The Goat Debate

[mick1303]

The weighted ranking thread (https://www.talkabouttennis2.com/viewto ... 444#p34444) is updated - now it has 1968-2021 data in the tail of the comments section.

[ti-amie]

The weighted ranking thread (https://www.talkabouttennis2.com/viewto ... 444#p34444) is updated - now it has 1968-2021 data in the tail of the comments section.

Thanks for doing this Mick.

[mick1303]

I've actually created a parser to collect doubles results from ATP site. It is in the beta stage now - it works, but I have to test it more. And this would be only a part of the story. To fully integrate doubles results, I will have to insert all the cross-references to the data environment. The doubles shall undergo the same validations that is now implemented for singles. So far imported only 10 tournaments from this year...

[ponchi101]

How are you going to judge the doubles? That is the truly critical question. For example, Pam Shriver has more Grand Slam doubles (21) than Hingis (13), but all but one of Shriver's titles came with Navratilova. How do you balance that data?

[mick1303]

I don't have an answer for this. Alas, it is a long way before I will have to decide anything on this matter.

[ponchi101]

Finding your proper balance to judge how the results truly make sense will be hard. Maybe:
If the player you won the doubles' slam has more slams than you, you get 3/4 of the points.
If you have more slams than your partner, you get 5/4.

Don't know, throwing options.
(Ah, we want this by Monday... )

[mick1303]

The most obvious thought - calculate doubles ranking in two passes. First pass - similar to singles weighted ranking. Then during the 2nd pass in the categories that rely on ranking, the ranking share for the tournament will be split between the doubles partners proportionally to the ranking calculated in the first pass. Then to apply this idea even further - multiple passes shall be applied until the ranking of the next pass for each player will differ from the previous pass insignificantly.

The more complicated variation of this method - to account for the ranking change over the time. Then the same method will be applied not to the whole duration of the data at once, but it will be split by time intervals. The smaller the time interval - the more precise result will be in the end.

I have to say - I'M NOT signing up to do such calculation ))) I will start with something much more lightweight.

[ashkor87]

I would start with the premise that one cannot compare across eras anyway, and certainly not by counting majors..in earlier times, majors were simply not considered all that important. Take one era at a time..there are only a handful of candidates anyway..for instance, today there are 3 at the top,- Nadal is the best doubles player among them so he gets the nod. in the era of McEnroe/Borg/Connors, it is equally obvious. Shriver was never a candidate anyway..and so on...

[ponchi101]

But aren't you tacitly comparing eras when you say that in the past, majors were simply not considered that important? Sorry Ashkor, but you are comparing eras when you say that "in the past, such and such thing". It is a comparison.
The debate is for fun. But comparing eras is part of it. You have other subtleties: in the past, 3 of the 4 slams were on grass. Laver never won a slam on a hard court, not because he did not but simply because there were none. The players had to go pro to make a living, today's players are full time tennis players, and so many other factors. You have the bias or recently, meaning that I never saw Kramer or Gonzalez play.
About Shriver: she got lost in the Martina/Chrissie era but she made on slam final and has several tournaments to her name. She was not just 25% of a doubles team. Still, you are correct. She was never a candidate for doubles GOAT, as we know who carried 2/3 of the court there.

[ashkor87]

But aren't you tacitly comparing eras when you say that in the past, majors were simply not considered that important? Sorry Ashkor, but you are comparing eras when you say that "in the past, such and such thing". It is a comparison.
The debate is for fun. But comparing eras is part of it. You have other subtleties: in the past, 3 of the 4 slams were on grass. Laver never won a slam on a hard court, not because he did not but simply because there were none. The players had to go pro to make a living, today's players are full time tennis players, and so many other factors. You have the bias or recently, meaning that I never saw Kramer or Gonzalez play.
About Shriver: she got lost in the Martina/Chrissie era but she made on slam final and has several tournaments to her name. She was not just 25% of a doubles team. Still, you are correct. She was never a candidate for doubles GOAT, as we know who carried 2/3 of the court there.

Yes, you are right..comparing across eras is key to the GOAT debate, and what makes it fun...! I guess my real problem is with numbers (I consult and teach that for a living and have a PHD in it) ..it is not right to simply count majors, making it an accounting exercise..I would much rather look at a player subjectively and say, really, I like x more than y because of whatever..so my favorite is Gonzales simply because he is so mysterious, I never saw him play either..then Federer, just for the grace and simplicity...
Because of this doubles twist, I rate McEnroe higher than, say Sampras..and Hingis higher than, say, Graf and Seles..

[ashkor87]

But of course, the GOAT debate is about who was better...so liking someone better is not enough..we have to ask, was Laver better than Federer? It is only fair to see how good they were at their best (they all have lows) and check if they played at that level long enough, not just once or twice (Adriano panatta) ..then ask, at his best, would Laver have beaten Federer..fun to debate but impossible to prove.. i would actually say yes, Laver would probably have beaten Federer..much as I like Federer!

[ashkor87]

And anyway, if we want a numerical rating, Elo is the best, it tells us Djokovic is GOAT

[3mlm]

But of course, the GOAT debate is about who was better...so liking someone better is not enough..we have to ask, was Laver better than Federer? It is only fair to see how good they were at their best (they all have lows) and check if they played at that level long enough, not just once or twice (Adriano panatta) ..then ask, at his best, would Laver have beaten Federer..fun to debate but impossible to prove.. i would actually say yes, Laver would probably have beaten Federer..much as I like Federer!

If they were both playing during the same era I'm sure each of them would have beaten the other multiple times.

[ponchi101]

Those are the fun questions. Can you imagine Laver with a modern racquet? I can see Federer, keeping his same style, being successful with wood. I can't see Rafa even having anything close to his same style with a wood racquet, because with that forehand and a 15 ounce Jack Kramer, his elbow would be destroyed after one set. He would be successful, with a different style. Novak the same: that quick, bunting BH passing shot would not make it to the net with a Borg Pro, much less pass Laver or Roger or Pete.
I can see Martina having even better groundstrokes (if that were possible) had she grown with synthetics. I would say Rosewall would have developed a topspin BH if he had grown up with a Pro-Staff 6.0 or Graphite Edge. Borg with a graphite racquet would have been even better, basically impossible to go through (he would have still have had his speed).
As 3mlm, Laver Vs Roger would end up like Borg/Mac (7-7), or something very close.

[mick1303]

I've already collected all the doubles results of main draws of the tour events and Davis Cup ties for the year 2021. Tweaked a parser code in process of doing so. Also I have an adaptation of a weighted ranking calculation for doubles players. But it is a bit raw and needs testing and possible adjustments. Continue to populate results going backwards in time