Mens Tennis Forums banner
Status
Not open for further replies.

The Rankings Database

51K views 194 replies 36 participants last post by  Slasher1985 
#1 · (Edited)
#107 ·
1979 has now been updated. Now to talk about the ranking system.

Starting with 1979, the ATP adopted the star system to be used for their tournaments. The minimum requirement for a tournament to be counted for the rankings was to have at least $25000 prize money. That margin would make the tournament 1 STAR. Every increase of $25000 in prize money would mean 1 new star for a tournament.

For the first time since the ranking system existed, draw size and difficulty would also count for a tournament's category. A draw of 32 players or less would not give the tournament any extra star, 48 players would mean 1 extra star, 64 players would mean 2 extra stars, and, only in the case of Grand Slams, 128 players in the draw, would mean an extra 3 stars. Draw difficulty would count in much the same way. The total entry ranking of the first 20 players inside the draw would be counted (rankings of the first 20 players added), and if the total would make more than 1550, the tournament would not grant an extra star. If the total would be between 850 and 1550, 1 extra star would be given, if the total would be between 450 and 850, 2 stars would be awarded. And if the total would be less than 450, 3 difficulty stars would be given. With this system, Roland Garros was the top tournament of the year, with 21 STARS.

The bonus points system was already restructured since 1978, so, no change was made to it in 1979. Defeating a player ranked 1-8 would award 3 points, one between 9-16 2 points, and one between 17-24, 1 bonus point. Defeating one of the first quarter of seeds would grant 4 bonus points, a player from the second quarter of seeds, 3 points, a player from the third quarter of seeds, 2 points, and finally, a player from the last quarter, 1 point. The bonus points between 1974 and 1977 were only given for defeating ranked players, the double of points awarded for the same thing, after 1978.

The ranking system, ranking bonus and calendar are in the first post, along with the weekly rankings for this year.

A final note, because I've seen everywhere people trying to establish Masters in the 70s, incorrectly. Judging by the stars awarded, I created a tournament ranking, to determine the most prestigious tournaments of the year. I'd say that every tournament with 12 stars or more can be counted as MASTERS SERIES.

TOURNAMENT RANKING 1979
 
#109 · (Edited)
One more information to add about the 1979 research.

As with 1973-1978, a serious number of errors have been discovered in the original ATP rankings, and have been corrected: tournament misses by players, extra tournaments added or doubled, tournaments getting points for an incorrect round, bonus or qualifying points missing, etc. Obviously, the categories of tournaments were not the ones expected through the rules mentioned at the start of the calendar document, specifically, the stars assigned for tournament difficulty. The anomalies have been discovered through a mathematical algorithm, which I built, specifically aimed at detecting the original star assignations the ATP used. The stars were not corrected in order to maintain the original ATP assignations.

This file contains the stars as they should have been (with the original draw difficulty assignation and the mathematical algorithm's finds:

1979 MODEL
 
#112 ·
The idea behind placing two columns there was that both are accurate, but it's all your choice to decide which you prefer. Some fans here prefer the original categories of tournaments not be changed, whilst others want 100% corrections of errors. The same went for 1973, where a side run was made with Wimbledon included, despite it not being part of the original rankings. :)
 
#119 ·
Slasher, how many weeks at #1 in total in the Open Era do you have for Laver and for Rosewall?

Also interested in weeks at top 10 for Gonzales in Open Era
 
#120 ·
Gonzales Top 10 Open Era: 148 weeks (52 in Top 5, 22 at #2).
Laver #1 in Open Era: 187 weeks.
Rosewall was not #1 in the Open Era (the previous iterations of the calculations were ignoring a lot of 1968 and 1969 results, they weren't too realistic), but he was a total number of 73 weeks at #2, 154 in Top 3, 272 in Top 5 and 414 weeks in Top 10.
 
#123 ·
Wow so Laver really ate up all of those weeks, didn't he? Wow. And Rosewall, holy shit, 414 weeks in top 10? In Open Era alone? That is like 8 years :lol: So Rosewall was top 10 from 68-75/76 or so? He was still top 10 at age 41 :eek:

How about Ashe? Ashe has several weeks, yes?

I am also interested if the ATP's 268 for Connors is realistic, I have him closer to 190 in my calculations, and Borg's 150 ATP weeks or so closer to 200

I know Vilas also got like 3-4 weeks, right?
 
#125 ·
Status update:

1980:
[PRELIMINARY RUN]

This preliminary run has the purpose of employing the rules of the rankings in order to determine a mockup of the stars assigned to tournaments in 1980. This initial model will then be run through the algorithm to determine the actual stars ATP awarded to tournaments. You will notice (if you access the files via OneDrive) that the points will differ significantly from the official figures, although placement of players may not differ by much. The succession of the #1 player is exactly the same as the official records show.
 
#137 ·
I've updated the database for 2015, and also upgraded the status with 1980 with the primary result of the mathematical model, the calendar, the ranking system and ranking bonuses. You can now find the weekly rankings inside the OP and here, but note they are NOT final, one major feature missing being satellite data.

1980: (RANKING SYSTEM|RANKING BONUSES - 1980 ORIGINAL SYSTEM --- CALENDAR)
[UNCONFIRMED]
07/01/1980
14/01/1980
21/01/1980
28/01/1980
04/02/1980
11/02/1980
18/02/1980
25/02/1980
03/03/1980
10/03/1980
17/03/1980
24/03/1980
31/03/1980
07/04/1980
14/04/1980
21/04/1980
28/04/1980
05/05/1980
12/05/1980
19/05/1980
26/05/1980
02/06/1980
09/06/1980
16/06/1980
23/06/1980
30/06/1980
07/07/1980
14/07/1980
21/07/1980
28/07/1980
04/08/1980
11/08/1980
18/08/1980
25/08/1980
01/09/1980
08/09/1980
15/09/1980
22/09/1980
29/09/1980
06/10/1980
13/10/1980
20/10/1980
27/10/1980
03/11/1980
10/11/1980
17/11/1980
24/11/1980
01/12/1980
08/12/1980
15/12/1980
22/12/1980
29/12/1980
05/01/1981 [YEAR END RANKINGS]

Also, you can see the category ranking of tournaments here:
Ranking of tournaments
 
#138 ·
Hi Marian,

Thanks for the preliminary version of 1980.
However, looking at the Excel sheet, the stars often do not add up to their total value (like in the case of Rome).
When do you think you can clarify all those stars?

It seems the ATP has been rather biased toward their own Cincinnati event, the quality of the draw does not justify giving it 3 draw difficulty stars.
 
#139 ·
There have been a few inconsistencies as data is not yet final, and I run a few more calculations before I update the calendars, but you should now see the latest versions of both calendars and weekly rankings, in sync with each other.

Regarding Cincy, in 1980 the non-final math model shows he got 1 difficulty star instead of worthy 3. You can see what the calculations show on how many stars events should have received here:

https://onedrive.live.com/view.aspx?resid=29EC224593D363F0!46435&app=Excel
 
#140 ·
The star system is now almost final. Hardly will there be any modification to tournament categories now, as the Top 100 model fits in with precision (sans the usual errors which were corrected). Satellite data is inserted for Top 100 players, so it can be seen at a semi-final state now.

The status of 1980 has been upgraded to
[CALCULATED]

and the weekly rankings can be seen as before in the links above or the OP.
 
#141 ·
The Top 200 is now calculated, a few modifications were made to the calendar's Challenger circuit, but the changes are not yet visible in the links above. I am now going to finalize the satellite input process, which must set all the minor players to their year end values. Take note of a minor problem with the ATP Challenger Tour of that year: The Japanese (April) and Italian (September) sets of tournaments were each part of a 5 tournament circuit, and although 3 of them each had 25000 dollars in prize money, both circuits were considered satellites in the official rankings and noted as such (counted as one single tournament).
 
#145 ·
The Top 200 is now calculated, a few modifications were made to the calendar's Challenger circuit, but the changes are not yet visible in the links above. I am now going to finalize the satellite input process, which must set all the minor players to their year end values. Take note of a minor problem with the ATP Challenger Tour of that year: The Japanese (April) and Italian (September) sets of tournaments were each part of a 5 tournament circuit, and although 3 of them each had 25000 dollars in prize money, both circuits were considered satellites in the official rankings and noted as such (counted as one single tournament).
Well, that bold Italian circuit was not a satellite after all, but accepted as Challengers. The inconsistency came from the fact that I assumed Challengers were awarding bonus points to players for defeating seeds. Seems that, starting with 1980, this wasn't true anymore. Of course, I need to confirm this, and since, it only became somewhat visible once minor ranked players were broken down to points, it's gonna take some time.
 
#144 ·
Any chance for pre-Open Era? :p

I use www.tennisarchives.com for some of that but they don't have all of it

Karoly Mazaks' Concise History of Tennis is also pretty good
 
#147 ·
Alright, the calendars should be final right now, with all stars in place. One last modification was made to Sofia. Although a $75000 tournament, it seems it had 2 stars, reason being unknown to me, but after 1979, I am out to preserve all of ATP's original categories, even the wrong ones.

Also, it is confirmed that Challengers didn't award bonus points for taking out seeds, and that Italian autumn Challenger circuit was treated as such, and not as a satellite.

Top 300 players are in place at year end. Now come the majority of players, the ones which scored 20 points or less.
 
#148 ·
Alright, the calendars should be final right now, with all stars in place. One last modification was made to Sofia. Although a $75000 tournament, it seems it had 2 stars, reason being unknown to me, but after 1979, I am out to preserve all of ATP's original categories, even the wrong ones.
That is great news.
I understand you want to preserve all of ATP's original categories, but there must be category errors because of the incorrect designation of draw difficulty stars.
Is the Excel sheet showing the correct stars in post #139 still up to date?

I can understand misplacing a draw difficulty star by one unit, but in the case of North Conway, Indianapolis and Cincinnati the ATP undercategorised them by 2 stars. It almost looks like the North American summer circuit was penalised in some way.
 
#149 ·
I was thinking the same thing actually. In both 1979 and 1980 there were differences to the expected star system. In 1979, a large numbers of tournaments didn't fit with it, in 1980 much less tournaments didn't fit with it, but still in both cases, the summer American swing had less stars than expected.

Yes, the #137 calendars and #139 calendar model are now in their final star assignment mode. The calendar will still need to fit the 25 satellite circuits of 1980, but their final position isn't yet determined (or even if there really were 25 or 24, or 26).

Also yes, I agree with you that there are Category Errors (Sofia is more than clear one such error). The other tournaments, not so clear, because they all earned extra stars for difficulty, and we can't identify if the error was made concerning draw difficulty or prize money category. Actually, this is why I would only focus on Sofia for a category error. I am not even confident if the draw difficulty calculations I made are what ATP would have made. Sure, the rules are the same, but the starting point is not. I used my ranking sheets to determine the values, ranking sheets which already altered in some way the official rankings, by correcting errors. It's actually better this time to preserve the original star categories than alter them, because I can't be 100% sure that the draw difficulty figures I have are historically good. Especially those that are close to one of the limits: 450, 850 or 1550. Better leave the original categories than to insert an error caused by uncertainty. When there's a 2 star difference though, it's obvious that the ATP had some reason for that tournament to be considered a lower/higher star category, we just can't be sure what the reason is (it could also be a prize money category error).

What I do expect to see though, is even less differences in star assignment for 1981, as 1980 is much more precise than 1979. I count 24 differences in 1980 and 45 in 1979.
 
#150 ·
It's actually better this time to preserve the original star categories than alter them, because I can't be 100% sure that the draw difficulty figures I have are historically good.
OK, I understand the difficulty.
So we cannot be sure where the ATP got the stars wrong. The draw size seems to be pretty obvious. How confident are you that you got the correct prize money?

Another question: the year-end rankings for 1980 are pretty close to the official rankings, only Solomon gained an extra tournament, but positions in the Top10 are not affected.
But in 1979 Ashe had a much higher average with the same number of tournaments, while the opposite is true for Higueras. These two players actually swap places in your rankings. Could you unravel what error the ATP has made with these two players?
 
#152 ·
I'm looking at the ranking sheet for the 1970 Australian Open, but the other early sheets seem to be the same. The average column is determined by points/tournaments. That works (this week, anyway) down to 8 tournaments. At 7 (and below) tournaments, however, the number of tournaments stays at 8. This makes perfect sense... so that a player has to play so many tournaments. My question is: how was the 7 to 8 cutoff established?
 
#153 ·
The minimum number of tournaments was established to aid the official rule of 1973 which stated a number of 10 tournaments was minimum and 12 after 1974. You will see in the official rankings that all players participating in a number of tournaments less than the minimum had the divisor be the minimum. This was done in order to eliminate players winning a Grand Slam and not participate in any other tournaments suddenly become Top 10 or even number 1.

The idea was to gradually increase the minimum between 1968 and 1973 in order to not create any sudden changes. 4 in 1968, 6 in 1969, 8 in 1970 and 10 from 1971. This is also determined from the number of tournaments considered for rankings, lower in 1968, and higher in later years.
 
#154 ·
Since there have been questions through the private channels.

1980 is in the process of being confirmed (all the calculations, even those for lower ranked players are done). The process is tedious, as all numbers must be good and mistakes eliminated. The process has also been slowed down (besides my 2 month break) by the fact that the ATP eliminated the system's automation possibility by converting the Spanish version of the website to the new format, so, along with the Live Rankings, the Historic Rankings system also had to be modified.
 
Status
Not open for further replies.
You have insufficient privileges to reply here.
Top