ELO inflation ha ha ha

[Nay Lin Tun]

In general elo is the winning percentage , 400 elo difference means better player A will get around 9 win and 1 loss vs weaker player B. 800 elo difference means better player A will win around 99 games and 1 loss vs weaker player B or 50 games win and 1 draw vs B( The same scenario of SF vs Delphi). If Sf win 100/100 games vs delphi without draw or loss, then Sf elo will be infinity and Delphi elo =0. In conclusion you can see 2% chances of this result. Second condition, SF is black. Average score of all human games database from 2015 showed white scores 55 percentage that means 35+ elo bonus for white. Your friend's assumption 2600 elo player is harder to beat 2200 vs 1600 and 1200 comparison is partially correct because of FIDE elo adjustments of K factors over 2400 rated player( In term of general elo , it is wrong, same winning percentage for same elo difference)
For reference
http://www.3dkingdoms.com/chess/elo.htm
http://www.fide.com/component/content/a ... e-k-factor
https://en.wikipedia.org/wiki/First-mov ... e_in_chess

[Henk]

I assume TCEC keeps K-factor of ELO rating constant. So one should give ELO rating plus K-factor. Otherwise information is incomplete.

[Nay Lin Tun]

1200 is average elo of United state chess players(1157 or something I could not remember) . So we can assume average elo of human chess players(720 millions' average elo) in the world is around 1200, though average elo of some countries may be higher cos of chess schools etc. 1600 elo is average elo of club players. On online chess rating chess24, chess cube start with 1600 elo while chess.com start with 1200 elo. So chess.com 1200 basic elo is realistic while other sites just overrate elo. If you register and play in both sites your elo will be inflated 300-400 in those websites rather than chess.com

[Henk]

My chess.com blitz rating is 140 points higher than my normal (no blitz) Dutch KNSB rating. So there really are differences. Otherwise I better play very fast in long time control games.

[Nay Lin Tun]

Higher Blitz rating than your standard rating is common for > 1200 elo players cos you play more games in blitz vs standard and you have more winning chances than standard games vs average 1200 players. On the reverse side, weaker player <1200 usually have more chance losing by playing more games in blitz games than standard games and their blitz elo is generally worse than their standard elo. (We have to assume, no play , no different in elo and 1200 elo for everybody). However some people may have worse blitz rate and good standard rate or vice versa depending on his/her playing style.

[Henk]

My KNSB rating is only updated after a year or so and I play far less of these KNSB games. But I only expect an increase of 40 ELO.

[Henk]

For KNSB rating if ELO is less than 2100 and old rating is based on less than N = 75 games then K = 216/SQRT(N). If more than 75 games were played K = 25.

If ELO > 2400 and N > 75 then K = 10

Otherwise N > 75 K = 25 – (R – 2100) / 20 where R is old ELO rating

So K is between 10 and 25 if R calculated from more than 75 games.

[Ajedrecista]

Hello Henk:

Someone told me that the chance of a draw from 2200 against 2600 player would be higher than a draw from 1800 against 2200 player. So that is nonsense.

By the way how do you convert ELO rating back into chance.

I guess that you want to say 'Elo difference into expected score'. The classical, well-known Elo formula is:

Code: Select all

(Elo difference) = (own rating) - (opponent's rating)
(Expected score) = 1/{1 + 10^[-(Elo difference)/400]} = (win ratio) + 0.5*(draw ratio)

(Win ratio) + (draw ratio) + (lose ratio) = 1

It is useful with thousands of games. You can not claim anything with 10 games or so.
There are error bars that are proportional to (games)^(-0.5).
There is much information about error bars in this forum.

Kai said something like 'the expected score is the same in a match of 3800 vs. 3000 or 1800 vs. 1000' but the draw ratio would be very different, of course.

Kai stated a 700-Elo difference and a 3% of draws IIRC. Just using the Elo formula:

Code: Select all

1/{1 + 10^[-(700)/400]}  ~ 0.9825 = 98.25% (for the stronger engine).
1/{1 + 10^[-(-700)/400]} ~ 0.0175 =  1.75% (for the weaker engine).

If the weaker engine can not win a single game after lots of games (win ratio = 0):
(Draw ratio) = [(expected score) - (win ratio)]/0.5 = (0.0175 - 0)/0.5 = 0.035 = 3.5%

Given a fixed Elo gap, the draw ratio is expected to be higher when the average rating is higher, that is, less blunders are expected; however, the expected score should be the same. You have an example here:

Draw Rate

I hope this info could be useful to you.

Regards from Spain.

Ajedrecista.

[Norm Pollock]

Elos are based on the past performance, which predicts, on average, future performance. When there is a result that greatly contradicts Elo expectations, like this game v Delphil, then it gives the SF team a heads-up to a deficiency. Perhaps they will find the cause and make the correction. Without these "hiccups", engine development will be much slower.

[Adam Hair]

Isn't the further away from average the less impact an ELO difference has. I guess ELO 1600 is average strength of a player or is it more like 1200.

I don't understand the question. With ELO, only the ELO difference counts, and the same ELO difference has the same winning percentage. So, if the ELO model is correct, then the score of 3800 against 3000 ELO is the same as 1800 versus 1000 ELO.

Someone told me that the chance of a draw from 2200 against 2600 player would be higher than a draw from 1800 against 2200 player. So that is nonsense.

By the way how do you convert ELO rating back into chance.

Kai's statement is correct. But the Elo model does not account for higher draw rates in matches between higher quality opponents.

Expected score of player A = 1/(1+10^((Rb-Ra)/400)) where Ra and Rb are the Elo ratings of players A and B.