How to classify rating in TT

FruitLoop said:

You are correct but the Ratings list for TTE doesn't include local league results I think? Also Scottish and Welsh leagues are not yet incorporated into the system but I heard there is hope they will be soon.

Click to expand...

Yes, Ratings list for TTE does not include local league results. IMO, it is useful to have a ranking based solely on tournament performance.

As to Scottish/Welsh leagues - well, many already use tt365 - e.g. https://tabletennis365.com/Aberdeen or https://tabletennis365.com/Cardiff. Full list is at https://tabletennis365.com/Sites. Basically, the tt365 website is free to use for any UK league or club, all the club/league needs to do is contact tt365.

vvk1 said:

Yes, Ratings list for TTE does not include local league results. IMO, it is useful to have a ranking based solely on tournament performance.

As to Scottish/Welsh leagues - well, many already use tt365 - e.g. https://tabletennis365.com/Aberdeen or https://tabletennis365.com/Cardiff. Full list is at https://tabletennis365.com/Sites. Basically, the tt365 website is free to use for any UK league or club, all the club/league needs to do is contact tt365.

Click to expand...

What I meant was they are not yet incorporated into the ranking system despite using the site. I assume the players have ghost rankings though.

They all got Westgate Las Vegas ghost rider ratings for the sports betting world...

Sent from my SM-N950U using Tapatalk

FruitLoop said:

What I meant was they are not yet incorporated into the ranking system despite using the site. I assume the players have ghost rankings though.

Click to expand...

There you go - TT365 has just added a global ranking list:

https://www.tabletennis365.com/Ranking/

What a fantastic resource. It's a shame that 1. It took so long. 2. It probably won't last now and I have no faith in TTE being able to recreate anything as good. This finally gives England a rating system like France/Germany/USA that is meaningful on national and down to local level. Fantastic work.

ttpshot said:

USA's rating system is fantastic and I wish we have it in Japan. Are there any cons for having ELO rating system?

Click to expand...

The US rating system is full of fudge factors.
I would use Ratings Central.
https://www.ratingscentral.com/
Many countries already use RC.

In Belgium, we used this one: https://competitie.vttl.be/rankings/elo

in France we have an ELO like ranking system as well ranging from top players to beginner amateur players

What a great thread, so many different ranking systems being shared.
I have worked in Table Tennis ratings and rankings for the past few years. If there are any national associations or clubs which require a system developed by someone with experience, I'd be happy to help. Even if you think your organisation might benefit from a conversation with me, feel free to get in touch.

I think a division ranking system is the most practical.

I will admit that just about any ranking system is better than none.
Never let perfection be the enemy of the good.

However, Many years ago I had a small exchange with David Marcus, the inventor of ratings central. I would have done it differently but my approach was like how an engineer would make a ratings system. David Marcus's approach is how a PhD in statistics and probability would approach the problem. I understood what David Marcus trying to achieve.

Before you guys dismiss this lightly, I want point out the it was my analysis of Emrathich's desired to include the number of games one or lost as being important in how the ratings are changed. Basically I saved Emrathich a lot of time trying to do what no body wanted.

I do not have a PhD in statistics or probability. I do not need one. I can see the brilliance of the rating central system.

brokenball said:

I will admit that just about any ranking system is better than none.
Never let perfection be the enemy of the good.

However, Many years ago I had a small exchange with David Marcus, the inventor of ratings central. I would have done it differently but my approach was like how an engineer would make a ratings system. David Marcus's approach is how a PhD in statistics and probability would approach the problem. I understood what David Marcus trying to achieve.

Before you guys dismiss this lightly, I want point out the it was my analysis of Emrathich's desired to include the number of games one or lost as being important in how the ratings are changed. Basically I saved Emrathich a lot of time trying to do what no body wanted.

I do not have a PhD in statistics or probability. I do not need one. I can see the brilliance of the rating central system.

Click to expand...

That system was rejected by the USATT during a board meeting, go back to the archives there. It appeared to have flaws in the algorithm which is not easy to fix.

Haha, the current system is very flawed too, in many ways. Largest faults are point deflation over time, allowing seriously under-rated players in comps, failure to properly adjust ratings. Both the young and geezers watch their ratings more intensely than Korean Dram Shows...

Still, the system does what it is supposed to do: classify players for ranking.

If USATT could figure out how to incentivize geezers to get bonus rating points by some manner, there would be 1000% more turnout for sanctioned tourneys. That spells sum moar cash for the bigwigs to go on TDY. I am all for a top official making 5% of revenue... if they can double/triple/quadruple income, then it is worth it.

Back when the USATT rejected RC at a board meeting, the system used was considered better. RC had some big supporters too. There were examples given of the RC system problems.

The way Koreans differentiate levels of players is by looking at average scores and assigning handicaps under the 2+1 system.

1st level of difference gets a 2 point handicap to make the play levels equal.
each additional level of difference gets an additional one point handicap... up to a max of 7 points.

So if you defeat a player an average of 11-8, then there is about two levels of difference in that head to head matchup.

This forms the basis of the Korean amateur system.

In Korean clubs, you have divisions 1 through 7 at the club level. Div 7 is a newbie. Div 1 is the 2000-2200 typical Div 1 city male player.

In the city/regional and national system, you have division 1 through 5... with accommodations at the top level to go to div 0 or div MINUS 1 if the top elite are levels better than average Div 1 player. They do that for city/regional, and make a "Champions League" division for those who win Div 1 titles. Any level above that is considered "pro Athlete" division, which means a player had pro training as a kid in school system. Those pro trained athletes are not allowed to compete in the "regular" (Life Athletic) tourneys.

This system is formed with judgment by the clubs… it prolly would not work in USA, because we all do not think alike or look at the situation the same way in agreement. If USA got rid of a numerical rating system, there would be bloody mayhem and mutiny exceeding the scale of the LA riots of the 60s/70s/80s.

Der_Echte said:

If USA got rid of a numerical rating system, there would be bloody mayhem and mutiny exceeding the scale of the LA riots of the 60s/70s/80s.

Click to expand...

It would probably double the tournament participation. It has probably discouraged more players over the years than the current active number of players. It is inadvertently used to embarrass adult players to quit playing sanctioned tournaments, especially if draws are rigged sometimes.

....

Metaxa said:

In Belgium, we used this one: https://competitie.vttl.be/rankings/elo

Click to expand...

Takkyu_wa_inochi said:

in France we have an ELO like ranking system as well ranging from top players to beginner amateur players

Click to expand...

Some explanation of Elo-ranking can be found here: http://tabt.frenoy.net/index.php?l=EN&display=MethodePointsELO_EN

For an improved version of Elo, you can look up the Glicko (2) rating system. It helps to alleviate the issues which are due to small pool size and erratic results.

Loopadoop said:

That system was rejected by the USATT during a board meeting, go back to the archives there. It appeared to have flaws in the algorithm which is not easy to fix.

Click to expand...

What flaws?
The flaw I see is that the USATT wouldn't have control over the rating system.
I doubt that the USATT board understands the RC rating system well enough to articulate a flaw.

brokenball said:

What flaws?
The flaw I see is that the USATT wouldn't have control over the rating system.
I doubt that the USATT board understands the RC rating system well enough to articulate a flaw.

Click to expand...

It is a matter of public record. RC couldn't defend their system. The flaws were clearly shown by rating error examples of players. You are trying to defend a position without any real facts, pure speculation which is wrong.

....