Another Rating System Much More Accurate Than the ITTF World Ranking

This user has no status.
This user has no status.
Member
Jan 2017
120
176
423
For someone who doesn't know the most accurate rating system story of Rating Centrals.

fatt
He might have lost interest. He wanted to be part of the American table tennis adventure but he also pursued a wider audience re. tt ratings: the world!When a PhD is willing to spend their time on a ping pong rating system for a small association like USATT, they should be given all the leeway they need and I suspect USATT wanted to keep DM on a leash (telling him what to program lol) which he declined.
It's a shame, his system is the best in the world; waste at its best! but he put the work in, it is there available and he can pass it on or just wait for a better opportunity.
More to come.

skip
As ITTF new ranking system completely screwed up players' playing level, higher ranked player doesn't mean higher playing level - world-wide TT fans will be more interested in RC system than ever before.


Don't know if RC no longer updates ITTF sanctioned events is due to the system's creator (Dr, David Marcus) lost interest on his own creation/technical difficulty to capture the match results from the new ITTF database or the funding problem.


If it is the former, not much we can do about it. If it is due to the latter, funding problem, then there is something we can do (or try): (A) The world-wide TT fans can write to TT equipment mfrs (Butterfly, DHS, Donic, Nittaku, etc.) asking them to advertise on RC's site. (B) And we can launch a donation program and get world-wide TT fans to donate to RC. (Personally, I will donate.)

David Marcus (AUTHOR OF RATING CENTRALS)

Don't see much, if any, advertisements by Table Tennis equipment
companies.

See

http://www.ratingscentral.com/SupportRatingsCentral.php

Hope they are alright.

Sean and I are fine.

I have no connection with the company that operates the site,

Unfortunately, there is no company. It is just a hobby for me and Sean. See

http://www.ratingscentral.com/SupportRatingsCentral.php

They probably got pissed off at the ABS ball and quit the sport completely

Sean still coaches and, I assume, plays. I haven't played in a couple of
years, so I haven't tried the ABS balls.

DM occasionally comes on the forum to chime in on something RC related
being debated... I would be patient.

Hope he will show up and let the TT fans know.

I don't read this forum. Using it to communicate with me is playing "Six[/FONT]
degrees of separation"

https://en.wikipedia.org/wiki/Six_degrees_of_separation. Someone told me
today that there was this thread.

If you want to communicate with me, I don't think I'm hard to find: I have a couple of websites with contact info:

http://www.davidmarcus.com/Personal.htm
http://www.ratingscentral.com/Problems.php

Posting to this forum is not a reliable method.

Might be something to do with all the ITTF's website changes and how this
makes the transferring of results to RC trickier.

About a year ago, ITTF unveiled their wonderful new site. It looked all
cool and was probably written in AngularJS. However, it was hard to view
the draw sheets: you could only see one match at a time. You wonder who
thought that was a good idea. Probably other things were messed up, too.
However, they left the old site up, so you could still find the results
there. And they soon added links from the new site to the draw sheets on
the old site.

In October, they sent out this press release:

http://mailchi.mp/d1e7fe0e4fcd/ittf-...l?e=57190ec215

So, another disastrous IT project bites the dust. At that point, the old
site disappeared and the new site stopped even having its one match at a
time results. Now, it only has PDFs of the draw sheets.

Scraping the PDFs wouldn't be a lot of fun, and it sounds like ITTF will be
trying to make changes to the site. We've seen a couple of other sites that
have ITTF data in a more usable format, but we don't know how official they
are or how they get their data. So, I'm waiting to see if things stabilize
before spending a lot of time trying to rewrite my scraper to grab data off
the current site.

We have no contacts in ITTF, so if anyone knows anything, feel free to pass
it on. I have no idea if ITTF intentionally makes the data hard to get or
it didn't occur to them that people would like to be able to get the data.

> For ITTF sanctioned events: This requires work on RC's part

It took some effort to write the scraper program. But, it generally only
took me about five minutes to submit each event (except when ITTF did
something weird with the data, like had matches where the winner was
neither of the players or made up new event codes).

> For League events: The League Director enters the data, with no work
> needed on RC's part

Yes, although it is impressive how many things event directors can do wrong
or need help with. For example, despite the instructions clearly saying
that event directors should make backups of their data, I get a steady
stream of directors telling me that they don't have the data for their old
events because they no longer have the computer they used to submit the
events.

> I suspect USATT wanted to keep DM on a leash

They mostly ignored us. The original project was suggested by Sean and the
USATT Ratings Committee. Before spending 2000 hours developing a new
system, I wanted some indication that it was wanted and would be adopted.
Some board members seemed supportive. But, words are cheap. I said that to
do the research and development, I'd need some better computing resources
(remember, this was long ago when computers were primitive). I said USATT
should pay for a decent computer and software that I could use for the
project; this would cost around $5,000. I would contribute the 2000 hours
for free. The board declined. I wasn't in the room when the Executive
Director and board discussed it, but I'm pretty sure it was because the
Executive Director didn't want to spend the money. The Executive Director
then said we could get a grant from the USOC. Perhaps foolishly, I agreed,
and we did that. Skip ahead a couple of years, and the system is ready and
the Ratings Committee recommends USATT adopt it. By this time, the few
somewhat-friendly people on the board had been voted out, and the board was
full of nutty people. We attended the board meeting where it was decided
whether to adopt the new system. The discussion was completely nutty with
board members saying things that were ridiculous (with ridiculous
PowerPoint presentations to back them up). I think the vote was 7-5 against.

In retrospect, I'm happy to not be involved with USATT. They are very hard
to work with. Tangential to Ratings Central is Zermelo, my tournament
software. This supports USATT tournaments. Twice in the last few years,
USATT has changed how they handle digital submissions of tournament
results. The first time, they emailed the three people who had tournament
software and said the system would be changing in one week (they were going
to go live with the change and announce it in literally one week) and they
were sure we would be able to modify our apps to support the new formats by
then. This was the first we had heard of this, and they didn't provide any
sample data or specs for the new formats. After several months of
discussion with them, I did manage to get them to tell us (decide) what the
new format would be and modify Zermelo to work with their new setup. Then a
year later, they changed it again, and again did not coordinate with us.

it is there available and he can pass it on

That might be hard. There are quite a few interconnected parts.

(A) The world-wide TT fans can write to TT equipment mfrs (Butterfly,
DHS, Donic, Nittaku, etc.) asking them to advertise on RC's site.

We have sometimes had people pay for ads, but not regularly. I doubt you
will be able to convince companies to spend their money by writing to them

(B) And we can launch a donation program and get world-wide TT fans to
donate to RC. (Personally, I will donate.)

We welcome donations:

http://www.ratingscentral.com/SupportRatingsCentral.php

We accept donations in any amount, so you are welcome to donate based on
what the site is worth to you. If you are curious what our expenses are
this may help:

We could cover our out-of-pocket expenses

* if each director who submitted an event in the past year donated
US$5.91,

* or if each director donated US$0.25 for each event that they
submitted in the past year,

* or if each player who played in an event in the past year donated
US$0.05,

* or if each player donated US$0.01 each time they entered an event in
the past year.

David
 
Last edited:
This user has no status.
This user has no status.
Member
Dec 2016
496
388
1,723
ERT,

You made a very interesting claim that that Ma Long has a 15% chance of beating Dima based on the current rankings. How did you arrive at this result?

Thanks.

Hi,
I think he meant that using ITTF's results and applying common probabilities analysis (much in the way ELO does), these figures would suggest that Dima will win five times out of six or so, which is obviously hard to believe and the basis for this very legitimate thread. It's like looking at FIFA results, and let's say 1st place Germany (no idea if they are) are facing 17th place Serbia (ditto). Rankings analysis would back your gut feeling if it suggested Germany was expected to win five out of six and in this case it does, but that's only if the rankings are trustworthy. Using the ITTF system, Serbia (sorry for picking on you!) might as well end up top of the table and be expected to win by the same margin, in any given year, if they bother to turn up at many more tournaments AND still get very good results (this last point being downplayed a bit in the original post and website entry of this thread, my only criticism) while Germany are re-painting their stadium or something...


One thing though, the ELO system is slightly dated (while still much in use). It'd be interesting to compare the results using Glicko-2 instead, which would account for erratic results (I'm quoting) and better take into the account the number of games played (still an important factor! - using the football analogy, Sam Allardyce of England, having been team manager for an only game - and won it, an all-time 100% record - shouldn't get a 3450 rating, in all fairness to him). Glicko also gives a fair idea of how trustworthy every individual rating becomes, based on some of the factors just mentioned and the resulting rating deviation. And it tends to reach viable results faster than ELO, with any given pool of games/samples.

http://www.glicko.net/glicko/glicko2.pdf
https://en.wikipedia.org/wiki/Glicko_rating_system

My apologies if any of this has been mentioned in this or the other thread, I haven't read it all (yet). Interesting, though.
 
This user has no status.
This user has no status.
Well-Known Member
Oct 2014
12,277
17,740
44,291
Read 17 reviews
Hi,
I think he meant that using ITTF's results and applying common probabilities analysis (much in the way ELO does), these figures would suggest that Dima will win five times out of six or so, which is obviously hard to believe and the basis for this very legitimate thread.

Maybe you can explain to me how and why these probabilities should be applied to a system that is not an ELO system? LEt's skip the why for a second and focus on the how.
 
Last edited:
  • Like
Reactions: zeio
This user has no status.
This user has no status.
Member
Dec 2016
496
388
1,723
I think again that it boils down to that "gut feeling". You may want rankings that give as fair an impression as possible of all player's relative strenghths, I'd imagine, especially (as mentioned by OP) if Olympics and such seedings will be based on that list. In this case ITTF encourage more participation in tournaments (no bother if you lose, just come up and score as many points as possible), which is also somewhat commendable.
 
Last edited:
This user has no status.
This user has no status.
Member
Dec 2016
496
388
1,723
Here is a good resource:
https://lichess.org/qa/6/how-does-the-rating-system-work-on-here

It's from a chess website and I don't feel I am pulling any plugs here, as as it all as free and openly run as you'll find [edit: not by me!]. (Props for chess too!)

Just think of Fan ZhenDong and Jean-Michel Saive as chess players (I hope they are / have been).
 
Last edited:
  • Like
Reactions: Andy44 and TTFrenzy
This user has no status.
This user has no status.
Member
Dec 2016
496
388
1,723
Maybe you can explain to me how and why these probabilities should be applied to a system that is not an ELO system? LEt's skip the why for a second and focus on the how.

Oh sorry ahah I read as though you asked for more of the "why", not less! ERT is best suited to explain his own figures, but in any way it's more a case of applying a "system" (ELO, ITTF's own, /else) to the same data (the players' results), although they will use different factors or stats within these results. As mentioned in the resource I posted, none of these rankings end up as being "absolute", and one could really use any number of systems which all have their own merits.


*edited typos
 
Last edited:
This user has no status.
This user has no status.
Well-Known Member
Sep 2016
1,024
1,960
3,016
ITTF new WR only counts best 8 results and ITTF old WR counts bonus points from every tournament results!
Give you an extreme example: under new system, the highest points that a player can get in 2018 are 6 Plat winner +World Cup winner + Grand Finals winner; winning some regular world tour does not add him/her more points and why attend? It is not the case in the old system, one can get bonus points attending all tournaments.

Again, as I said, there is no perfect ranking system. But as WR is important for eligibility and seeding, it should indicate something meaningful. I support the structure of old system - bonus points (more participation are awarded, but balanced once losing) + rating points (levels of the opponents and tournaments, in other words, the difficulty of a win are taken into consideration). It can be adjusted and improved. People are concerned that senior established players who do not play a lot a year keep high ranking (mostly from rating points that never expire) is not good for the sport and not fair to other players. It can be simply solved by removing rating points long time ago (say 10 years ago) or from matches against retired players over 5 years.

One major problem of the new WR system is that the ratio of winner points /loser points is too low, compared to the ratio of old WR bonus points/World Tour standing points (still in use)/prize money/tennis ranking. The supporters of new WR system kept saying it resembles tennis and ignored the fact loser in early round of tennis gets very few points and the ratio of winner/loser is much higher, not to mention that the prize money of tennis events is huge and tennis players do not need to play leagues for a living.

I have no problem accepting that the new WR is in use and will be used for qualification/seeding. By the end of the day, it is the players who suffer or thrive from it.

I think again that it boils down to that "gut feeling". You may want rankings that give as fair an impression as possible of all player's relative strenghths, I'd imagine, especially (as mentioned by OP) if Olympics and such seedings will be based on this list. In this case ITTF encourage more participation in tournaments (no bother if you lose, just come up and score as many points as possible), which is also somewhat defensible.

*edit: typo
 
This user has no status.
This user has no status.
Member
Jan 2017
120
176
423
Here you are.
This is the probability of upset function
.
ProbabilityOfUpset.jpg

the difference is about only 250 points. With the difference about 400 points, the Upset chance is almost zero.

For Ma Long, has only 13000 points, and Ovtcharov has 16000 points. You can apply on the Distribution curve yourself, and apply the PDF function to find the winning chance.

ITTF ranking is a very bad design.
No loosing-point, so everybody can abuse the system. Mass play, and don't care about the loss, because you can only win the point, but never loose your points.
 
This user has no status.
This user has no status.
Member
Jan 2017
120
176
423
And if you look at the points of Ovtcharov and Fang Bo. Ovtcharov on the new ITTF ranking has 16000 points, Fang Bo has only 9000 points. Ovtcharov has almost 2 times higher points than Fang Bo.

So apply on US table tennis rating for example, with the same ratio, for example, give Ovtcharov has 2200 points, and Fang Bo has only 1200 points equivalent (9000/16000*2200). Is it logic? Feel yourself!


So in the US, 2200 is the semi-pro players? and Fang Bo is a new beginner players! LOL!

current-ITTF-world-ranking-01-2018-500x351.jpg
 
Last edited:
This user has no status.
This user has no status.
Well-Known Member
Sep 2013
7,259
6,225
15,295
Read 3 reviews
What I don't understand is ITTF always had a world tour/pro tour standing
The same standing/ranking for that year that leads to the finals (world/pro tour finals)

So ITTF always had win/loss ranking and then they always had participation ranking
So now, both are participation ranking

What I think is the most silly is now due to the new method, the age groups are counted separately.
so in Juniors section - waiting to see how it will be (only release on 15 Jan)
I suspect your number 1 seed of Harmoto and Hirano will not be top 3 or 5 or maybe 10.
Simply they didn't play enough in the junior sections lol

And vice versa with Moregard ranking disappearing by 700 spots on the senior mens is also another disadvantage

I guess people only think about who is world number 1
but people don't think enough on how this change effects hundreds or not thousands of international particpants

Timo is right in saying Tennis pros don't play club and only play on the world tours
Try that will TT, the player will end up on the streets

edit
here is the WJC 2017 standings, so I guess ranking would therefore be similar (obviously your loosers will gain more than double the points with 2018 calculation) and WJTTC would be added
boys: https://d3mjm6zw6cr45s.cloudfront.net/2017/11/ittfjc_standings_boys.pdf_0.pdf
girls: https://d3mjm6zw6cr45s.cloudfront.net/2017/11/ittfjc_standings_girls.pdf_0.pdf
 
Last edited:
  • Like
Reactions: matzreenzi
This user has no status.
This user has no status.
Well-Known Member
Oct 2016
1,659
1,140
5,311
So what IS your age?

jesus christ the net never seizes to amaze me. I suggest we make a new sport with boosting/speed glue allowed, super spin pips allowed,same color rubbers allowed and we can call it "PING IAMNOTINDENIAL PONG"

p.s. Once it seemed really funny that NL wanted me to post a video of me playing to verify whatever he wanted to verify but now I totally get it. Im suggesting that it is mandatory for everyone posting here that his age will be visible
 
  • Like
Reactions: matzreenzi
says Shoo...nothing to see here. - zeio
says Shoo...nothing to see here. - zeio
Well-Known Member
Jan 2018
7,227
9,317
18,297
Here is rating system 101.


There're three common types of rating system - subjective, accumulative and adjustive. Examples for subjective rating system are combat sports like mixed martial arts and wrestling where experts judge the athletes. Accumulative and adjustive rating systems are considered objective since there is a set of rules that applies equally to all players. Adjustive rating goes up or down after each match based on the opponent and outcome, whereas accumulative rating goes up based on success during a set period.


Adjustive is the most sophisticated and hence the most complicated as it keeps track of the most information. Because of that, it is also the most accurate in predicting the outcome of a match. Accumulative pales in comparison, yet tennis, badminton and now table tennis have turned to it. The biggest reason is that accumulative encourages participation, which in turn enhances ticket sales. From a marketing standpoint, it is also easier for a laymen to follow due to its simplicity.


Tennis(ATP) first turned to purely accumulative in 1990(first best-14 results in a year, then best-18 in 2000, then a new points system in 2009), badminton followed in 2007(first best-10 results in a year, then best-12 with a new points system in 2018) and now table tennis(best-8 results in a year with a points system similar to tennis and badminton).


Now, the biggest mistake one can make when comparing two fundamentally different rating systems is simply converting points from one to the other, and vice versa. Apples and oranges. You're basically asking the question why they look and taste different.
 
Last edited:
This user has no status.
This user has no status.
Well-Known Member
Sep 2013
7,259
6,225
15,295
Read 3 reviews
Here is rating system 101.


There're three common types of rating system - subjective, accumulative and adjustive. Examples for subjective rating system are combat sports like mixed martial arts and wrestling where experts judge the athletes. Accumulative and adjustive rating systems are considered objective since there is a set of rules that applies equally to all players. Adjustive rating goes up or down after each match based on the opponent and result, whereas accumulative rating goes up based on success during a set period.


Adjustive is the most sophisticated and hence the most complicated as it keeps track of the most information. Because of that, it is also the most accurate in predicting the outcome of a match. Accumulative pales in comparison, yet tennis, badminton and now table tennis have turned to it. The biggest reason is that accumulative encourages participation, which in turn enhance ticket sales. From a marketing standpoint, it is also easier for a laymen to follow due to its simplicity.


Tennis(ATP) first turned to purely accumulative in 1990(first best-14 results in a year, then best-18 in 2000, then a new points system in 2009), badminton followed in 2007(first best-10 results in a year, then best-12 with a new points system in 2018) and now table tennis(best-8 results in a year with a points system similar to tennis and badminton).


Now, the biggest mistake one can make when comparing two fundamentally different rating systems is simply converting points from one to the other, and vice versa. Apples and oranges. You're basically asking the question why they look and taste different.

This doesn't explain why TT reward the looser 90% of the winner
SF gets 80% of the winner
QF gets 70% of the winner
ATP and badminton as well as TT own prize money doesn't use such high looser awarding ratio

If this was rectified to a more fair ratio, it would allow the ranking to be more fair and not inflated
 
This user has no status.
This user has no status.
Well-Known Member
Sep 2013
7,259
6,225
15,295
Read 3 reviews
This doesn't explain why TT reward the looser 90% of the winner
SF gets 80% of the winner
QF gets 70% of the winner
ATP and badminton as well as TT own prize money doesn't use such high looser awarding ratio

If this was rectified to a more fair ratio, it would allow the ranking to be more fair and not inflated

Infact let me reply to my own post haha
I know why ITTF made it 90%
because if they made it 60%, then your best players don't need to play 8 times a year
5 event win will be enough to be number 1

So there we go, if any one agrees that 90% is ridiculously high, then the only reason for it is ITTF is forcing (not encouraging) top players to play 8
imo, such methodology to get top players to max out is unfair. It gives winners not much incentive over loosers
 
Last edited:
This user has no status.
This user has no status.
Well-Known Member
Sep 2016
1,024
1,960
3,016
ITTF old WR does not put a limit to the bonus points one can get and World Tour standing points count all the tour events (the best 12). I wonder how ITTF comes up with the best 8 results not best 10 or more. Players play a lot of continental events but only one result is counted for WR. It is not fair, imo.

Infact let me reply to my own post haha
I know why ITTF made it 90%
because if they made it 60%, then your best players don't need to play 8 times a year
5 event win will be enough to be number 1

So there we go, if any one agrees that 90% is ridiculously high, then the only reason for it is ITTF is forcing (not encouraging) top players to play 8
imo, such methodology to get top players to max out is unfair. It gives winners not much incentive over loosers
 
This user has no status.
This user has no status.
Well-Known Member
Sep 2013
7,259
6,225
15,295
Read 3 reviews
ITTF old WR does not put a limit to the bonus points one can get and World Tour standing points count all the tour events (the best 12). I wonder how ITTF comes up with the best 8 results not best 10 or more. Players play a lot of continental events but only one result is counted for WR. It is not fair, imo.

yeah, there is too much inconsistency with the new change
I am okay with the new system in terms of accumulative, but with the following conditions:
1) looser points at 60% of winner
2) up to 10 events count
3) bonus scale for player participating in more events
4) junior points can be used in seniors ranking as seniors is known as "open", meaning all (ITTF points) should contribute to it
5) Prize money to increase 10 folds :)
6) ITTF to make a separated rating list so we have both
7) ITTF to explain all these "trophy" purpose, as I have been following TT all my life and I can't explain why we have so much trophies:
- world team cup?
- world cup (limited players??)
- WTTC
- WTTTC
- World Tour grand finals
I rather have WTTC and WTTTC, and then maybe 4 major tours in a year, so making it 4 major tours and 1 world championship every year and these 5 events having major action

I don't see FIFA with world championships, they only have world cup :)
 
Last edited:
says Shoo...nothing to see here. - zeio
says Shoo...nothing to see here. - zeio
Well-Known Member
Jan 2018
7,227
9,317
18,297
It was not the intention of my post to address the question on points system.


Now, you're comparing valencia and navel. Both are oranges, yet slightly different. The former for juice and the latter for pulp.


ATP has tweaked their points system multiple times. Started out with a lower ratio in the early 90s, then it got roughly 20% higher for a few years, with bonus points added for upsets, then it got slashed in the next few years again, and with bonus points taken away, then a further reduction for several years with the best-18 system, and yet another reduction for the past several years.


The ITTF have stated they will adjust it as feedback rolls in.


The bottom line is you're not giving it time to pan out. Cross your fingers it won't take them like the BWF to make an adjustment after 11 years.
 
Top