delpo - yes! Like Murray he has serious momentum that the numbers understate a bit. Not surprisingly Delpo is worse on 2nd serve and return than his career average. The big, big surprise is his first return which is a whopping 3% above his career average (the slice is working.)
5. Monfils - having a career year, but the guy is surprisingly bad on 2nd return points.
6. Thiem - way, way up from 2015. Went from 49.9% of points won to 52.2% and most of this is from a giant leap in first return.
7. Nishikori - doing well this year. Kei has a relatively weak first return game.
8. Nadal - stats don't show Rafa doing much of anything, but he's solid accross the board with a surprisingly weak return game given his total dominance on clay.
The above are the all time great hard court players since 1991.
March 20, 2017: I've added games won for the ATGs above. Tiebreakers are not included in serve or return games won. The serve games won is merely the average of return and serve games. This adds a little inaccuracy for tiebreakers. Also good players often win the last game serving, so this average probably favors returners a bit.
This also includes @Gary Duane 's check,
(Game%-50)/(point%-50)= ~2 (should be around 1.8-2.2)
Most of the Olympics data fiasco has been completely compensated for in the 2016 numbers (ATP site is a disaster for the year until they fix this.) I've fixed most of the top players completely, but some of the lesser Olympians will have wacky inflated DR numbers because no return data was entered for the Olympics. Not counting DFs as 2nd serve attempts were also an issue so some may have inflated 2nd serve points won numbers. Look at the Dominance% number which is the most accurate. The return points won number for these players is what they had coming into Olympics. Delpo, Rafa, Ninja, Monfils, and Murray have been fully compensated.
The data for these tables is taken from the annual stats leaders on each surface which has the raw points involved allowing for an exact calculation of the percentage rather than the ATP's retarding rounding and misordering of all of their stats (the morons round and then order by the number of points, so on first serve Felicano Lopez is shown behind Sam Queerrey despite having won 0.8% more first serve points, 79.5% versus 78.7% for Querrey, so Feliciano is shown 3 spots lower than he should.)
In the all time greats table, some of the players and numbers are in bold. For the ATGs the ATP keeps the top 200 so if one of the players was particularly low an exact estimate was made which involved algebra and other methods including data from tennisabstract (which painfully does not keep 1st return and 2nd return data distinct, but only an exact return points won which can be undone using algebra to get the distinct numbers.) The break point numbers are shown with a .0% and in reality are just the crude numbers from the ATP player stats as that was the only number possible.
The 1st return weight DR is based on a normal dominance ratio used for tennis which is simply:
(percent points won on return) divided by (percent points lost on serve)
Dominance ratios always favor servers and you can see the values for normal dominance in the above tables.
The 1st return weighted DR used here to order the above tables is exactly the same for the denominator (percent points lost on serve).
The top number (numerator) is where all the changes are done for return. Its weighted completely to first return, but does the following adjustments to reflect 2nd return:
1. If greater than 51% (0.51), the margin (as a ratio) is added to DR. So Djokovic at 60% 2nd returns (0.60) gets a healthy 0.09
2. If greater than 54%, more is added in a similar manner. Djokovic picks up another 0.06.
3. If the 2nd return is below 49%, double the ratio is dropped. The might Isner is at 44.2% (0.442) which is 4.8% below 49% (0.490) so Isner gets .048 twice for a drop of 0. 096 or nearly 0.1. (Call this the Serve Bot factor)
I think for Servebots, this DR formula may be a tad inaccurate, but they are known for blowing up any DR formula, so if this one sometimes drops them too much so be it. In the all time hard court greats the only players affected are Fernando Gonzales and Tsonga.
@Gary Duane
In ordering these players difficulties arise when you try to use something like points won. This year Isner is a very strong 5th by points won and if you use a traditional dominance formula he moves up to third (see discussion below for details on the formulas). Isner is not the third best hard court player this year, so after much perusal of the three lists and their accomplishments the key ingredients for hard court success have been unearthed:
1. Not losing points on serve (1st or 2nd) is key, so the higher the serve points won the better and no advantage is accrued by winning these on first or 2nd serve points though 1st serve is best because the player can get hot on first serve and win tournaments (reverse also true.)
2. For return, first return is critical
3. 2nd return is a factor if the player is extremely bad or extremely good. A favorite match stat that is often a key to success is 2nd serve points won. Here the reverse, 2nd return points won really shows the quality of the player in the rallies and general level. Servebots generally don't do well on 2nd return.
For the All Time greats, Sampras is always an issue because of his performance outside of majors, so no model can fully capture Pete, but several are shown in the tables. These are career numbers so its little surprise that the consistent Murray pulls in at 3rd versus the inconsistent Agassi. We see how great Roddick was over his career, but his return game really held him back (Roddick won more points on serve than Sampras including a very high 2nd serve points won number.) The stats start in 1991 so Edberg and Becker's numbers don't reflect their earlier years. Feel free to lookup the hard court stats of your favorite players in their peak years and compare.
For 2015, the big story is Federer who had tremendous stats overall, and some of the best first serve and first return numbers of his career. As discussed above these are important ingredients to success and so its little surprise that Federer is close to Djokovic with a huge gulf with the rest of the players. Nadal had a fine Fall, but basically the gap is huge between the top two and the rest of the tour.
For 2016, sadly Fed does not have enough stats to make the detailed numbers, but his numbers through Australia on hard courts were superlative. The big stories are the improving players:
1. Djokovic - return numbers really high in 2016 which has compensated for a lesser first serve game, but his 2nd serve numbers even better in 2016.
2. Murray - the big mover in 2016 and his stats have been best recently so he's probably stronger than the numbers indicate. Murray's 2nd serve and return numbers are up dramatically and this may reflect his general play in rallies, but the 2nd serve number is much, much higher than Murray's career average.
3. Raonic - look out. Milosh has credible return numbers this year that are better than Roddick's career return numbers and of course Raonic has a serve game in line with Roddick's career average.
4. Delpo - yes! Like Murray he has serious momentum that the numbers understate a bit. Not surprisingly Delpo is worse on 2nd serve and return than his career average. The big, big surprise is his first return which is a whopping 3% above his career average (the slice is working.)
5. Monfils - having a career year, but the guy is surprisingly bad on 2nd return points.
6. Thiem - way, way up from 2015. Went from 49.9% of points won to 52.2% and most of this is from a giant leap in first return.
7. Nishikori - doing well this year. Kei has a relatively weak first return game.
8. Nadal - stats don't show Rafa doing much of anything, but he's solid accross the board with a surprisingly weak return game given his total dominance on clay.
The above are the all time great hard court players since 1991.
March 20, 2017: I've added games won for the ATGs above. Tiebreakers are not included in serve or return games won. The serve games won is merely the average of return and serve games. This adds a little inaccuracy for tiebreakers. Also good players often win the last game serving, so this average probably favors returners a bit.
This also includes @Gary Duane 's check,
(Game%-50)/(point%-50)= ~2 (should be around 1.8-2.2)
Most of the Olympics data fiasco has been completely compensated for in the 2016 numbers (ATP site is a disaster for the year until they fix this.) I've fixed most of the top players completely, but some of the lesser Olympians will have wacky inflated DR numbers because no return data was entered for the Olympics. Not counting DFs as 2nd serve attempts were also an issue so some may have inflated 2nd serve points won numbers. Look at the Dominance% number which is the most accurate. The return points won number for these players is what they had coming into Olympics. Delpo, Rafa, Ninja, Monfils, and Murray have been fully compensated.
The data for these tables is taken from the annual stats leaders on each surface which has the raw points involved allowing for an exact calculation of the percentage rather than the ATP's retarding rounding and misordering of all of their stats (the morons round and then order by the number of points, so on first serve Felicano Lopez is shown behind Sam Queerrey despite having won 0.8% more first serve points, 79.5% versus 78.7% for Querrey, so Feliciano is shown 3 spots lower than he should.)
In the all time greats table, some of the players and numbers are in bold. For the ATGs the ATP keeps the top 200 so if one of the players was particularly low an exact estimate was made which involved algebra and other methods including data from tennisabstract (which painfully does not keep 1st return and 2nd return data distinct, but only an exact return points won which can be undone using algebra to get the distinct numbers.) The break point numbers are shown with a .0% and in reality are just the crude numbers from the ATP player stats as that was the only number possible.
The 1st return weight DR is based on a normal dominance ratio used for tennis which is simply:
(percent points won on return) divided by (percent points lost on serve)
Dominance ratios always favor servers and you can see the values for normal dominance in the above tables.
The 1st return weighted DR used here to order the above tables is exactly the same for the denominator (percent points lost on serve).
The top number (numerator) is where all the changes are done for return. Its weighted completely to first return, but does the following adjustments to reflect 2nd return:
1. If greater than 51% (0.51), the margin (as a ratio) is added to DR. So Djokovic at 60% 2nd returns (0.60) gets a healthy 0.09
2. If greater than 54%, more is added in a similar manner. Djokovic picks up another 0.06.
3. If the 2nd return is below 49%, double the ratio is dropped. The might Isner is at 44.2% (0.442) which is 4.8% below 49% (0.490) so Isner gets .048 twice for a drop of 0. 096 or nearly 0.1. (Call this the Serve Bot factor)
I think for Servebots, this DR formula may be a tad inaccurate, but they are known for blowing up any DR formula, so if this one sometimes drops them too much so be it. In the all time hard court greats the only players affected are Fernando Gonzales and Tsonga.
@Gary Duane
Last edited:
In RIO and was a bit tired for the Murray final for which they were scheduled a day of rest before playing. With that first return Delpo is all kinds of dangerous right now.
