- Joined
- Apr 14, 2013
- Messages
- 10,117
- Reactions
- 5,767
- Points
- 113
INTRODUCTION: The Changing Shape of Careers
One thing that I find fascinating is how different players have different "shapes" to their careers. Some peak early, some late; some decline early, some late. Some have anomalous great years, then nothing much after. Some are steady and consistent in their performance, others up and down.
These patterns are most likely due to "real world" things that don't show up in the stats. Maybe a player goes through a bad breakup and loses focus for half a year. Or maybe they have a nice combination of luck and health. All we can really do, without being on the "inside circle" of players--not to mention the limitations of hindsight for past players--is look at the results: how their performance changes over time.
The focus of this thread is just that: the career trajectories of different players, and how they compare. But to do that, I need to share my background work - the "source material," if you will.
PART ONE: Tournament Mega-Chart (and some stat nerdery)
I've been working on a mega-project for the past few months, basically a huge chart that includes every tournament result for as many players as I get around to chronicling. As of this writing, I've done 146, which includes all Slam winners and year-end top 5 players; most 10+ title winners, big title winners and top 10 players; also, a bunch of active players that might end up in one of those categories. I'm not sure where (or if) I'll draw the line, but I've been tentatively planning on including every player who has done at least one of the following: top 5, year-end top 10, big title, 10+ titles, Slam final. But we shall see.
Anyhow, this is a sample of the chart to see how it looks:
As you can see, it is challenging to translate that to a screen for readability. But that is how 2022 and 2023 look; I've got the whole thing, going back to 1968.
Notice the separating columns: those include ATP ranking as well as four statistics of my own design. If you just want to view the results, you can ignore these - though the ATP rank is useful (I've thought of doing an alternate version with only ATP ranking, just so it is "cleaner" - but I like the nerdy stuff). I won't bore you with the formulas, but each measures something different:
PEP or Premier Event Points (pink): This is a slightly newer version of what I've shared in the past, that gives points for "premier" results at tournaments. What is different is that I've added fractions for lesser results (so maybe it should be considered just Event Points), because I found that it separated the elite too much from the pack and didn't differentiate between lesser players (e.g. whether a player reaches a Masters QF or 2R). It basically is an accumulative stat that roughly correlates with ATP ranking points.
P% or PEP Percentage (blue): The above stat but as a percentage of total possible points. For example, if a player has 20 PEP out of 90 possible, he has a 22 P%. If PEP is number of hits in baseball, P% is batting average. It is the quality of performance.
TS or Title Shares (green): This is similar to PEP, but gives points only for titles won and is more heavily weighted towards Slams and, to a lesser extent, Tour Finals. I call it "Title Shares" because it is based on 100 possible shares in a given season from big titles, plus extra points for lower titles. Meaning, in any given year, there are 100 TS for big titles, so a player's TS roughly equates with what percentage of the total he has earned.
SD or Season Dominance (purple): This adds up the three numbers into one, which gives a rough sense of how dominant a player was in a given year.
Note: For both TS and PEP, I use smaller numbers for the Australian Open in 1968-82, as it wasn't really close to equivalent to the other Slams. It become closer in 1983, but wasn't really coeval until the late 80s.
For PEP and TS, 10 is a very good season and roughly equivalent to a top 10 player; 20 is elite (top 5), 30 is great, 40 is historically great, and 50 is among the best of all time. For P%, it is more variable depending upon playing time, but a bit higher than PEP and TS.
PART TWO: Performance Shapes
I don't have a good term for this, but basically I started creating "shapes" for the stats above: PEP, P%, and TS, with the overall total being SD. This gives a visual translation of the numbers, so that you can get a sense of not only how good a player was based upon a visual form, but--to get back to the initial premise of this post--it depicts the trajectories of different players.
Given the announced retirement of Rafael Nadal, I thought I'd use him as an example:
The dark magenta is PEP, slight blue is P%, and green is TS - which are the relative size for the number of shares each title earns*. Bold means the player was the leader in that category. So for example, you can see that Rafa led all three categories in 2008, 2010, and 2013--easily his three best seasons--but led PEP in 2017, but not TS or P%, which were both led by Roger. Similarly in 2019, when he led two categories but Novak had more TS; and in 2022, he led two, but Novak had a higher P%.
You can clearly see how Rafa rose meteorically in 2005, plateaued for a couple years, then had his first peak year in 2008. He dropped again in 2009, then had another peak year in 2010. Another drop in 2011 (Novak's first peak year) and 2012 (injury) before what I consider his best year in 2013. He fell off substantially in 2014, had his weird decline in 2015-16, then resurged in 2018 to 2022. Actually, the shape is sort of similar to Andre Agassi's -- although quite a bit mightier than Andre's!
We know all of this, right, so what does this chart add? Well, you can compare seasons. For instance, you can see that while he only won two Slams in 2013, his quality of play was at his very best: he had a 71 P%, meaning he won 71% of PEP in tournaments he played in - an astonishing number. It also interestingly notes that his P% in 2019 was 60, slightly higher than 2008. Meaning, his overall quality of performance in 2019 was the third best season of his career, even if he took home fewer titles.
In later posts I'll share more players, with an eye on comparisons. For example, I'll compare Rafa to Novak and Roger, to give a sense of how each are different in terms of career arcs. One thing you'll note is how Rafa's "baseline level" of performance was extraordinarily high and consistent, due to his clay dominance. Roger's shape is a lot more "v-shaped" due to his high peak and then slow decline.
But before that, I'm going to share the shapes of some "second tier" guys, to give a sense of how different players born in the same year can be, in terms of career arcs. Stay tuned....
--------------------------
*NOTE: Here I'll share the formula for Title Shares, so it makes sense:
14 TS for Slams 1983-present (x4 = 56)
7-8 for Tour Finals, depending upon whether it is undefeated or not (x1 = 7-8)
4 for a Masters (x9 = 36)
56 + 8 + 36 = 100 TS. Olympics (5), ATP 500s (2 each), and ATP 250s (1 each) are added.
I had two goals in mind when I created this formula. One, I wanted the total for big titles to add up to 100; two, I wanted Slams to be about 50% of total shares or a bit more, but not too much more. I dabbled with the idea of an even 15 shares for the Slams for a total of 60 TS, but wasn't able to make the rest of the formula work as well.
One thing that I find fascinating is how different players have different "shapes" to their careers. Some peak early, some late; some decline early, some late. Some have anomalous great years, then nothing much after. Some are steady and consistent in their performance, others up and down.
These patterns are most likely due to "real world" things that don't show up in the stats. Maybe a player goes through a bad breakup and loses focus for half a year. Or maybe they have a nice combination of luck and health. All we can really do, without being on the "inside circle" of players--not to mention the limitations of hindsight for past players--is look at the results: how their performance changes over time.
The focus of this thread is just that: the career trajectories of different players, and how they compare. But to do that, I need to share my background work - the "source material," if you will.
PART ONE: Tournament Mega-Chart (and some stat nerdery)
I've been working on a mega-project for the past few months, basically a huge chart that includes every tournament result for as many players as I get around to chronicling. As of this writing, I've done 146, which includes all Slam winners and year-end top 5 players; most 10+ title winners, big title winners and top 10 players; also, a bunch of active players that might end up in one of those categories. I'm not sure where (or if) I'll draw the line, but I've been tentatively planning on including every player who has done at least one of the following: top 5, year-end top 10, big title, 10+ titles, Slam final. But we shall see.
Anyhow, this is a sample of the chart to see how it looks:
As you can see, it is challenging to translate that to a screen for readability. But that is how 2022 and 2023 look; I've got the whole thing, going back to 1968.
Notice the separating columns: those include ATP ranking as well as four statistics of my own design. If you just want to view the results, you can ignore these - though the ATP rank is useful (I've thought of doing an alternate version with only ATP ranking, just so it is "cleaner" - but I like the nerdy stuff). I won't bore you with the formulas, but each measures something different:
PEP or Premier Event Points (pink): This is a slightly newer version of what I've shared in the past, that gives points for "premier" results at tournaments. What is different is that I've added fractions for lesser results (so maybe it should be considered just Event Points), because I found that it separated the elite too much from the pack and didn't differentiate between lesser players (e.g. whether a player reaches a Masters QF or 2R). It basically is an accumulative stat that roughly correlates with ATP ranking points.
P% or PEP Percentage (blue): The above stat but as a percentage of total possible points. For example, if a player has 20 PEP out of 90 possible, he has a 22 P%. If PEP is number of hits in baseball, P% is batting average. It is the quality of performance.
TS or Title Shares (green): This is similar to PEP, but gives points only for titles won and is more heavily weighted towards Slams and, to a lesser extent, Tour Finals. I call it "Title Shares" because it is based on 100 possible shares in a given season from big titles, plus extra points for lower titles. Meaning, in any given year, there are 100 TS for big titles, so a player's TS roughly equates with what percentage of the total he has earned.
SD or Season Dominance (purple): This adds up the three numbers into one, which gives a rough sense of how dominant a player was in a given year.
Note: For both TS and PEP, I use smaller numbers for the Australian Open in 1968-82, as it wasn't really close to equivalent to the other Slams. It become closer in 1983, but wasn't really coeval until the late 80s.
For PEP and TS, 10 is a very good season and roughly equivalent to a top 10 player; 20 is elite (top 5), 30 is great, 40 is historically great, and 50 is among the best of all time. For P%, it is more variable depending upon playing time, but a bit higher than PEP and TS.
PART TWO: Performance Shapes
I don't have a good term for this, but basically I started creating "shapes" for the stats above: PEP, P%, and TS, with the overall total being SD. This gives a visual translation of the numbers, so that you can get a sense of not only how good a player was based upon a visual form, but--to get back to the initial premise of this post--it depicts the trajectories of different players.
Given the announced retirement of Rafael Nadal, I thought I'd use him as an example:
The dark magenta is PEP, slight blue is P%, and green is TS - which are the relative size for the number of shares each title earns*. Bold means the player was the leader in that category. So for example, you can see that Rafa led all three categories in 2008, 2010, and 2013--easily his three best seasons--but led PEP in 2017, but not TS or P%, which were both led by Roger. Similarly in 2019, when he led two categories but Novak had more TS; and in 2022, he led two, but Novak had a higher P%.
You can clearly see how Rafa rose meteorically in 2005, plateaued for a couple years, then had his first peak year in 2008. He dropped again in 2009, then had another peak year in 2010. Another drop in 2011 (Novak's first peak year) and 2012 (injury) before what I consider his best year in 2013. He fell off substantially in 2014, had his weird decline in 2015-16, then resurged in 2018 to 2022. Actually, the shape is sort of similar to Andre Agassi's -- although quite a bit mightier than Andre's!
We know all of this, right, so what does this chart add? Well, you can compare seasons. For instance, you can see that while he only won two Slams in 2013, his quality of play was at his very best: he had a 71 P%, meaning he won 71% of PEP in tournaments he played in - an astonishing number. It also interestingly notes that his P% in 2019 was 60, slightly higher than 2008. Meaning, his overall quality of performance in 2019 was the third best season of his career, even if he took home fewer titles.
In later posts I'll share more players, with an eye on comparisons. For example, I'll compare Rafa to Novak and Roger, to give a sense of how each are different in terms of career arcs. One thing you'll note is how Rafa's "baseline level" of performance was extraordinarily high and consistent, due to his clay dominance. Roger's shape is a lot more "v-shaped" due to his high peak and then slow decline.
But before that, I'm going to share the shapes of some "second tier" guys, to give a sense of how different players born in the same year can be, in terms of career arcs. Stay tuned....
--------------------------
*NOTE: Here I'll share the formula for Title Shares, so it makes sense:
14 TS for Slams 1983-present (x4 = 56)
7-8 for Tour Finals, depending upon whether it is undefeated or not (x1 = 7-8)
4 for a Masters (x9 = 36)
56 + 8 + 36 = 100 TS. Olympics (5), ATP 500s (2 each), and ATP 250s (1 each) are added.
I had two goals in mind when I created this formula. One, I wanted the total for big titles to add up to 100; two, I wanted Slams to be about 50% of total shares or a bit more, but not too much more. I dabbled with the idea of an even 15 shares for the Slams for a total of 60 TS, but wasn't able to make the rest of the formula work as well.
Last edited:
