Research into the 70s

[Slasher1985]

Sorry, but I can't figure out this one.

There is really no problem. No need to apologize. Some questions just don't have answers.

 

[Slasher1985]

Next question. Thank you for your help so far, tented.

Can you provide me with the correct draw to this tournament:

Lakeway

Unfortunately, it is broken on the ATP site.

 

[tented]

Next question. Thank you for your help so far, tented.

Can you provide me with the correct draw to this tournament:

Lakeway

Unfortunately, it is broken on the ATP site.

If you work backwards, you can figure it out. [Note: Numbers do not reflect seeding or ranking. They're only used to identify the number of people in each round.]

1 Richey
2 Laver

3 Cox
4 Smith

5 Nastase
6 Stockton

7 Okker
8 Metreveli

9 Riessen
10 Taylor

11 Ashe
12 Drysdale

13 Tanner
14 Kodes

15 Alexander
16 Pilic

---

1 Richey
2 Cox

3 Nastase
4 Okker

5 Riessen
6 Ashe

7 Alexander
8 Tanner

---

1 Richey
2 Nastase

3 Alexander
4 Riessen

---

1 Richey
2 Alexander

---

1 Richey

 

[Slasher1985]

^
Thanks.

I am now done converting 1973 points to the 1974 system and ready to start a first weekly pass through 1974.

 

[tented]

^
Thanks.

I am now done converting 1973 points to the 1974 system and ready to start a first weekly pass through 1974.

That's great news. Please keep us informed of your progress.

 

[tennisville]

Slasher your links in the main site are not working :huh:

 

[Slasher1985]

Slasher your links in the main site are not working :huh:

Yeah, those are old links when I was still using images. When Brit will have time, he will update them to how they are suppose to work.

 

[britbox]

Hi Slasher,

Is there an interim solution that I can put in place before I look at it more closely in a few weeks?

 

[Slasher1985]

Hi Slasher,

Is there an interim solution that I can put in place before I look at it more closely in a few weeks?

This should work:

<iframe width="700" height="900" frameborder="0" scrolling="no" src="https://skydrive.live.com/embed?cid=29EC224593D363F0&resid=29EC224593D363F0%21587&authkey=&em=2&wdAllowInteractivity=False&wdHideGridlines=True&wdHideHeaders=True&wdDownloadButton=True"></iframe>

 

[britbox]

Cheers Slasher - I've put that in as an interim solution. Thanks for your efforts.

 

[Slasher1985]

The ATP Chain Error: The story of improperly applied bonus points

The ATP rankings have been a tremendous effort ever since they were invented. Their elegance and simple nature have made it possible to create an objective system to show us the leaders of the tennis world.

Ever since the important data has been lost or forgotten, a quest has begun to check if ATP got the facts right with their own system. They did in most of the part, but there have been a few mistakes which will soon show as turning ugly.

It all starts rather simple and it involves luck. For instance, how lucky Newcombe should be that ATP decided to publish the post-Wimbledon 1974 rankings on July 29th. He managed to score 3 extra weeks as World Number 1. Although ATP's publication of the rankings once in a few weeks is completely justified, because of the difficulty of calculations, this has caused the rankings to be not as objective as they were intended, for you see, many tournaments were played and dropped in-between publications, and each change could show a different World Number 1 just for a few weeks, or one week, without ATP seeing it or telling us about it. The Wimbledon 1974 example is just one of the examples I could find right now, but I have not yet moved beyond 1974. Connors should objectively have scored 3 extra weeks as World Number 1 in July 1974.

Now comes the difficult part. Not knowing of the consequences, ATP decided to create a Bonus Point system completely dependent on the rankings just before a tournament is played. With it, beating Top 24 players would award you points. The bonus point system had two factors: a static one (points for eliminating seeds) and a dynamic one (points for eliminated ranked players). Once we try to calculate the rankings weekly, the dynamic factor plays havoc and inserts a chain error into the official rankings, since we are discovering that if we try to be objective and calculate rankings as per week of tennis played, which was a measure back then, we discover that the much faster modifications of the Top 24 lead to larger and larger differences, creating a Chain Error.

The chain error extends beyond 1974, because points start to drop out having different values than the original, and new ones are added depending on the altered rankings. In a way, the bonus system completely butchered the objectivity of the rankings. It's not ours to judge though, because as I said, ATP was completely justified, but it's ours to correct what once went wrong, and without altering the original system in any way, view the rankings as they should have been.

You can find my 1974 repaired rankings in the rankings database.

 

[GameSetAndMath]

Actually, I like the bonus points awarded to players for upsetting big-name players.
It adds an extra incentive for lower ranked players to beat the top ranked players.
I almost wish they bring it back.

 

[Slasher1985]

Actually, I like the bonus points awarded to players for upsetting big-name players.
It adds an extra incentive for lower ranked players to beat the top ranked players.
I almost wish they bring it back.

They were the best thing they could have thought up to be exact.

Once used properly, the rankings provide a "short-cut" to the Top echelon.

 

[Kieran]

That's fantastic, Slasher, a tremendous database!

As for the bonus points, it was a great idea in terms of giving added incentive to knock down a champ, but maybe getting rid of it was fair, because it means each match has the same at stake for both players...

 

[GameSetAndMath]

That's fantastic, Slasher, a tremendous database!

As for the bonus points, it was a great idea in terms of giving added incentive to knock down a champ, but maybe getting rid of it was fair, because it means each match has the same at stake for both players...

I disagree. For example in Wimbledon this year in the first round, Federer defeated
Victor Hanescu and Steve Darcis defeated Nadal. Who would in the right mind think that
Federer deserves the same credit (for defeating Victor) as Steve Darcis (for defeating Rafa).
Federer did something that he was expected to do and Steve did something that nobody
imagined he could do. So, I think it is fair that Steve be given more credit for his 1st
round match than Fed for this 1st round match.

In fact, the rating system works exactly like this in Chess (and I believe even in
club level tennis). If you are a 4.0 player, you are expected to defeat a 3.0 player
and so your rating will not go up no matter how many 3.0 players you defeat.
On the other hand if the 3.0 player defeats a 4.0 player and consistently holds his
ground agasint 2.5 players etc, his rating would go up to say 3.5 or so. I know for
sure this is how it works in Chess ratings.

To summarize, dog biting man is not news; man biting dog is.

 

[Kieran]

That's true, except both Federer and Darcis both won a single tennis match. Of course Darcis got more credit for it: his name is alongside Rosil and Stakhovsky as one of the modern great upsets. Does he deserve more points than Roger? Well, both Darcis and Roger were knocked out the next round. Did they get the same prize money?

I don't totally disagree with you and I think it adds spice to things, but I also think that players should hoist themselves up by their own bootstraps. I remember Sampras playing some guy years back and the commentators were describing how many bonus points the other guy would get for victory and I thought, how unfair! Sampras dedicates himself to keeping the wolf from the door both day and night, his standards have been beyond reproach, and a dilettante gets more points for winning the match than him!

I know, it's good for motivation etc, but really, who needs more motivation when they're playing Sampras or Federer?

 

[Slasher1985]

Well, the bonus points had a great purpose at the start. But when all became ABOUT the bonus points, they simply lost some of the charm. At the height of their power, they brought too much spotlight to the rankings themselves and made the game revolve only around them. They could also help flukes rise to Top 10 easier. Like for instance, if a qualifier would beat all top seeds and get a Grand Slam final after which fade into nothing, he could very well rise to Top 10 for one year and stay there. I don't really remember a case right now, but when I get to the 90s I shall study this problem. Nowadays, a Top 10 position is a measure of endurance and consistency, which can hardly be "fluked" IMO.

 

[Slasher1985]

Here's one error that I have no idea how ATP committed. Metreveli's official tournament count for 1974 was 281, yet, I have him at 353 and 17 tournaments. So, what am I missing? Simple it appears, the Nottingham final. ATP have not counted it, and there is simply no reason for it.

If you open up the 03.06.1974 rankings on the ATP website, Metreveli's 9th. On 29.07.1974 he is 13th.
On my corrected rankings, Metreveli is 10th on June 3. On July 29th, he is still 10th.

The clear conclusion: Metreveli's final in Nottingham from later June did not count for ATP at all. There is no reason for this.

May look like a simple enough error, but when you deal with a bonus system depending on rankings, it increases the error factor. With what I am encountering so far, it is possible that no rankings were correct until the year 2000.

 

[GameSetAndMath]

Well, the bonus points had a great purpose at the start. But when all became ABOUT the bonus points, they simply lost some of the charm. At the height of their power, they brought too much spotlight to the rankings themselves and made the game revolve only around them. They could also help flukes rise to Top 10 easier. Like for instance, if a qualifier would beat all top seeds and get a Grand Slam final after which fade into nothing, he could very well rise to Top 10 for one year and stay there. I don't really remember a case right now, but when I get to the 90s I shall study this problem. Nowadays, a Top 10 position is a measure of endurance and consistency, which can hardly be "fluked" IMO.

Of course, how much bonus points are given should be carefully figured out to eliminate
someone rising all the way up just because they beat a top player once in a blue moon.
At any rate, it appears that you are agreeing to it in principle. Now the details have to be
worked out.

How about something like this. The bonus will never be more than say 40% of regular
points. How much of the 40% bonus you actually get should be dependent on the differential
in the current rankings.

Something like that can be easily made to work fine.

 

[tennisville]

Here's one error that I have no idea how ATP committed. Metreveli's official tournament count for 1974 was 281, yet, I have him at 353 and 17 tournaments. So, what am I missing? Simple it appears, the Nottingham final. ATP have not counted it, and there is simply no reason for it.

If you open up the 03.06.1974 rankings on the ATP website, Metreveli's 9th. On 29.07.1974 he is 13th.
On my corrected rankings, Metreveli is 10th on June 3. On July 29th, he is still 10th.

The clear conclusion: Metreveli's final in Nottingham from later June did not count for ATP at all. There is no reason for this.

May look like a simple enough error, but when you deal with a bonus system depending on rankings, it increases the error factor. With what I am encountering so far, it is possible that no rankings were correct until the year 2000.

wait because of 1 error, you feel rankings were iincorrect till 2000 ? I dont get this . WHy should that be

I disagree with bonus points actually. No win should be greater than the other , its wrong for the superior player