Tennis Forum banner
1 - 11 of 11 Posts

·
Registered
Joined
·
1,645 Posts
They are absolutely ridiculous - Greg Rusedski at #16 while Mark Philippoussis is at #35!?! Hewitt at #4 after a dreadful year. Calleri in the Top 10 (who??!).

People will ALWAYS moan about the rankings. FACT. And the only rankings fact you will ever need.

The ATP Champions Race was installed so that people couldn't go 'how come Ferrero lost but he got to #1' and EVERYBODY ignores it.

You can't win.

Rankings are fun, fun, fun. Bottom line - you have to work damn hard to get to #1 - that's why so few have on the WTA Tour, and no-one can say Kim hasn't worked hard, slam or no slam. But at the end of the player's careers it's Slams that will count overall, that's why Jennifer Capriati and Lindsay Davenport would be smiling way more than Kim Clijsters if they all retired now.
 

·
Registered
Joined
·
5,563 Posts
I think it's interesting to get a different perspective on the rankings from an outsider...but I agree, every ranking system is going to have its flaws and you can't keep changing it everytime it doesn't come out perfectly
 

·
Plainclothes Division
Joined
·
5,873 Posts
Brilliant! Let's replace a ranking system in which you can be #1 without winning a slam title, with a ranking system in which you can be #1 without winning a slam match!

And Wertheim complains about Serena's slams not carrying enough weight now, when they are awarding more than double the points of any tier I? Newsflash, bub, under this new system, there's no difference between a slam and a tier V, since women play best of 3 year 'round.

Under this new system, if Chanda were to beat Kim and Serena in the semis and final of, say, Filderstadt, she'd get rewarded more for it than Kim or Serena would if they beat the other and Chanda in the final and semis of the USO.

For someone that wants to put more importance on the slams, that seems a bit odd.

I can understand wanting to put emphasis on quality of losses, sets won, etc. Especially considering that my fave rarely loses to anyone but top players, and often in tough matches. But there definitely needs to be a higher threshold than 10 matches. And what's to stop someone who is, say, a legit top 15 player, from dropping down to tier IV's and V's and waxing lesser competition? Sure, she wouldn't get rewarded as well for beating lower-ranked players, but she'd make up a lot of that difference in set scores. It probably wouldn't get her from #15 to #5, but it might get her to around #9.

The only way to avoid this is to have a detailed reward system in place, which would be a lot like the QP system currently in use. But much more involved. For writers always complaining about how "complicated" the ranking system is (it requires what, 4th grade math?), they should really hate this system. And when will these morons get it through their head that Clijsters won't move up because she played "twice as much" as Serena. She will because she's played a full schedule AND put up good results. There's a reason why it's Clijsters, and Clijsters alone, who's about to overhaul Serena. Hantuchova and Dokic play more tournaments, yet they aren't even close. And to use an analogy even a sports writer could understand, there's a reason platoon players don't win batting titles in baseball. There's a minimum # of plate appearances required, to prove you can produce those kind of results over a complete schedule against a variety of opposition, instead of just in favorable matchups. There is a mechanism to allow a rare possibility of it happening (adding 0-fers until reaching the minimum, then if they are still the highest, they win). If someone plays less than a full schedule in tennis, they are gambling heavily on putting up stellar results every time out. If it doesn't happen, and/or if their schedule is just too light, and someone else happens to put up results nearly as good, only more of them, then this challenger gets to be #1.

I'd be curious to see a full detailed explanation of Sagarin's ranking system. He doesn't have to worry about overwhelming me. And I'd like to see how his women's rankings pan out. (Sprem at #5!! :))
 

·
Registered
Joined
·
1,517 Posts
Brian Stewart wrote:

Brilliant! Let's replace a ranking system in which you can be #1 without winning a slam title, with a ranking system in which you can be #1 without winning a slam match!
Well, that's possible with the current system. Suppose some player wins the year-end Championships, all nine Tier 1s, and seven Tier 2s. That's ~450RP for the Championships, ~2500RP points for the Tier 1s, and ~1400RP points for the Tier 2s, or a total of ~4350RP. I haven't added Quality Points in, but if we presume beating two Top 10 players in the semis and finals, that's 86 x 17 = 1462 Quality Points, plus whatever they get in the quarters/R16. If the four Slams were won by four different players, I doubt you'd have anybody above 6000 points.

In practice, it's not going to make a difference. The best players are going to win most of the time, whether it's the Slams or not.

And Wertheim complains about Serena's slams not carrying enough weight now, when they are awarding more than double the points of any tier I? Newsflash, bub, under this new system, there's no difference between a slam and a tier V, since women play best of 3 year 'round.
Well, I run a system similar to Sagarin's for the women, as I mentioned in a previous post. Playing Tier 5's would hurt you, because you're going to end up playing weaker players and have a worse strength of schedule. I just ran a Perl script to spit out the 18 players with the strongest schedules (minimum of 12 matches) over the past year, and here's what I got:

Code:
 1  WILLIAMS Serena             2539.7
  2  CLIJSTERS Kim               2529.3
  3  MAURESMO Amelie             2524.1
  4  HENIN-HARDENNE Justine      2522.5
  5  WILLIAMS Venus              2521.7
  6  SELES Monica                2515.3
  7  DAVENPORT Lindsay           2509.2
  8  RUBIN Chanda                2500.0
  9  CAPRIATI Jennifer           2491.7
 10  HINGIS Martina              2484.6
 11  DOKIC Jelena                2477.9
 12  MYSKINA Anastasia           2477.6
 13  HANTUCHOVA Daniela          2476.9
 14  RITTNER Barbara             2475.8
 15  DEMENTIEVA Elena            2474.9
 16  DECHY Nathalie              2473.2
 17  SUGIYAMA Ai                 2470.7
 18  PANOVA Tatiana              2465.4
I'm not certain why Rittner and Panova are there, but the other 16 seem like logical choices.


Under this new system, if Chanda were to beat Kim and Serena in the semis and final of, say, Filderstadt, she'd get rewarded more for it than Kim or Serena would if they beat the other and Chanda in the final and semis of the USO.

For someone that wants to put more importance on the slams, that seems a bit odd.
I agree with this assessment, although I should point out that unlike Sagarin and Wertheim, I don't care about the importance of Slams. I think it's somewhat of an anachronism with the present Tours. If you look at golf, it still makes sense to have the four designated Majors: There's a PGA of America and a PGA of Europe, and the top players from those and the other PGA Tours don't play all that often. The Majors are among the few events where they all do. This is unlike tennis, where it's not uncommon to get six or more Top 10 players at an event. In 2002, when the Top 10 was stable, there were six Top 10 players at Sydney, Stanford, and Filderstadt amongst the Tier 2's; Key Biscayne, Charleston, Berlin, Rome, the Canadian Open, and I believe Zurich all had six Top 10 players. (Actually Filderstadt had seven Top 10 and about 15 Top 20 players.) This year's Indian Wells was missing three of the top four players, but virtually everybody between #5 and #35 was there other than Mauresmo, and injured Seles, the retired Hingis, and an injured Kremer.

I can understand wanting to put emphasis on quality of losses, sets won, etc. Especially considering that my fave rarely loses to anyone but top players, and often in tough matches. But there definitely needs to be a higher threshold than 10 matches. And what's to stop someone who is, say, a legit top 15 player, from dropping down to tier IV's and V's and waxing lesser competition? Sure, she wouldn't get rewarded as well for beating lower-ranked players, but she'd make up a lot of that difference in set scores. It probably wouldn't get her from #15 to #5, but it might get her to around #9.
I for one don't believe in using set scores. I wonder if Sagarin even considered that winning a set 6-3 could mean that you've broken your opponent either once or twice, depending upon who served first. Besides, there's enough data as is to get a reliable ranking based solely on match results. Empirical evidence from chess shows that about two dozen matches is enough to get a pretty accurate assessment of a player's performance.

The only way to avoid this is to have a detailed reward system in place, which would be a lot like the QP system currently in use. But much more involved. For writers always complaining about how "complicated" the ranking system is (it requires what, 4th grade math?), they should really hate this system. And when will these morons get it through their head that Clijsters won't move up because she played "twice as much" as Serena. She will because she's played a full schedule AND put up good results. There's a reason why it's Clijsters, and Clijsters alone, who's about to overhaul Serena. Hantuchova and Dokic play more tournaments, yet they aren't even close. And to use an analogy even a sports writer could understand, there's a reason platoon players don't win batting titles in baseball. There's a minimum # of plate appearances required, to prove you can produce those kind of results over a complete schedule against a variety of opposition, instead of just in favorable matchups. There is a mechanism to allow a rare possibility of it happening (adding 0-fers until reaching the minimum, then if they are still the highest, they win). If someone plays less than a full schedule in tennis, they are gambling heavily on putting up stellar results every time out. If it doesn't happen, and/or if their schedule is just too light, and someone else happens to put up results nearly as good, only more of them, then this challenger gets to be #1.

I'd be curious to see a full detailed explanation of Sagarin's ranking system. He doesn't have to worry about overwhelming me. And I'd like to see how his women's rankings pan out. (Sprem at #5!! :))
Well, now we get to the details of how the rankings I've developed work (and I think I've addressed most of the problems you seem to have with Sagarin's system in my system). Basically, it's based on the system developed by Dr. Arpad Elo (do a Google search on the name for scads of information) for chessplayers almost 50 years ago, except with some modifications.

First, I entered all the results from the WTA Tour in the year I began (1999) into a spreadsheet. (I didn't do Challengers at the time because Alain Vincent's site and this one weren't around then!) Each player got a winning percentage, and I also entered who played whom. In addition to the winning percentage, I calculated the average (mean) of one's opponents' winning percentages (which I called 'strength of schedule' or SOS), and iterated a second time by taking the average of ones opponents' SOS numbers. I used Elo's formula to determine a "performance rating" for each player. (The beauty of the Elo system as regards chess is that it's actually predictive: if you take the players in a round-robin tournament and look at their ratings, you can use the formula in reverse to determine a player's expected winning percentage. The "performance rating" version of the formula is fairly commonly used to figure out the initial rating of new players.)

As you can guess, a player who's playing the Top 10 players over and over and winning 70% of the matches will end up with a better rating than somebody who's playing the same players and only winning 50% of the matches. And the player winning 70% of the matches against Top 10 players should be better than a player winning 70% of her matches, but amongst players ranked between #11 and #20. But is this third hypothetical player better than the one who plays the #1-#10 players and only wins 50% of the matches? That, of course, is what the ratings try to figure out.

What I started doing in the first week of 2000 was to make the system a rotating 52-week system like the WTA. Events from the first week of 1999 were removed and replaced by the events from the first week of 2000 and a new set of ratings were calculated, and so on for 150 or more sets of ratings and counting.

This makes winning tournaments very important: If you play around with some numbers, you'll discover that the difference in winning percentage between winning a (single-elimination) WTA event and being the beaten finalist is much bigger than the difference between the beaten finalist and the beaten semifinalist.

I've also added a few refinements to the system: I've set a minimum number of events for players to play (14); players who don't play enough get "penalty losses" which effectively lower a player's winning percentage, and which are a fraction of a loss for each event less than 14 a player has failed to play. (Fractional for several reasons, but mainly because a player who's played three events and reached the second round of both is generally likely to be better than a player who's played the minimum fourteen events and racked up a 3-14 record. If you required a minimum of 14 losses minus tournament titles, that would make the 3-3 player and 3-14 player have effectively the same record, which I didn't think was fair.) Venus Williams was #2 until 's-Hertogenbosch, when Clijsters passed her -- but only because of the penalty losses imposed on Venus for playing just 12 events. Clijsters overtook Venus' unadjusted rating after Stanford; and with San Diego removed Venus fell behind both Hénin-Hardenne and Mauresmo. Venus' rating withough the "penalty losses" would still be ahead of Mauresmo, but actually behind Hénin-Hardenne. (She has, after all, only won two titles in the past 52 weeks. Should being in the opposite half of the draw from Serena and losing to her in the final be worth that much more than being in the same half of the draw as Serena and losing to her in the semifinal? This is not a rhetorical question.)

Thanks to this site and Alain's, I'm able to include Challengers as well, although these (and Qualifying of WTA Tour events) are weighted less than the main draws of WTA Tour events. I've also got some dodgy adjustments made for playing players who have lost all their matches in the past year, mainly done to keep the ratings from deflating too quickly. As it is, they've deflated noticeably in the past three-plus years, and I'm looking at some adjustments for that.

Now, the system isn't perfect, but there are a few interesting results I can give:

Venus first became #1 under my Elo-like system when she won the 2000 US Open. She kept the #1 ranking until her title at Zurich 1999 was removed from the rankings, at which point Hingis took it over again. Hingis remained #1 until Amelia Island 2001.

Davenport was #1 at the end of 2001, just like in the WTA rankings. :) Venus was #2, and Serena was #3, and Capriati #4. Don't forget that besides her two Slams, Capriati only won one other event (Charleston) in 2001. Serena and Venus each had better winning percentages. Indeed, Capriati never got beyond #4.

Serena first became #1 after Wimbledon 2002; she'll be #1 regardless of what happens with Clijsters this week. Clijsters might be able to get to #1 if she wins two of her remaining three hardcourt events -- it would depend on her opposition. It's somewhat more likely that Clijsters could get to #1 after Shanghai and Leipzig.

Well, I see I've gone on far too long. If you've got any questions, you can visit my website http://tedstennis.tripod.com/index.html, although I think it's been quite some time since I updated the FAQ. Feel free to ask questions either here or by email at [email protected]. Just make certain not to use words like viagra, prescription, or digital cable since I've got active spam filters in action. ;)
 

·
Plainclothes Division
Joined
·
5,873 Posts
Interesting explanation, Ted. I've seen your rankings on occasion, and found them an interesting read.

I've never had the "slam worship" that most of the tennis media has. Which is why I find it so amusing that Wertheim would champion a system that doesn't put any more emphasis on them than on "regular" tournaments. (Although he didn't seem to have such a problem with a non-slam-holder grabbing the men's #1 ranking last year, when he said if Agassi did so, it would "validate the tour".)

So, anyone want to discuss the NFL's passer's rating formula while we're here? :)
 

·
Registered
Joined
·
5,684 Posts
One of the difficulties in using a 10 or 12 match minimum for the professional tours is that it doesn't encourage players to appear a a number of tournaments........ that is, if I am understanding this correctly, and there is every chance that I am not... Part of the reason that the ATP and the WTA have instituted their systems is to assist tournament directors around the world to attract players.

I think for the health of the sport there must be some mandatory tournaments and that the players have to have some kind of meaningful penalty if they do not play.

Of course the scheduling of the Masters' tournaments is completely crazy and needs to be rethought. As for the champions race -- lol -- no one pays any attention to it. The entry rankings are all that anyone cares about.

To me, the important difference between football rankings and tennis rankings is that football rankings have no effect on the game. Tennis rankings have an impact on the game. The football teams play opponents based purely on a preset schedule. This in not true of tennis.

I think Navratilova was right in her remarks about the number one ranking. It is not so terribly important. It is good to finish the year number one, but one's career will be judged on slams won. I also don't recall such emphasis placed on the number one status 20 years ago.
 
1 - 11 of 11 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top