pgt: I have a different solution to the ratings problem: play 3-wins matches. It would be good if we could hold tournaments of such matches.

Andersp: Well we have two of the four problems resolved, but until there is a more realistic rating system, people who chase ratings will clsuter at the top and play each other rather than risk losing a lot of points in a single game. But once this is fixed and the cube is implemented, ther will be a lot of very happy people, I know!

Andersp: Ah, I don't think that issue will go away!

redsales: I will miss the sportmanship discussion too, especially from those who dont play too low rated players fearing they should lose rating

thanks filip, working flawlessly now :)

i'll miss the sportsmanship discussion...but apparently, it is still an issue with bid fixing in ponds.

Fencer: Thankyou for this enhancement. I look forward to seeing a game in which it applies. This will save a lot of unnecessary discussion about poor sportsmanship and the like, I'm sure.

I've been known to introduce a new bug without fixing the old one ;-)

WhiteTower: You can classify me as one of those programmers.

I don't think I have ever introduced a new bug while fixing an old one. (Usually I introduce MANY new bugs while fixing an old one...)

Fabien Letouzey! Perhaps it surpasses the "King" Richard Lang.......

alanback: He sure is - who knows ONE perfect programmer who doesn't introduce a new bug fixing a old one? ;)

Fencer: Hats off to you, you must be quite a programmer!

Should be fixed.

Mine work as well. Thank you.

Fencer: ah! i checked for you but didnt see you online .. i guess you heard all our souls cry for help ;)

Whoah, real-time troubleshooting ;-)

Wait a minute.

Hrqls: i posted the bug in the bug tracker

i have another game with the same bug : http://brainking.com/nl/ShowGame?g=1084960

i want to move the 4 first, so i can bear off with the 6

alanback: I mean, he obviously worked in the shadows while we were all jeering at him, and now his surprise has come out slightly faulty :) SURELY his pride will not allow this issue to remain alive for much longer ;)

WhiteTower: I'm sure you're right :-)

UzzyLady: I believe Fencer will fix this relatively trivial bug [pointing and shaking index finger at Fencer!] as soon as he is made aware of it :)

I'm not sure which big one you are referring to, but I can't say the present situation is an improvement if the reports below are accurate.

alanback: I'll gladly suffer through the fixing of a few bugs, if it means we get the big one taken care of!

Not uncommon, but it needs to be fixed!

http://brainking.com/en/ShowGame?g=1090894&i=18

I wanted to swap the dice so I could take the 4 first, but there was no swap dice button.

http://brainking.com/en/ShowGame?g=1084875 and
http://brainking.com/en/ShowGame?g=1084875 are examples of the bearing off problems. In both cases I want to take the smaller die first so that I can start bearing off.

And of course, now that I'm looking for it, I can't find the other game I had a problem with. I'll keep searching.

UzzyLady: If you see them, post about it here - a link to the game, and what you want to do...... That way Fencer can see it and fix it.

UzzyLady: If it is as you describe it is definitely a bug.

BIG BAD WOLF: I've run into several situations where it won't let me use the smaller die first. These aren't games where I will only use one die, but want to move the smaller number first, to set up the larger number. In one case it didn't give me the option of swapping my dice. Is this a bug or have the rules changed?

Marfitalu: Looks like the system currently does not see that you can bear off chips soon, and in the process - trying to make you use both your dice.

Looks like Fencer will have to add the code of "if using the smaller dice leads to a point where you can start to bear pieces off the board, then allow it - otherwise ...."

AbigailII: It's in effect immediately even for running games. It would be too complicated to separate them.

Fencer: Ah, good. I presume that the new rule is in effect immediately even for running games? Or do running games keep using the old rules, which didn't enforce maximum die usage?

alanback: Right, I changed/fixed the model.

Fencer: Thank you, Fencer. I interpret this to say you have changed the software, not just the published rules?

I've just added these two points to the rules:

• If it is possible, both dice must be used. It means that some pieces can become "frozen" in certain positions because making a move with these pieces would create a situation where the second dice couldn't be used.
• If only one die can be used, the one with the higher number must be chosen.

Please let me know if it really works. I hope I've fixed this issue at last but I could easily overlook some special cases :-)

What is that "cube" and "bug" all about?

jolat: Indeed, there's no cube on this site. The reason is that it isn't implemented - a Fencer decision. ;-) At some, yet unknown, time in the future, a backgammon version with a cube will be implemented.

Some put a line on top of this page with the message that cubes will be implemented in the future.

I am new on this site and I would like to play backgammon but it seems to me that it is not possible to use the cube.
Is this well that?
If so, do you know why this possibility does not exist here?

I've no idea, but I hope so! Sounds like a very sensible tourney to me. ;-)

Brainking board's perhaps best for this as someone might have done it for a different game type.

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?

And a good calculator once you get past airplane-cockpit syndrome, lol.

Javascript calaculation:
function afRatingChanges (fRatingP1, fRatingP2, uiExperienceP1, uiExperienceP2, uiMatchLength)
    {
    var fRatingDiff = Math.abs (fRatingP1 - fRatingP2);
    var fRootMatchLen = Math.sqrt (uiMatchLength);
    var fBasicMatchValue = 4 * fRootMatchLen;

    var fProbUnderdogWins = 1 / (1 + Math.pow (10, fRatingDiff * fRootMatchLen / 2000));

    if (fRatingP1 < fRatingP2)
{
     fProbWinsP1 = fProbUnderdogWins;
     fProbWinsP2 = 1 - fProbWinsP1;
     }
    else
   {
  fProbWinsP2 = fProbUnderdogWins;
    fProbWinsP1 = 1 - fProbWinsP2;
  }

    var fExperienceFactor1 = Math.max (1, 5 - (uiExperienceP1 + uiMatchLength) / 100); // Will be 1 when
    var fExperienceFactor2 = Math.max (1, 5 - (uiExperienceP2 + uiMatchLength) / 100); // Exp >= 400.

    var fRatingChange1W = fBasicMatchValue * fProbWinsP2 * fExperienceFactor1;
    var fRatingChange1L = fBasicMatchValue * fProbWinsP1 * fExperienceFactor1;
    var fRatingChange2W = fBasicMatchValue * fProbWinsP1 * fExperienceFactor2;
    var fRatingChange2L = fBasicMatchValue * fProbWinsP2 * fExperienceFactor2;

    return new Array (
       fRatingChange1W, fRatingChange1L,
      fRatingChange2W, fRatingChange2L);
    }

FIBS rating - good for backgammon - http://www.fibs.com/ratings.html

grenv: The following is from US Chess ratings

In any good rating system, if two players with the same rating played a large number of games, one would expect each to win half of the games that were not a draw. As the difference in their ratings increases, the probability that the higher-rated player will win increases. In the U. S. system the difference in ratings at which the better player will win 90.9% of the time is arbitrarily set at 400. A player with a rating of 1100 will win 91% of his games with a player with a rating of 700, and a player with a rating of 2000 will win 91% of her games with a player with a rating of 1600.

For any particular match, it should be possible to calculate from the difference in the player's ratings the probability that one of the players will win. Taking “We” to be the “win expectancy” and “ΔR” the difference in the players' ratings,

We (underdog) = 1 / (1 + 10 ^ (ΔR / 400))
[The formula on the original web page is incorrectly formatted. The one above is correct. ^ is raise-to-the-power-of]

For example, using this formula, if two players differ by, say 90 rating points, the probability of a win for the higher-rated player is 0.627, and for the lower-rated player, 0.373. If the results of a series of games bear out this expectation, the players' ratings are “correct,” and shouldn't change. Players' ratings change only when the results of a match are not what the difference in their ratings led one to expect, and the extent of the change in ratings is based on how far off the expectation was.

So, according to the US Chess formula, the 63% point is a difference of 90 points.

In Backgammon 65% is the difference between a top player and an average player. I believe BKR formula is based on the the one referred to above so we could expect the entire ratings spread to be maybe 100 or so points each side of average!

So, okay, you're right - a difference of zero is exaggerated but with such a small spread and a volatility of up to 10% of that per match? ... they might as well be the same, lol.

I can't wait to see the results of the recalc....

playBunny: That doesn't make sense, there must be a point at which equilibrium is reached where players of different ability reach different ratings.

For instance if I win 62.5% of the time against a weaker opponent, we would balance out at a point where I stand to gain 6 and lose 10, whatever that difference is.

I agree in principle that this system is not good for backgammon, but the ratings would not balance to everyone equal since there is some skill in the game.

grenv: We'd eventually have the same rating.

The formula will consistently pull the higher rated player down by more than they can possibly win because it doesn't account for luck.

So the lower rated player will go up by more than they should and the higher player down until the two have the same rating.

Apply that to all possible pairs and everyone has the same rating!

grenv: In any event, because Fencer is recalculating the ratings from the beginning, we will be able to tell tomorrow whether we all wind up with the same rating!

playBunny: We won't all have the same rating, since if you lose and your rating goes down, you will gain more next time you win. Since there is some skill in the game the list will still be meaningful, though not as much as another system may be.

In fact winning 65% of games against average players is the problem, and there are undoubtedly players with inflated ratings as a result (for all I know I may be one :)

I took my thinking about this a bit further but posted it to the main BrainKing board where arpa raised the issue.

It's serious: This change will have a severe flattening effect. If left alone we will all - high and low - eventually have the same rating!

http://brainking.com/en/Board?bc=1&bscx=403725

(Bear in mind that the link puts my message at the top of the page but doesn't show anything posted afterwards. There may well be replies to it on the "previous" pages.]

alanback: I think you're right about the compression. In fact, because the points awarded favour the lower player so greatly, I believe it'll be over-compressed.

Playing all and sundry in single-pointer tourney matches is going to kill the ratings of the higher players who join a lot of them. (Score one for the Knights vs the Rooks, for a change, lol). Winning 60-65% of tourney matches, as a top player would, will not be enough compensation for the 7 to 1 points differential against players within 300 points and certainly not for the 15 : 1 against those further down.

After the adjustment all those games which previously won 8 useful points will only, in fact, have earned beans while the lost matches will be harshly felt.

Another aspect of this formula is that the ranking system will be highly volatile. A player rated 120 below will have a 2 : 1 advantage in the points awarded though the skill differential is minor and the luck differential is zero. Any player to reach a height can be tumbled merely by winning 50-50 against weaker players.