PDA

View Full Version : Hybrid Tennis Ratings


Wuornos
03-24-2008, 09:05 AM
Deleted as found unuseful by board members. :cry:

A.Davidson
03-24-2008, 09:24 AM
Nice! Can't wait...Wish it went back to the times of Budge and co., though.

Wuornos
03-24-2008, 09:26 AM
Deleted as found unuseful by board members. :cry:

SgtJohn
03-24-2008, 09:37 AM
Hi Wuornos,

Ambitious project indeed, the output will certainly be very interesting.

I think that your other ELO ratings used data from every match played in the Open Era. How did you solve the riddle of the pre-Open era? I and many others have lists of tournament winners and runner-ups, but certainly not of every match played in every tournament... I also wonder about the way the Pro/amateur divide is taken into account.

Anyway, I'm sure the explanation will soon be posted.
Thanks,
Jonathan

Wuornos
03-25-2008, 04:14 AM
Deleted as found unuseful by board members. :cry:

SgtJohn
03-25-2008, 08:36 AM
Hi Tim,

I studied statistics in the past so I know what you mean. Is it possible to evaluate the confidence interval of your results for each era? Because there could be statistical ties in your rankings then if two players happen to be close enough...

Actually I was referring more to the pre-1968 era. We don't have every result for the ten years following 1968, but we have most of them. Before 1968, we have almost no complete drawsheets...basically just Slams, and pro tournaments through McCauley's book. So it seems hard to use an ELO-based method in a statistically satisfying way for the 1945-1967 period.

Thanks,
Jonathan

Wuornos
03-25-2008, 09:35 AM
Deleted as found unuseful by board members. :cry:

jeffreyneave
03-25-2008, 02:07 PM
two questions

Is this ranking system solely dependent on the quality of opposition you beat

In otherwords does winning matches at wimbledon get you more points than winning an ordinart tourname4nt. The current atp system has a factor of approx 4 for winning matches at wimbledon

and does winning a the final match give you more points than winning a 1st round match. In a 64 man draw the current atp system progresses:

7,15,25,45,70,100.

In other you gain 7 points for winning a 1st round match, but an additional 30 for winning a final match.

jeffrey

llgc8080
03-25-2008, 03:56 PM
I'll be waiting man!

Wuornos
03-26-2008, 02:50 AM
Deleted as found unuseful by board members. :cry:

Wuornos
03-26-2008, 04:43 AM
Whoops duplicate post

Wuornos
03-26-2008, 04:59 AM
Deleted as found unuseful by board members. :cry:

jeffreyneave
03-26-2008, 11:29 AM
I think you rather avoided answering the question. in a grand slam you arguning that you are likely to meet a better opponent. However, Wimbledom has twice as many points as a super nine even though the strength of the field is the same (all the best players played at palm springs last week but djokovic only earned 100 points for his win compared to 200 at the aussie open. ie super nine progresses
3,7,15,25,45,70,100
a slam
7,15,30,50,90,140,200.

The point about winning a final match is that its worth more than any other win in the event, regardless of the opponent's quality; remaining underfeated is worth something extra. In otherwords if a player plays 2 super nines and wins one earning 100 points, but loses in the first round in the other earning zero, he has a match record of 7-1. Another player plays the same 2 and reaches the semi-final in both, he earns a total of only 90 points. He's won more matches with a 10-2 record, but most people would regard player 2 as having a worse record.

Do you lose points if you lose or are you just awarded zero for the match as in most ranking systems ?

jeffrey

hoodjem
03-26-2008, 11:46 AM
The advantages are as follows:
4. Cross era comparisons will be possible allowing for the underlying assumption of a stable playing population.

5. I have found a solution to the pre/post open era problem and data will be included back to the end of WWII.


This should be amazing: a true and accurate comparison of great players across eras. I too can't wait.

Wuornos
03-26-2008, 12:07 PM
Deleted as found unuseful by board members. :cry:

jeffreyneave
03-26-2008, 12:58 PM
Of course I would not give the Aussie open in 1980 the same as today, although it would have more points than an ordinary event with the same quality field. However, 1969 the Aussie open would have the same points as wimbledon in 1969 and 2008

About finalists some people go to the extreme of

3,6,12,25,50,100,200 (rounded) for a slam instead of

7,15,30,50,90,140,200

because yuo derseve the points of all points your opponent has defeated as well as the points for just defeating him.

I would not go that far, but since this is an incremental system match by match you have to give more points for winning the final than earlier round wins. Going from the last 32 to 16 (15) is worth far less than winning the final (60 points).

As to wimbledon/slams not beimg worth more than any other events your going completely against tennis history, every grand prix or atp points system of the open era, and nearly every journalist whose
written or done world rankings. Look at history books. The only wins you can find easily are grand slams. Finding info or ordinary events before 1968 is almost impossible, but you'll have no problem listing grand slam winners.
Slams have from time to time being downgraded with weak fields such as Wimlbledon 1973 or the aussie open in the 70s, but Wimbledon has always been the top tournament in the world if had a strong field. Sampras's greatness is based on winning 14 slams (Always quoted as the record to beat for modern players), particularly his 7 wimbledons, not his tournament record or Agassi's success at super 9s. Every world no1 player of the open era regarded winning wimbledon as his greatest achievement. And the two who failed, rosewall and lendl, badly wanted to win the event and purposedly skipped the french in some years to prepare and rest for wimbledon.


Your going close to the italian statiscian tommasi whose rankings treated every tournament the same. That system produced connors as the world no1 in 1975. every other journalist choose Ashe because of his two majors wins at wimbledon and dallas. the atp ranking system of the time because of its low weight on wimbledon choose connors, but the members of the ATp voted for Ashe. I have redone the 1975 rankings based on today's atp system and ashe is a comfortable winner, with connors only 4th.

jeffrey

Wuornos
03-26-2008, 01:27 PM
Deleted as found unuseful by board members. :cry:

Wuornos
03-27-2008, 01:12 AM
Deleted as found unuseful by board members. :cry:

jeffreyneave
03-27-2008, 07:36 AM
Yes your ranking is likely to produce higher points for a slam than other events because all the other best players are entered and you play seven rounds. Its also likely to produce runs closer

3,6,12,25,50 ,100 rather than

7,15,25,45,70 100

because of your emphasis on the qual;ity of player you beat.

Howver, palm springs which has the same field as a slam has every chance of scoring as many points as a slam . looking at the results I suggerst that nadal would probably score more points at palms springs than the aussie open even though he reached the semi at both and everybody would regard his aussie performance as far more signiicant. Also look at fish's result in palm srings where I suspect he'll score more points under your system than the winner Djokovic and would come pretty close to Djokovic's score at the Aussie open because of the quality of fish's wins. Such results are ridiculous in the world of tennis. The slams/majors are far more important than other events which have all the best fields.

In 1969 and 1970 Laver won philadelphia against all the world's best players,but nobody regarded it as anywhere near as important as Laver's and Newcombe's wins at Wimbledon in each year. Many journalists (tingay, mcCauley, collins) gave Newcombe the world no1 in 1970 because he won Wimbledon , but very little else (he only won 3 other small events; In contrast to Laver who won fourteen often against strong fields; Laver beat Rosewall 5-0, Newcombe 3-0 and beat every other member of the top 12 at least twice; in comparison newcombe was 1-5 against Rosewall)

Sharapova probably does well because of wins at WTA canampoionship, where under your system I suspect her runner-up is worth more than Kuznetsova's runner-up at the us open

Rosewall And Lendl are the 2 great modern number ones (nobody regards courier, Muster or Moya as all time greats) not to
have won Wimbledon.

In 1975 Dallas was a major. more important than The Aussie open. Ashe had to win Dallas as well as Wimbledon to secure the number one ranking according To Ashe himself and Tingay.

the 0.5 probability ratio would not see such a large ratio progression if all the players you beat were of the same level : eg ranked between 90 to 99. Its the supposed rising quality of the opponents which raises the points more quickly (ie the 3,6.12,,25,50,100).

How long a time span do these ratings cover 1 year or 2 or 3 ?

Is it players an average rating ie total points divided by number of tournaments played

Could you give the points (up to and including Aussie 2008) for beating Henin (no1), the number 10 ,number 50 and the number 100 to see how much the points system is dominated by beating top players

Ther's nothing wrong with empasising the quality of the opponents you beat, but in the world of tennis its not the only thing. Wining matches, particulartly at the the majors is just as important. Fish's runner up at palm springs is not as good as Tsonga's runner-up at he Aussi,e although I suspect under your system it actually is.

jeffrey

Wuornos
03-27-2008, 09:41 AM
Deleted as found unuseful by board members. :cry:

Nickognito
03-29-2008, 12:12 AM
Even if I have in my mind a different method to evaluate tennis history and create an all-time (or at least an open-era) ranking - just in my mind, because i don't have a complete database to work with - I'm very grateful to you all for these kinds of threads.

They always show not only ability and knowledge about tennis, but also a real wish to find the best method, not for your own 'glory', but just for your involvement in analyzing our sport by numbers.

I would like to point out the differences between 3-setters and 5-setters, and i agree that slam tournaments have to be judged in a 'special' way.

I'd like to clarify that Tommasi's rankings assign the same points for wins and losses in every tournament, but give bonus-points for achievements in Grand Slam tournaments as well.

Anyway i hope Wuornos can solve his problems, cause i can't wait to see comparitive rankings between pre-open and open era. Good luck ;-)

c.

Wuornos
03-29-2008, 01:13 AM
Deleted as found unuseful by board members. :cry:

Tennis old man
03-30-2008, 08:59 AM
Thank you Tim, like always, great idea.

Wuornos
04-08-2008, 03:47 AM
Deleted as found unuseful by board members. :cry: