cOMe gIVe Me 100 pOiNTz iN hYpEr.

Doerdich: That's only in the first 25 games and it's a false impression given by the calculation of interim provisional ratings.

The established rating after the 25th game is effectively the sum of the 25 opponent's ratings plus 8 for each win and -8 for each loss. These 8s are the result of averaging the +/- 400 over 25 games.

Rather than wait until the 25th match has finished, the calculation keeps up a moving average. This can produce big swings from game to game but the gains and losses aren't "real" because the rating, within the first 25 matches, is not real either.

This message may be illustrative.

@matarilevich yes, you are right, K can have a greater value. But usually it is 10 or 15. But compared with the BKR: Here it is possible to lose about 100 points in game, too much I think.

Doerdich:
The formula Fide for variation of rating is:

NR = OR + (SG - SW) * K

NR is New Rating
OR is Old Rating
SG is Score Gotten
SW is Score Waited
K is a Constant

K can have three values :
K=25 when the player hasn´t finished 20 games
K=15 when the player has finished 20 games and his rating is lower than 2400
K=10 when the rating of player is upper than 2400

Score Waited is calculated with a list of percentage, but you can calculate approximately in this way :
SW ~= 0.5 + [(RP - RO) \ 800]
RP = Rating Player
RO = Rating Opponent

The highest value of [RP - RO] is 350 and the lowest is -350. If the different is higher than 350 or lower than -350 then we have [RP - RO] = (+-)350

All this means that the highest variation can be:
Rating Player = 2000
Rating Opponent = 1500
Different above to 350
Score Waited ~= 0.5 + [(2000 - 1500) / 800] =
= 0.5 + [500 / 800] =
= 0.5 + [350 / 800] =
= 0.5 + [0.4375] =
= 0.9375
If the player rated 2000 loses the game then
New Rating = 2000 + ( 0 - 0.9375 ) * K =
= 2000 - 0.9375 * K
With K=10 => New Rating is 2000 - 9.375 = 1990.625
With K=15 => New Rating is 2000 - 14.0625 =
1985.9375
With K=25 => New Rating is 2000 - 23.4375 = 1976.5625

The highest variation is (+-)23.4375 for players with rating lower than 2400 and no more of 20 games and for strong players is only (+-)9.375

sUp

Hi, I developed a New Chess Variant, I want to suggest (I call it "Capture-Chess"): There are the same pieces and the same board like in normal chess, but additionaly it is allowed to capture own pieces. If someone captures his own king, he cannot be checkmated. But then he has to checkmate the other in a maximum of 50 moves (this is of course variable, but I took 50 in analogy to the draw-rule of 50 moves), otherwise he is lost. If someone has already captured his own king, the opponent may not capture his own king.

Doerdich: There is a special board about chess variants, you can post the details there.

To 2): Could you also consider my idea? I also found some interestng others in the internet. Of course it is not necessary to realise them immediately but I just want to show them to you.
To 3)Fide Rating is more "fair" in my view. Here, no strong players will play vs. weaker rated players, because they simply lose to much points. Some low rated players even don't play so bad because they are maybe new to this site.

Doerdich: 2) Yes. About 40 or more. 3) I don't know, I never used FIDE rating formula.

Well, Fencer, I read that, but it doesn't give much information. Seems that there is a problem with the creator or something like that?!
What about question 2. and 3. ? ^^

Doerdich: Please read this.

Doerdich: Just hours before you posted it, two new chess games have been added to the site - bringing the total of chess variations to 25.

New questions:
1. Why has Gothic Chess been removed?
2. Will there be new chess variations in the future? I developed one by myself...
3. I think the BKR calculation should be more like e.g. FIDE-ratings. Here you can lose to much points in one game I think. In FIDE_Rating, the maximum you can win in one game are 12.5 points, the maximum you can lose of course are also 12.5 points.

Andre Faria: Thank you, now it is.

Fencer, the link for Los Alamos Chess is not the right one...

Hrqls: old BKR = 1859, new BKR = 1895 (+36)
alanback: old BKR = 2001, new BKR = 1980 (-21)
I have 390 counted games in Nackgammon -- still getting adjustments as high as 21 points

I still maintain the dice roller should be destroyed painfully

Obviously it is the method for establishing an initial rating that is the most absurd in backgammon. It probably is fine for games of pure skill, but makes no sense at all in backgammon where the worst player can beat the best player by pure luck. Even in a game of skill, it could be distortive if a highly rated player happened to lose for some reason other than being outplayed, e.g. by timing out. But in backgammon, distortion is almost guaranteed in a signficant percentage of cases.

I don´t know much about how works ratings here but it´s not very similiar to the system of the federation international of chess (Fide) based on the method of the mathematician Arpad Elo.

I think the rating initial (provisional but initial) in brainking is established after of the fist game finished using the rules of the system of US chess federation [http://www.sizes.com/sports/chess_ratings.htm]



This system works in this way
rating initial = rating opponent + {400(win), 0(draw), -400(lose)}

However Fide system is different :
rating initial = rating opponent + {12.5(win), 0(draw), -677(lose)}

The variation of rating in the second and next games works different too.

This system applied here favours to the players win his first game. If you win the first game against an opponent of 1900 then you will have 2300 as an initial rating and you can keep this rating playing against weak players

Pythagoras: It's a rest of a membership which was cancelled on player's request. A rare situation.

Yep, it can't be a discussion about your priorities on this site....Only wishes (not suggestions because that would mean automatically a discussion about your priorities) for things should be done.......

A question: In this tournament there is a prize of 5 months membership. How the creator managed to offer a prize that is not 6 or 12 months? I mean what is the price for 5 months he paid.........?

All right, all opinions are noted and considered. There is no point to continue in a discussion about my priorities, it would only lead to a new argument.

What i see, with all these discussions about BKR, i think you should rethink your priorities about new features and spend the most time in:
1)Improving the BKR-system of Backgammon and of other games......
2)Fixing the bug in Backgammon.....

This is my humble opinion and it's up to you of cource to decide, but by doing this you will satisfy a nice number of members......

pgt: Your opinion is noted.

Fencer: With the greatest respect, and judging from the excellent input from so many wise backgammon players who have made contributions to this discussion, it would be a HUGE pity if Maxxina was indeed right. Such a flawed rating system for backgammon is not likely to endear the backgammon population to this rating system.

Somebody recently spoke of "contempt" for backgammon players, and I can only agree. There has been a huge delay in implementing the long-promised "pro-backgammon", and no action at all on the bug (play both dice if possible, and if impossible, play the higher die) which has been acknowledged for well over two years, and which has been the subject of numerous acrimonious discussions for a long time. This latest lack of acknowledgement of the deficiencies of the rating system for backgammon is almost the last straw.

I hear all the accolades and statements about "how busy Fencer is" and "what a great site" and I concur - but you are running a business, and running a business means KEEPING THE CUSTOMERS HAPPY.

Now I am one of those customers, and I joined BK as a paying member in the hope that my support would help to get some of the deficiencies corrected. I have waited patiently for that to happen for about two years, and there has been no action. I have been perfectly content to support BK and your business in the belief that there would be some action in addressing the deficiencies in the implementation of my favourite game, but I am rapidly losing confidence that anything will actually change.

My feeling at the moment is that unless I can see some concrete evidence of some REAL progress towards implementing some of these much-promised enhancements to the backgammon experience. I will not be renewing my subscription when it falls due. I imagine there will be many more backgammon
players who feel the same as I do. I trust that you will have sufficient non-backgammon players to keep the business viable.

Jason: Maxxina is right.

Jason - What can i tell and see on czech board . YEAA THIS IS A FINAL VERSION !!

Have you done messing with the bkrs once and for all now ? everytime you play with them, i get further down the scale . to me it looks like all games i have played and won a few points from have now taken points away lol.
It just seems everytime i get to a position that im happy with you play with bkrs and i lose 60+ positions

It doesn't seem to be possible to create a tournament of multiple-point matches. That is, one in which each player would play the other a 3-point match, for example. Am I missing something?

playBunny: The 100 points was what A and B agreed the game was worth ... if they had agreed 200 points then the 3 results would have been:
A wins - A goes to 2050, B goes to 1450.
Draw - A goes to 1950, B goes to 1550.
B wins - A goes to 1850, B goes to 1650.

The purpose is to make it worthwhile for the better rated player to actually play the game. I have games now (particularly in Tank Battle) where I will either stay the same (if I win) or go down if I draw or lose. Hardly an incentive to play as there will always be a game or two you will draw or make a silly mistake and draw. As you say, possibly not sufficiently independant to be a rating system.

AbigailII: I understand what you are saying about being unratable without a draw or loss (likewise someone who only ever losses is also unratable) ... no doubt it is to do with trying to divide by zero. But it is not that they exceed the rating system, it is that they are outside it! Maybe they should not be even given a provisional BKR until they do have a contrary result or have reached the stage where they have completed 25 games and can get an established rating. I know of only one or two players here who are so exceptional in their particular game types that their would warrant such a high rating from scratch.

So here is another suggestion ... A player's rating may never be more than say 10% higher than the highest player they have beaten or equal with the highest player they have drawn with. This would encourage them to play the higher rated players and they would not be able to obtain unmerited BKR levels by playing moderate players.

WQ:)

I'm not sure what function the 100 has in the following. It doesn't seem to affect the resulting ratings.
They agree the game is worth 100 points
If A wins - A gets 100 (+25 now 2025) B gets 0 (-25 now 1475).
If a draw - A gets 50 (-25 now 1975) B gets 50 (+25 now 1525).
If A loses - A gets 0 (-75 now 1925) B gets 100 (+75 now 1575).


But that aside, it's an interesting idea.

I'd start, though, by saying that it's not a rating system. The purpose of ratings is to be able to compare players (as accurately as the model will allow). The model should be self-consistent, ie. there should be no player input required or allowed.

By having players agree to varying amount that they "put up on offer", (or bet? lol) - is to have a currency system where the "rating" is more like a purse or wallet. That's actually an attractive idea which could work in parallel with real ratings. At TrueMoneygames you can play Backgammon for money. But they also have the concept of play money for those who like safe betting. I was watching a player the other day who had amassed an amazing $1,000,000 of this play money. (Given that games are $100 that's quite an achievement - very aggressive doubling seems to have been the key. He won 3200 in a single game because of this. But, lol, I digress..) The point is that risking a chosen amount of your points is fun but cannot be part of a rating system.

I agree with you that a smaller K factor for tournaments could help encourage higher rated players to join. I haven't given much it study in Chess because I'm a Backgammon man. I do believe that the best way to encourage Backgammon players to risk their rating in an open-to-all tournament is to use the proper formula, one which is fair to all players.

WhisperzQ: That depends on the results. In an ELO system, the true rating of someone how has won all games can't be determined - any rating will be an estimate that's too low. Ratings are not absolute, and they are certainly not points you collect over time by winning. Difference in rating gives an expectation of the outcome of a game (or series of games). A rating of 2600 means that it's expected someone wins 50% of the games against someone else of rating 2600 - and N% of the games against someone of rating 1200, of some value of N strictly less than 100. If someone has won all his games, even if there are only 4 games, he has exceeded all expectations of the rating system - and 2600 will even be too low.
<p>
Once again, there are reasons why there are provisional ratings, and why there are established ratings.

WhisperzQ: Maybe if the games were 4 wins against the top 4 players..but those would not be ordinary games. LOL.

AbigailII: Renaming the consequence does not solve the problem. Surely you do not think that a 2600 rating (by any name) after 4 ordinary games is right!

playBunny: While you were writing your latest note, I was also putting together my thoughts (interrupted by a few phones calls). I think the K factor is also a good idea, then a tournament, for instance could be set up to have K factors built in to encourage higher rated players to play and put less of their points at stake against lower rated players. Maybe a combination of both ideas?

Or another idea is that players could agree the number of points they will "put up on offer" (there would need to be a max.) at the beginning of the game and the respective splits they might get out a win draw and lose. An example might help:
A is rated at 2000.
B is rated at 1500.
The (historically) expected result from 100 games might be (for A) 70 wins/10 draws/20 losses.
They agree the game is worth 100 points (actually probably too much but lets move on).
A risks 70 + (1/2 of 10 = 5) 75 points.
B risks 20 + (1/2 of 10 = 5) 25 points.
If A wins - A gets 100 (+25 now 2025) B gets 0 (-25 now 1475).
If a draw - A gets 50 (-25 now 1975) B gets 50 (+25 now 1525).
If A loses - A gets 0 (-75 now 1925) B gets 100 (+75 now 1575).

WhisperzQ

WhisperzQ: But to be there in 4 (and sometimes way beyond) is, I believe, ridiculus.

That's why there's Established BKR. Perhaps if 'BKR' was named 'provisional BKR' and 'Established BKR' just 'BKR', people would stop having sleepless nights of someones BKR after four games.

BuilderQ: If the top players have only played say 50 or 100 games and they have say a 75% win rate then someone at a 90% win rate might only need 30-60 games. But to be there in 4 (and sometimes way beyond) is, I believe, ridiculus.

WhisperzQ: 200-300 games can take a long time to play in a turn-based environment. Many game types don't have any players who haved played that many games (eg Jungle, Fast Espionage, One Way Checkers).

Grim Reaper: Sumerian has only 29 counted games because three losses are by time out in one or two moves, so his score is +28=0-1.
The reason of Sumerian is the top of Gothic is the same by you are the top (very) provisional of chess : few counted games can be enough for get the highest rating.
Maths are beautiful!!

playBunny (and others): The problems with the BKR system are not limited to Backgammon ... that someone can play 4 games and have a rating of over 2600 is ridiculus ... even if they play against top opponents (which in a number of types of games they haven't).

A while back you gave a formula which I think would work excellently ... takes into account the relative abilities of both players and requires someone to work their way to the top if they are good enough. If it takes 200 or 300 games to get near the top then so be it, at leat then it is a true reflection on concerted effort, not an anomily in the system.

I play here for fun but I also play to have interesting hard competative games, ... I am, indeed, competative perosn by nature. I do not mind losing where I am soundly beaten, but I also like to have my abilities (or lack thereof) recognised and the BKR is one way to do that.

I think the current system stinks and has me, for one, now very disallusioned.

WhisperzQ

alanback: I'm wordy enough, so I only said 16 to keep things simple. ;-)

In the US Chess formula (Established rating section) there are bands defining how much a game is worth:

There is also a K factor that determines the number of rating points that can change hands as the result of a single match, and that depends to some extent on the player's rating: 32 for ratings from 0–2099, 24 for 2100–2399, and 16 for 2400 and up. (There are also so-called ½K and even ¼K events where the number of points that can change hands is reduced as the fractions suggest, that is, 16, 12 and 8, and 8, 6 and 4 respectively.)

I'd guess that different bands are in use here, but does that tie in with what you're seeing?

alanback: It could be 16 pts because you have played so many games, as I haven't played as many games say in nack as in backgammon the pts difference is in excess of 30

playBunny:

>>The new system is that the 16 points are now apportioned according to the rating difference.

I'm seeing adjustments in excess of 16 points in some games -- why should that be?

First, resigning on the first move doesn't count as a loss for rating purposes.

Second, besides the number of games that have been played, another factor is the ratings of those who were played... especially for the first few games. This has worked in my favour in Anti Line4. :)

from a player who can't post here because he's been banned:

"I was looking at the Gothic Chess ratings today after the recalculation was done.

We have a player with 28 wins and 4 losses being rated higher than a player with 157 wins, 3 draws, and 0 losses.

I won 129 more games than Sumerian, without losing, yet I am rated 32 points below him.

How is this possible?"

Pythagoras: or, if you meant "1500" things, here in the US we use a comma. 1,500.

but that's here in the US, where we do all *kinds* of backwards things.

Yeah, it's a bad habit of me putting a dot "." on integer numbers to make them more easy to read.......The correct rule is to put a space or nothing at all.......

Pythagoras: 1½ things a day, wow he gets a lot done. Aside from getting out of bed I rarely get 1 more thing done. :)

Yep, no one should complain to Fencer about delaying of taking actions. He has to do around 1.500 things per day. And that's only the Brainking-related one's......

Let us not forget how much we get from Fencer for little or no investment on our part!