A Tennis Frontier All-Time Greatest Players List - Help me!

[El Dude]

I have an idea - let's come up with our own, shared, All-Time Greatest Players list. Rather than debate other different lists, I think the best way to proceed - and for everyone to be happy about the results - is by coming up with a system. If we can agree on the system, even if in a general way, then we can agree upon the results that the system leads to? Right?

First of all, the system shouldn't be too complex. I mean, we could get really, really detailed, but I for one don't have the time to run every player's stats through a complex formula. Basically we'd need something that we could figure out by looking at Wikipedia and atpworldtour.com.

To get things started, think by far the most important criteria is Grand Slam results, and everything else is secondary. I also think anything from a QF or higher should be considered - basically second week results. A QF is sort of the baseline for a great player. As for other tournaments, its tempting to only count victories - I mean, who cares or remembers if you made it to the SF or even the Final at an ATP 500 or 1000?

Also, if you're going to go this route I think you have to look at ranking separate. Rankings and event results are two separate systems.

Maybe a preliminary system would look like this:

Grand Slam Results:
10 Win
5 Runner-up
2 SF
1 QF

(Pro Slams and pre-Open Era Grand Slams would receive half the above, or 5, 3, 1, 0)

Titles
5 WTF
3 Olympics
4 ATP 1000 (or its equivalent)
2 ATP 500 (or its equivalent)
1 ATP 250 (or its equivalent)

Or something like that. What do you think? If people can help me craft this I'd be happy to run the numbers and come up with lists.

A couple problems that come to mind are:
1) By not counting non-win results in other events, you end up with an ATP 250 being worth more than an ATP 1000 runner-up, when the latter is obviously harder to accomplish. So it may be that, at the least, appearances in Finals in all events other than ATP 250 need to be counted.
2) This system counts all Slams from all eras equally (except pre-Open Era). We all know that the Australian Open in 1973 wasn't the same as it is in 2013. I personally would prefer to just let this one go, otherwise we have to figure out when the AO become as important as the rest of the Slams, and really we'd probably have to count Wimbledon a bit higher than the rest. I'd rather say, "it all evens out."
3) This doesn't account for greatness on all surfaces. It might be that there could be a list of "bonus points" - like winning all four slams gives a player an extra 10 points, or even adding in ranking somehow.

OK, weigh in folks.

 

[tented]

Grand Slam Results:
10 Win
5 Runner-up
2 SF
1 QF

(Pro Slams and pre-Open Era Grand Slams would receive half the above, or 5, 3, 1, 0)

This seems like a justification for eliminating the possibility of any pre-Open Era player becoming the GOAT, right? Which means Roger is GOAT, because obviously his numbers would put him higher than anyone else using your system.

I know you mean well, but this reminds me of Britbox's comment on another GOAT thread:

"The usual policy tends to be

a) Decide who you like best
b) Pick the criteria to make a) come out on top."

And since you're a Federer fan ...

 

[El Dude]

Come on, tented, give me a break. Really? You think I'm that pathetic?

Maybe the half value should only be given for 1963 through the AO of 1968, the time when pros were banned from the Grand Slams. Roy Emerson won 9 of his 11 Slams when Laver and Rosewall weren't even allowed to play in Grand Slams. There were great players in both sides, but the talent was diluted.

 

[El Dude]

By the way, I think using the above system Ken Rosewall ends up being #1. So much for your "theory."

 

[britbox]

I think it could be interesting. I think pre and open eras should be included and there should be various formulas used for slam values in the pre-open era and the AO in certain time periods. We'd also have to figure comparable tournaments to the Masters Shields etc... I'm in El Dude - and will help contribute... give us a week or so though when I have some time.

As Tented said though, these things don't mean a whole lot definitively, but it could be fun to and see where the chips fall.

 

[tented]

Maybe the half value should only be given for 1963 through the AO of 1968, the time when pros were banned from the Grand Slams. Roy Emerson won 9 of his 11 Slams when Laver and Rosewall weren't even allowed to play in Grand Slams. There were great players in both sides, but the talent was diluted.

That would certainly change things a lot, but the "weak competition" aspects, so to speak, would be difficult to resolve, if I understand you correctly. On one hand, I think you've improved the system a lot with this change, and on the other, I think it makes it less objective.

By the way, I think using the above system Ken Rosewall ends up being #1. So much for your "theory."

If you mean in terms of doing the math based on the changes in values outlined above, then, yes, I think Rosewall might become GOAT. Or at least GOAT-ish. Roger's Open Era quarters, semis, finals, and wins would be heavily weighted, though, not to mention all of his Masters, WTFs, 500s, and 250s.

And since Federer is still an active player, whereas Rosewall is not, can we (or should we) compare them?

Also, I'm not convinced throwing out results at non-GS tournaments unless it's a win is the right thing to do. You state "I mean, who cares or remembers if you made it to the SF or even the Final at an ATP 500 or 1000?" The "who cares" immediately introduces a subjective element, but it's the use of "remembers" which is particularly troublesome, because memory is subjective in a different sense.

Some people are going to remember results others won't. I actually do remember quite a few results for non-winners of 500s and 1000s, but I couldn't tell you the same info. for tournaments in the 1950s. I'm sure there are people out there who would be in the reverse situation. Who's right?

 

[tented]

One other thought --

Do you plan on using this system only for the men? If the system is to be truly objective, then shouldn't it also be transferable to the women? Does this change things?

 

[El Dude]

The key, I think, is to find a balance between usable (and thus relatively simple) and as comprehensive as possible. Another problem with looking at non-Slam tournament results other than wins is just finding the data. As it is, there are some players that I can't find more than Slam finals for. And once we get into the pre-Open Era I'm not even sure the data exists. Tennis isn't baseball, where stats have been meticulously kept for over a 100 years.

Anyone know of a good tennis statistics site? I use Wikipedia and atpworldtour.com, but I'd love to find the tennis version of baseball-reference.com. I don't think it exists, but I would think something is out there...

And yes tented, I'd like to be able to apply it to women's tennis.

But as a starter, how would people weigh the importance of Slam results? I'd like to do two things: One, simmer numbers down to as small as possible; 2) work only with whole numbers.

Ultimately there's no way around the subjective element in terms of valuation of different results and different tournaments.

But let me give it a shot, with my process explained. We could look at the ATP point values for each as a ratio to work with, and then find an internal ratio to apply to each. So we get:

Grand Slam: 2000
WTF: 1500
ATP 1000: 1000
ATP 500: 500
ATP 250: 250
Olympics: 750

Then divide those by the lowest common denominator, which is 250:

Grand Slam: 8
WTF: 6
ATP 1000: 4
ATP 500: 2
ATP 250: 1
Olympics: 3

Then we use the same formula that I described above, which gives relative value of 1:2:4:8 for QF, SF, F, and W. Of course for non-Slams that don't have that many rounds we'd have to apply something else. I'm tempted to only give points to being a runner up, at least for ATP 250s and 500s, because that data will likely be findable for most players, while it will be very difficult to find QF and SF results for most tournaments. So then we get:



What do people think? I guess my biggest caveat is that it doesn't value Grand Slams enough. I just can't see a WTF victory being worth 75% of a Grand Slam.

 

[ClayDeath]

nobody gives a damn about wtf. they have been trying to save that event for years by moving it around.

next stop: Madison square garden.

the sport has become too demanding. those who dominate during the early part of the year don't have much fuel left during the latter part of the year.


you just cant dominate all surfaces.

they keep throwing money at this event and they keep trying to move it around. it is what it is.

it is a useless event that is not given any weight by the true tennis historians.

 

[ClayDeath]

its still a cool exercise. we will have fun with it.

 

[britbox]

I'd approach it differently if we include pre-open era, but it would require a lot of research.

Top 4 Tournament (not always a major, as sometimes the Masters was a bigger tournament than the AO.)
Open Era: Winner: 100 pts Runner-up: 25 pts
Pre Open Era: Pro Major Winner: 60 pts Runner-up: 15 pts
Pre Open Era: Amateur Major Winner: 40 pts Runner up: 10 pts

5th Tournament in any given year (i.e. Olympics, WTF, AO)
Winner: 40 pts Runner-up: 10 pts

6th Tournament in any given year:
Winner: 30 pts Runner-up: 6 pts

7-14th biggest tournaments in any given year
Winner: 15 pts Runner-up: 3 pts

I wouldn't award any points for semis or quarters.
Pre-open era points would be split 60/40 on equivalent tournaments for pros and amateurs.

If pre-open era is taken into account, it makes it a much bigger research project and we can't use a simple methodology.

 

[tented]

Another problem with looking at non-Slam tournament results other than wins is just finding the data. As it is, there are some players that I can't find more than Slam finals for. And once we get into the pre-Open Era I'm not even sure the data exists. Tennis isn't baseball, where stats have been meticulously kept for over a 100 years.

Anyone know of a good tennis statistics site? I use Wikipedia and atpworldtour.com, but I'd love to find the tennis version of baseball-reference.com. I don't think it exists, but I would think something is out there...

I've hunted around for such a site, but haven't found anything better than those two. An archeologist will find the intact tomb of another pharaoh before we find a comprehensive reference for tennis data.

 

[tented]

But let me give it a shot, with my process explained. We could look at the ATP point values for each as a ratio to work with, and then find an internal ratio to apply to each. So we get:

Grand Slam: 2000
WTF: 1500
ATP 1000: 1000
ATP 500: 500
ATP 250: 250
Olympics: 750

Then divide those by the lowest common denominator, which is 250:

Grand Slam: 8
WTF: 6
ATP 1000: 4
ATP 500: 2
ATP 250: 1
Olympics: 3

Very intelligent. Point systems will always be subjective, of course, but using an established, universally accepted system as the basis for yours introduces a level of consistency which lends credence to the project.

I'm tempted to only give points to being a runner up, at least for ATP 250s and 500s, because that data will likely be findable for most players, while it will be very difficult to find QF and SF results for most tournaments.

I agree. Such data are probably either lost, or buried in the ATP archives, and will never be put online.

What do people think? I guess my biggest caveat is that it doesn't value Grand Slams enough. I just can't see a WTF victory being worth 75% of a Grand Slam.

Since the WTF winners can get anywhere from 1,100 to 1,500 points:

http://en.wikipedia.org/wiki/ATP_Entry_Ranking#Current_points_distribution_.282009_-_present.29

You could use 5, instead of 6, which would bring up the value of the GS wins.

 

[ClayDeath]

wtf is being given too much weighting.

it is harder to win a masters series event than it is to win wtf.

 

[El Dude]

britbox, I like your ideas but it veers into too complex - especially with trying to figure out 5th, 6th, 7th etc most important tournaments. That would be a very detailed analysis - I'm really looking for a system that can be figured out by looking at the Wikipedia career statistics page, and atpworldtour.com website.

There are some necessary simplifications that I think have to be followed: like equal value for Open Era Slams, even though the 1970 Australian Open and the 2013 Wimbledon are very different creatures.

Not including SF and QF results for a top 20 list is fine, but if we want to see beyond that it is problematic. Not to mention that it softens the greatness of a Chris Evert, who appeared in so many SFs. That said, we could increase the Wins and Finals by a bit, which was how I did it in a system I used privately, so it would be something like 1, 2, 5, 10 for Grand Slams.

So while I agree with you, tented, that having some basis in objectivity (the ATP point system), it doesn't accurately portray all-time greatness because the point system's primary purpose is to determine rankings within a given year, not all-time. For that wins and runners-up should have greater proportionate weight, imo, just as Grand Slams should have greater proportionate weight than other tournaments.

So here's another stab at it, with those factors weighed in:

Grand Slams: 1 QF, 2 SF, 5 F, 10 W
WTF: 1 SF, F F, 4 W
ATP 1000: 1 SF, 2 F, 4 W
ATP 500: 1 F, 2 W
ATP 250: 1 Win
Olympics: 1 SF, 2 F, 3 W

I reduced the WTF to be equal to the ATP 1000. I don't agree that its a completely joke like someone said - you still have to get by three top players to win - but it shouldn't be worth more than an ATP 1000.

Speaking of which, weren't there two year-end tournaments for awhile? In the 70s, I think?

As for pre-Open Era, I agree that Pro Slams should be worth more than amateur Slams. Who wouldn't have rather faced Roy Emerson and Fred Stolle than Rod Laver and Ken Rosewall? So I'll take britbox's basic idea and convert it to these point values:

Amateur Grand Slams: 1 SF, 2 F, 4 W
Pro Slams: 1 SF, 3 F, 6 W
Early Grand Slams (before split): 1 QF, 2 SF, 5 F, 10 W

What do folks think?

 

[El Dude]

OK, just to try it out I ran a few players through it - Federer, Sampras, Nadal, and Lendl, and got the following results:

434 Federer
424 Lendl
328 Nadal
317 Sampras

First of all, it seems odd that Lendl would rank so high but the reason he does so is because he played so many tournaments, especially mid-range (ATP 500 equivalent), of which he was 42-18 in finals! So he gets 102 points just for that level of tournament, compared to a range of 23-30 points for the other three.

I also tried to run Rod Laver but the simple fact of the matter is that its impossible to understand his record - he won a whopping 200 tournaments in his career, and I have no idea what is what.

I hate to say it, but I think a detailed analysis like this across eras is simply impossible. The game has changed to so much and, more so, the TOUR has changed - the schedules of players - so that even if we had all the necessary data (which we don't), it wouldn't make sense to compare it across eras.

In major league baseball, for most of history the schedule was 154 games and then it switched to 162 games. That's a pretty small difference, so records are basically intact across the two eras. But in tennis it is much wider.

The one thing that we CAN compare are Slams. They've always been around in one form or another, even if they've changed, split, and re-joined.

I would suggest that the only way to come up with a system is to do what I did about a year ago - a greatness quotient based solely on Slams. This doesn't mean the other tournaments aren't valid, just that A) Slams are far and away the most important mark of greatness, and B) other tournaments and the schedule change so much across time to make it impossible to compare.

So if we could keep to Slams as the basis of all-time greatness, then use other factors as "eyeball" tie-breakers if two players are close.

Using the above numbers, what if we used the following:

Early Slams/Open Era: QF 1, SF 2, F 5, W 10
Pro Slams: SF 1, F 3, W 6
Amateur Slams: SF 1, F 2, W 4

I'll run some numbers and see what I come up with.

 

[El Dude]

OK, I ran a bunch of players using those numbers and came up with this list:

253 Ken Rosewall
230 Roger Federer
209 Rod Laver
176 Pete Sampras
159 Ivan Lendl
157 Jimmy Connors
155 Rafael Nadal
147 Andre Agassi
143 Pancho Gonzales
141 Bjorn Borg
118 Stefan Edberg
113 John McEnroe
108 Novak Djokovic
102 Mats Wilander
101 Boris Becker
99 Roy Emerson
84 John Newcombe

I didn't run any pre-Open Era players other than Gonzales, Emerson, Laver, and Rosewall (and accounted for part of Newcombe's record being amateur Slams, thus the lowered total).

There are a few problems when looking at this list. One is that it really benefits longevity, so you have Rosewall better than Federer and Laver, or Lendl and Connors better than Nadal, or perhaps worst of all, Agassi better than Borg. Roger has a good chance of passing Rosewall, and Nadal will almost certainly pass even Sampras to enter the "GOAT Four," but it penalizes players like Borg, who was as good as anyone there for a few years.

I suppose this system - or any system, really - merely points out that you can only really measure all-time CAREER greatness, and that PEAK greatness is a different thing altogether, which could come about through looking at best year, best three years, and best five years, and weighing those results with career greatness to come to an overall greatness score.

 

[Riotbeard]

I think the WTF should be more than a master's personally. It's a different exercise, but to win you will have to beat at least 4 top 8 players. I dont get how that isn't tougher than a masters. I don't have much to add but am enjoying reading it.

 

[tented]

I think the WTF should be more than a master's personally. It's a different exercise, but to win you will have to beat at least 4 top 8 players. I dont get how that isn't tougher than a masters. I don't have much to add but am enjoying reading it.

Theoretically, couldn't someone win having only played 3 different players, because they play one of them twice? I don't know if it has ever happened, but I think it is possible.

For example:

- In one group, Player A beats Players B & C, then loses to Player D.

- Two wins is enough to advance Player A to the SF, and Player B also advances.

- In the SF, Player A beats Player E from the other group, and gets to the final.

- Also in the SF, Player B beats Player F from the other group, and gets to the final.

- In the final, Player A beats Players B again, and wins the tournament.

So Player A wins after only beating Players B, C, and E.

 

[Riotbeard]

I think the WTF should be more than a master's personally. It's a different exercise, but to win you will have to beat at least 4 top 8 players. I dont get how that isn't tougher than a masters. I don't have much to add but am enjoying reading it.

Theoretically, couldn't someone win having only played 3 different players, because they play one of them twice? I don't know if it has ever happened, but I think it is possible.

For example:

- In one group, Player A beats Players B & C, then loses to Player D.

- Two wins is enough to advance Player A to the SF, and Player B also advances.

- In the SF, Player A beats Player E from the other group, and gets to the final.

- Also in the SF, Player B beats Player F from the other group, and gets to the final.

- In the final, Player A beats Players B again, and wins the tournament.

So Player A wins after only beating Players B, C, and E.

Still 4 wins over top 8 players even if two are against the same guy.