What would top players Usatt rating be?

[PoppaChubby]

There is no way that properly rated 1800 players beat properly rated 2000 players 15% of the time - this is simply wrong.

It's like a 1300 beating a 1500 in french regionale leagues, PR, R3 or R2... it happens almost each week

 

[Tony's Table Tennis]

It's like a 1300 beating a 1500 in french regionale leagues, PR, R3 or R2... it happens almost each week

I think the main problem here is tb doesn't play any tournaments (where rating gains/drops counts).
so he has a lot of practice matches and thinks winning up is normal - because he can do it, everyone else can.

if winning up is normal, then why does a 2950 Dima struggle with a 3100 Ma Long?

Game (Set) scores:
19 vs 76

 

[PoppaChubby]

This is what Qwen says:

If a player has a ELO rating 200 points below his opponent, what is his mathematical probability of winning a match?

Qwen3-235B-A22B10:24 pm
To determine the mathematical probability of winning a match based on the ELO rating difference , we use the formula from the ELO rating system , which predicts the expected score (probability of winning) for each player.

ELO Probability Formula:​

For a player A with rating RA , facing an opponent B with rating RB , the expected score (probability of winning) for player A is:
EA=1+10(RB−RA)/4001

---

In Your Case:​

• Player's rating: RA
• Opponent's rating: RA+200

So the rating difference is −200 points for the lower-rated player.

Plug into the formula:
E=1+10(200)/4001=1+100.51

Now calculate:
100.5≈3.162
E=1+3.1621=4.1621≈0.24

---

Final Answer:​

The weaker player (200 points lower) has approximately a 24% chance of winning the match.

My coach always told me 1 chance over 4, and then he described all the 4 chances as the 4 crucial moments in a chess game:
- openings, with whites, mostly how much you've read books and historical games, thinking deeper here doe not matter that much, it's mainly about rote learning that's why many players can play blind games.
- defenses, with blacks: are you confortable accepting a gambit ? playing a scottish or italian opening like the textbook and being able to know the disrupting 2 or 3 moves ? same here, reading a lot, rote learning, and a bit of creativity to disrupt the whites
- middle of the games: you have to think deeper now, it's mostly about experience also
- finals, when there are less pieces on the board and now it's the harder part of the game, only single wrong move can decide from a draw, a win or a loss. It's like doing multiball play when you train: you have to master all the variations from the easiest ones (King+ pawn vs King) to the trickiest (mostly knights and bishops finals, symetric or asymetric: knights vs bishops, bishop white squares vs bishop black squares, etc... with variations in starting positions).

 

[TensorBackhand]

I think the main problem here is tb doesn't play any tournaments (where rating gains/drops counts).
so he has a lot of practice matches and thinks winning up is normal - because he can do it, everyone else can.

if winning up is normal, then why does a 2950 Dima struggle with a 3100 Ma Long?

View attachment 36664

Game (Set) scores:
19 vs 76

See, i actually appreciate this. Because it is an actual proven data. So i agree with you that 200 pts is a big difference. So youre right ML has a stranglehold over Dimitri. That is a very good data point.

However, on the flip side, i can think of several other 3000+ players losing to much lower players.

WCQ lost to 14 year old.

Harimoto lost to WR50 players.

LGY has many bad losses.

SYS lost to wr 100? Indian players.

Again, 2% is a really small number. Yes ML has 20 straight wins. 50 is still far away

 

[TensorBackhand]

It's like a 1300 beating a 1500 in french regionale leagues, PR, R3 or R2... it happens almost each week

Yes, i imagine this happens much more than 2%

 

[Tony's Table Tennis]

See, i actually appreciate this. Because it is an actual proven data. So i agree with you that 200 pts is a big difference. So youre right ML has a stranglehold over Dimitri. That is a very good data point.

However, on the flip side, i can think of several other 3000+ players losing to much lower players.

WCQ lost to 14 year old.

Harimoto lost to WR50 players.

LGY has many bad losses.

SYS lost to wr 100? Indian players.

Again, 2% is a really small number. Yes ML has 20 straight wins. 50 is still far away

the really appreciation comes when you play in a ranking tournament and test it out for yourself.

 

[Tony's Table Tennis]

Yes, i imagine this happens much more than 2%

I didn't quite follow every post
your whole fancy calculation has it at 24%
then it became 15%
is it now at 2%?

 

[TensorBackhand]

I didn't quite follow every post
your whole fancy calculation has it at 24%
then it became 15%
is it now at 2%?

No. Another user suggested that the 200 pt deficit would suggest a 2% win rate for the lower player. The higher ranked player could win 50 matches in a row, is his suggestion.

My personal estimate is 10-15% win probability.

GPT suggested a 20% win probability.

Qwen suggested a 9% win probability.

I have not found a definitive source to estimate the win rate, so I just listed a few data points.

 

[Tony's Table Tennis]

No. Another user suggested that the 200 pt deficit would suggest a 2% win rate for the lower player. The higher ranked player could win 50 matches in a row, is his suggestion.

My personal estimate is 10-15% win probability.

GPT suggested a 20% win probability.

Qwen suggested a 9% win probability.

I have not found a definitive source to estimate the win rate, so I just listed a few data points.

i think the win rate is 0%
the loosing rate is probably the numbers you have there.

that means, 2800 can't win
3000 can loose.

hope that gives you a better understanding on what actually happens

 

[TensorBackhand]

i think the win rate is 0%
the loosing rate is probably the numbers you have there.

that means, 2800 can't win
3000 can loose.

hope that gives you a better understanding on what actually happens

sure, I agree with your meaning.

But that's why probabilities exist. The higher player can't play to his "normal" level every match. He chokes, he has bad matches, there is randomness, etc. So the real statistic answer is higher than 0%.

 

[zeio]

In the USATT system, a 250 points advantage means you should at least beat the other player 50 times in a row before they win once. Upsets are far less common the higher the rating as the playing is more stable. Would Ma Long or FZD beat the 2800 rated players 50 times in a row if they were motivated to? I would say absolutely. So their minimum rating would be 3050. Probably higher in reality.

According to David Marcus, "the numbers in the rating chart were basically just made up by the people who developed the current system."

 

[zeio]

3000 is about right for the cream of the crop.

For reference, ITTF rating and Ratings Central for select top CNT players:
https://www.tabletennisdaily.com/fo...st-peak-elo-rating-all-time.35800/post-492031

David Marcus wrote in 2000 that world class players would be around the 3000 level.

https://www.ratingscentral.com/Doc/NewTTRS.pdf

The law N(1400, 450 2) is assigned to players entering their first tournament and is called the
unrated prior. Many players playing in their first tournament have been playing with their friends
in their basement. These players often turn out to be at the 800–1200 level. Young children can be
much lower. At the other extreme are world class players from abroad (e.g., the world champion)
who come to the USA to play in the US Open. These players may be at the 3000 level.

 

[piligrim]

According to David Marcus, "the numbers in the rating chart were basically just made up by the people who developed the current system."

Who is Dadid Marcus?

 

[boretennis]

this guy i assume?

https://www.davidmarcus.com/TableTennis.php

https://www.ratingscentral.com/

 

[Dr Evil]

The Elo formula implies that expected win rate is purely a function of the rating difference, regardless of absolute rating, but in practice this is clearly not case. In tournaments it's common for 1200 to beat 1400, but a surprise when 2200 beats 2400. For obvious reasons there's more variance in the match to match performance level of lower rated players, and greater variance compresses win-loss probabilities towards 50/50. It works the other way, too, so that at higher levels the win-loss probability curve becomes steeper and even small ratings differences can imply big differences in skill level. Pure Elo ratings don't capture this at all (as well as a number of other factors), and so can be quite misleading if you take them as absolute skill level measurements.

 

[TensorBackhand]

The Elo formula implies that expected win rate is purely a function of the rating difference, regardless of absolute rating, but in practice this is clearly not case. In tournaments it's common for 1200 to beat 1400, but a surprise when 2200 beats 2400. For obvious reasons there's more variance in the match to match performance level of lower rated players, and greater variance compresses win-loss probabilities towards 50/50. It works the other way, too, so that at higher levels the win-loss probability curve becomes steeper and even small ratings differences can imply big differences in skill level. Pure Elo ratings don't capture this at all (as well as a number of other factors), and so can be quite misleading if you take them as absolute skill level measurements.

Makes sense. So what would the win probability be for a 1600 player vs a 2200 player with a 200pt deficit

 

[Dr Evil]

Makes sense. So what would the win probability be for a 1600 player vs a 2200 player with a 200pt deficit

Maybe a typo? In practice you almost never see a genuine +600 win. When a 1600 beats a 2200 it's usually only the slight coincidence of an underrated junior playing an out of practice senior. This is yet another limitation of pure Elo ratings; limited ability (beyond adding a volatility component to the formula) to reflect changing skill levels.

 

[TensorBackhand]

Maybe a typo? In practice you almost never see a genuine +600 win. When a 1600 beats a 2200 it's usually only the slight coincidence of an underrated junior playing an out of practice senior. This is yet another limitation of pure Elo ratings; limited ability (beyond adding a volatility component to the formula) to reflect changing skill levels.

Sorry, what I meant to say is:

What is the win % for a 1600 with a 200 deficit (1800 opponent)?

What is the win % for a 2200 with a 200 deficit (2400 opponent)?

 

[McSmasher]

Maybe a typo? In practice you almost never see a genuine +600 win. When a 1600 beats a 2200 it's usually only the slight coincidence of an underrated junior playing an out of practice senior. This is yet another limitation of pure Elo ratings; limited ability (beyond adding a volatility component to the formula) to reflect changing skill levels.

This is what I like about the Ratings Central ratings. It has some consideration to this volatility (measured as a standard deviation for those maths gurus) such as:

• Juniors / quick improvers that are consistently beating players with a higher ranking
• Players who don't often play leagues / tournaments vs those that play leagues multiple times a week

If a player is consistently beating others with a higher ranking, or don't play leagues often, then their standard deviation will be higher. For instance, if the standard deviation is 50, and the player's rating is 1050, then there is a 68% change that their ranking is between 1000 and 1100, and a 95% chance of being between 950 and 1150. In practice, those with a smaller standard deviation have a more accurate rating than those with a higher standard deviation.

Also, players that win/lose, if they have a higher standard deviation, their ratings points will go up or down further than those that play league multiple times a week.

Ratings Central have an explicit probability of an upset, based on the rating (playing strength). For instance, a player ranked 200 points lower has roughly a 5% chance of winning. Obviously the volatility mentioned above will make this much less perfect than this function portrays:

Some may argue that this is what happens when you let mathematicians run the show lol, but at least it attempts to factor in the volatility.

More info for those that are curious:

How the Rating System Works

How the rating system works.

www.ratingscentral.com

 

[McSmasher]

Sorry, what I meant to say is:

What is the win % for a 1600 with a 200 deficit (1800 opponent)?

What is the win % for a 2200 with a 200 deficit (2400 opponent)?

Theoretically, the win % would be similar... however, as Dr Evil mentioned, there's no factoring in for volatility... besides the examples he gave, there's also the case where lower ranked players have more obvious weaknesses that mean some matchups don't suit them (e.g. not handling pimples or slow spinny loops). Whereas higher ranked players are more robust to different styles, so you'll see less upsets.

Also, the USATT ratings system is about as simple as you can get, getting an explicit win % is futile without considering volatility. You could infer the win % based on the USATT formula, but it's extremely crude. For instance, if the difference in ratings is between 188 and 212, if the expected winner wins, then the players exchange 1 point, and if they lose, then 40 points is exchanged. You could do 1/40 = 2.5% as the winning percentage, but really, this is definitely not reliable because of the volatility.

So to answer your question, 2.5% for both scenarios, with between 0-100% accuracy 🤣

More info on the USATT ratings system:

https://usatt.simplycompete.com/info/ratings