frolind: I enjoy a good debate. "Good" means that I have someone worth debating with. Redsales, Alanback, Grenv, Walter, Abigail, ... all these and others people are worth debating with. You have yet to go beyond single-sentence abtract utterances. Please explain your reasoning and how I'm wrong.

Alternatively, given that I don't think you're interested in actually debating anything meaningful, let us continue this conversation in the Flame Pit. There is a message waiting for you there.

Fencer: 1) Sure. I may state what I believe though.
2) Yep. I took a liberty there at the risk of making you blush.
3) Does that mean you're going to ignore the opportunity to make this minor but useful improvement?

frolind: "LOL. Well, at least now we know why you were clueless about the subset thing."

No, frolind. You think you're making sense, somewhere in your mind, but nobody else knows what you are talking about.

Please join me in the Flame Pit so that we can conduct an investigation into your mentalstate....

Andersp: Well done that man .

But you've had the benfit of reading these messages and knew what to look for. We are talking about those who needlessly learn the lesson the hard way. It's avoidable with a small change to the message.

Fencer: "It's not my fault they don't read the warning messages."

I believe it is your responsibility as a web designer to follow human interface guidleines and ensure that the possibility of such occurences is minimised. I pointed out to you how this message is part of a larger body of text and can be passed over by the eyes when scanning that area. I mentioned it because that's what happened when I first saw it (though I was just testing the link rather than resigning). And it's happening again as in the case of alanback and his and pgt's opponents.

If I may take the liberty of repeating a snippet of our conversation in our test match...

playBunny: (15. October 2005, 15:08:04) No problem. [The decision not to implement a resign/accept dialogue] In effect it means that resigning is not an option and games must be finished the long way. That can be lived with. There will be those unfortunate enough to resign without realising what they've done but a few sharp slaps will teach them. [emphasis added]

I'd certainly recommend that the warning be in red and/or in the next font size up, or add to it. At the least the warning line should be separated from the text above by a blank line.

Even better, I'd use the text: "Are you crazy??? Can't you see how many points you're giving away?!!"

(I'd also suggest taking out the "according to the backgammon rules" unless you're going to make it part of your own rules, because nowhere awards gammons or backgammons in this manner.)

Fencer: (15. October 2005, 15:09:32) Hmmm, maybe "according to BrainKing backgammon rules" would be better.

playBunny: (15. October 2005, 15:16:35) Aye, that would be better. You'll get no arguments then, lol.

Fencer: (15. October 2005, 15:18:20) Done.

playBunny: (15. October 2005, 15:32:10) Super. And the bit about the warning? The "crazy" bit was obviously a joke but the highlighting suggestion was serious. I'd say the in-red idea is best, with the extra blank line being a close second.

Fencer: (15. October 2005, 19:46:50) Yeah, something will be done.

alanback: "The Crawford game is always cubeless; also, any game that is played when both players are 1 point away from winning the match is cubeless."

That's missing the point, or rather using the wrong perspective. "Cubeless" is a synonym for chequer play. But every backgammon game - in cubeless matches, cubeful games, Crawford and DMP games - has chequer play. However, no cubeless match has the cubing, gammons, game sequence, etc in it. It's the addition of the cube to chequer play that creates the new variation.

There aren't separate ratings among golfers for driving and putting, or among tennis players for serving and volleying;

For sure! There are no putting-only tournaments or driving-only championships. Tennis players don't play three sets of volleying. That's because these things are only parts of their game and are inseparable.

"so why should there be different ratings for cube play and checker play? "

What we are talking about is where an entire match either does or doesn't include a set of features. It's more akin to sprinting and hurdling. Hurdling requires sprinting but not vice versa. They make totally different events, ones which do have separate ratings and rankings. Cubeful Backgammon requires chequer play but cubeless Backgammon is sufficient unto itself. The two are substantially different for all the reasons previously given and are similarly justified in having separate ratings.

"... However, that doesn't mean that separate ratings should be maintained; it just means that a complete backgammon player has to be able to use the cube properly."

This actually highlights another factor in favour of separate ratings: the so called "incomplete" backgammon players! Or are you suggesting that the cube be obligatory and players can like it or lump it?

alanback: Indeed, by that definition I'm definitely not a "real" backgammon player. Hands up all the "real" players! Speak out - your hypothetical future is being discussed.

Do these cubeless games actually occur as part of the tournament and are they counted and rated?

BIG BAD WOLF: Where do the 10 cube choices come from? Do you mean match length? Ratings shouldn't be separate in that case. Ladders should, of course, because their games need to have the same parameters.

alanback: I guess I'm not a "real" backgammon player then?

or: Don't "real" backgammon players think that playing cubeless is silly? Having separate ratings is then meaningless because the cube should be mandatory.

or: IYT has cubeless and cubeful. Same with Gold Token.

and: I'd like there to be separate ratings simply so that I can enjoy twice the challenge - to aim for #1 in chequer play and #1 in cubeful play. ;-)

frolind: For some reason you've got my name in that message but you were talking toyourself. Duh?

frolind: Well spotted frolind. Perhaps you'll share with everyone here what I'm like. You arecapable of actually explaining things rather than using vagueness, generalisation and innuendo, aren't you?

No, I thought not.

alanback: It's item 1) on the Requested Features list. BBW was the proposer at that time. Support for this one is getting strong. I'm forever scrolling down to see what the state of the match is (and what type of match it is) so I'll add my vote, too.

redsales: Okay, so we agree that the doubling cube (especially when taken to the extreme) involves different thinking. At the more basic levels it means assigning a probability to a position, a skill which is quite dissimilar to the major one of chequer play, ranking the possible moves). At the higher levels it involves mathematics and memory (eg. remembering match equity tables).

We also agree that both the doubling cube and the castling rule introduce additional skills into their games.

So now I wonder about the degree of these additional skills. The discussion started with the idea that doubling introduces sufficient difference to warrant a separate identity.

Let's look within the backgammon arena. Hypergammon has exactly the same rules as Backgammon but has fewer men. This adds "additional skills" but I'd be hard pressed to argue that they are sufficient to claim that the thinking is different from that used in Backgammon. The tactics and strategies may be different but the thinking is still a question of ranking possible moves. Deciding that order is the same as in Backgammon but with more emphasis on probabilities and less on creating structure (ha ha, structure?). I'd say that any claim regarding additional skills can only be made in terms of emphasis rather than newness. Yet, while I can't put forward any additional cognitive functions, the game is obviously very different to Backgammon.

Considering Nackgammon we find that the additional skills argument is even more tenuous. It's almost pure Backgammon but with two men moved. Again it's a very different game, not drastically so as in Hypergammon, but still enough to give it a separate identity.

Now let's consider Chess. The castling rule adds additional skills in a quantitative way however I can't see that additional ways of thinking are required to tackle a castled king. As in Hypergammon there are new tactics involved, but no new brain function.

Chess has its nackgammons, too, with Fischer Random, Gothic, CRC. These may change strategic and tactical thinking but make little, if any, demand on new cognition. Yet these are branded as distinct variations. The different setups change the course of the game, taking the player through different areas of the game space, and that provides them with a distinctive feel.

The castling rule doesn't affect the whole game in the may that having only three does men in Hypergammon, or a different setup. Maybe that's why it's not a separate game from non-castling chess.

So where am I going with this? Well, the "additional skills" examples are almost all the same cognition but different game space. They still, however, make for separate games.

The doubling cube requires skills additional by an "order of magnitude", different cognition; doubling is pervasive, the option to double has to be considered by a player for every move while there's access to the cube; it means that a set of matches is no longer a collection of discrete games but an ordered sequence (cube decisions in later games are based on the match score). I haven't even mentioned the effect that gammons and backgammons have on the game, though they are a change that goes hand in hand with the introduction of the cube.

All of these things mean that Backgammon with the cube is a substantially different game to that without. Merely changing the position of two men has been enough to make a new variation. I suggest that doubling more than qualifies Backgammon for similar treatment.

The final point; we can elect not to have the cube in a game. In what other game is there the option to remove a complete chunk of rules and game play? Take Cylinder Chess, for example. The only difference from standard Chess is the wrap-around of the board. Can you imagine making that an option of standard Chess which players may switch on when they create a game invitation? They would both be called Chess and all games would contribute to the same BKR. You could look at a players played/won/drew/lost stats and BKR and have no idea what contribution was made by the two variations. Some players would never play the Cyclinder sub-variety, others may specialise. What would the Chess community's reaction be to that?


Given the above arguments I believe that Backgammon with the cube sharing the same identity as Backgammon sans cube has no basis in terms of the game itself. It's a pragmatic decision which keeps the number of Backgammon variations down for simplicity. There are 26 variations of Chess, only 6 of Backgammon. If we were to have doubling cube variations of those 6 then Backgammon would be in danger of catching up with Chess, especially as each new Backgammon would introduce 2 variations. [Lol, sorry Fencer, I couldn't resist that one! ]

frolind: Lol. I've said nothing of the sort. It's a good idea to stop; you'll be talking complete gibberish if you carry on.

frolind: Carry on, lead me from generalities to specifics....

frolind: Lol. Don't despair. Here's an opportunity to explain to a poor dumbo like me how one thing being a subset of another means that they are not significantly different, how the extra that the superset contains doesn't matter, doesn't serve to distinguish the two sets.

redsales: Aye, castling opened up new and interesting areas of game space to be explored... but did it cause a difference in the kind of thinking that was required?

Have a look a Doubling Strategy in Backgammon and see if that bears any resemblance to the thinking involved in chequer play! ;-)

(I'm cheating slightly here. That kind of stuff is obligatory for professionals and world champions but you can get quite far with the cube without having to learn all that. It's very useful to understand the concepts involved though, even if you don't get as far as playing with the equations during a match)

alanback: Aye, it's worked for me in this match.

redsales: I'll be happy enough with the ladders too but I can't agree with likening the cube to minor additions to the moves in chess.

There might be similarities in the rationale behind their origins but after that the comparison breaks down. The cube requires additional skills. Gaining true expertise with the cube is more of a challenge that attaining the same level in chequer play, I would suggest. The en passant and castling rules hardly stretch the chess mind in a new direction.

frolind, alanback: Saying that cubeless backgammon is a subset of cubeful backgammon is the same as saying that hurdles is a subset of sprinting. It may be true but nobody thinks of them as being comparable. The are treated as separate events.

Czuch Chuckers: I'd like to see separate rating scales too; it would be another challenge and it would mean that the ratings are "cleaner" in as much as some people will play only with the cube while some others will play only without and being in the same rating pool makes then incomparable.

Fencer will at some stage be introducing Ladders to BrainKing. It'll be possible to have separate Ladders for chequer play and cube play.

grenv: Yeah, it's like a drug! You can't beat the feeling of being 4-1 down in a 5-pointer and winning gammon with a 2-cube. Better than sex, Ecstacy, crack..

(Okay, I exaggerate. It can't beat a nice cup of tea)

Walter: You could always agree with your opponent that neither will double.

The Backgammon rating formula values a match as 4 x Sqrt (Match length). This is then apportioned according to the probabilities of each player winning (which is calculated from the rating difference).

grenv: At the top you specify the Match type as N points with cube. Down towards the bottom, below the time controls, you specify the Final likewise.
(It makes more sense to me that the match types go together.)

frolind: Those are great. For those who like life to be easy, here they are (minus Question 4 which couldn't be found):

[Q_1]   [Q_2]   [Q_3]   [Q_5]   [Q_6]   [Q_7]   [Q_8]   [Q_9]   [Q_10]   [Q_11]   [Q_12]   [Q_13]   [Q_14]   [Q_15]   [Q_16]   [Q_17]   [Q_18]   [Q_19]   [Q_20]   [Q_21]   [Q_22]   [Q_23]   [Q_24]   [Q_25]   [Q_26]   [Q_27]   [Q_28]   [Q_29]   [Q_30]   [Q_31]   [Q_32]   [Q_33]   [Q_34]   [Q_35]   [Q_36]   [Q_37]   [Q_38]   [Q_39]   [Q_40]   [Q_41]   [Q_42]   [Q_43]   [Q_44]   [Q_45]   [Q_46]   [Q_47]   [Q_48]

Hope this is fun....

It's a set of questions posed on the Usenet newgroup rec.games.backgammon and viewable via Google Newsgroups. As such it uses text format, so the board setups are a bit awkward to read, especially as chequer movement is the "wrong" way and the home tables are on the left.

People discuss their answers but the final answer is uisually in terms of a GbnuBg rollout. For those not familiar with that it will probably look like so much gobbledegook. Don't worry too much about that; the discussions should be interesting.

Questions 1, 2 and 3
Question 4
Question 5
Question 6
Question 7
Question 8
Question 9
Question 10
Questions 11 and 12
Question 13
Question 14
Question 15
Question 16
Question 17
Question 18

Fencer: Ah, beg your pardon. That one came to me as I was waking up this morning so it's untested. Naughty me made an assumption. I'll delete the list just posted as there's nothing else new on it.

The list so far.

1) Would be nice to have the current score up next to the playing board instead of buried down in the "game information". Or maybe next to the Points, have the current match score for each player as well.

1a) Change "Points" to "Pips".

2) In the table of "Finished games of this match" which displays who won each game, it would be nice to see how many points were won in each game (without having to select each game and counting the "D"s!)

3) In the list of games which make up a match, the winner is highlighted with a yellow background, no matter which player won. (Example) It would be easier to read that table if each player had a different colour. So Marfitalu could be in yellow (when he won) and Tulip in cyan (when she won). The loser would always be in white, though.

4) A time out forfeits the whole match, not just the current game.
[playBunny: i don't like this one. My opponent in a 5-pointer is about to time out. I don't like winning that way. We're bearing off and I've effectively won already, so it's okay, but I'd hate to get the whole match just because he missed the end of this game.]

5) No draw offers on single games.

6) Show the cube on the correct side of the board after a double has been accepted. That is, with the owner, the one who is able to offer a double.
Done

7) Change the "D" and "+" in the moves list to "Double" and "Take". [Fencer says that's a biggie because notation is language-neutral and the the notation code was designed without translation requirements in mind. Might be okay if other languages don't mind "Double", "Take" and "Drop", perhaps?]

8) For completeness, Add an explicit "Drop" to the moves list (though it would be "-" in the current notation).

9) Auto-roll. Doesn't give the Dice/double choice but goes straight to the dice. Useful for players when they know they won't be doubling for a while.

10) Buttons for [Roll] and [Double] instead of links.

11) Use the words "Take" and "Drop" on the "double-offered" buttons rather than "Accept" and Reject". (Dropping the cube makes more sense to me anyway because it also means dropping the game.)

12) A note at the end of the moves list of a finished game saying "XXX wins P points (Cube 2 x N)" where N is the "degree" of the game, 1, 2 or 3.

13) Display the dice roll in the moves list. [This one's a high priority for me]

14) Implement the request-resign-accept/deny dialogue.
14a) Or highlist the "Warning: N points will be lost" message, for example, in red.

15) Allow the whole match to be resigned.

Vikings: That's correct. Even if it's obvious that you won't be gammoned, a resignation will cost a gammon or even a backgammon unless a man's been taken off first.

frolind: I can see how that would be annoying in turn-based games. At Vog, where it's real-time, that sort of foolishness is rare (perhaps because it's more hassle asking for the resignation than it is to say no each time). I'm okay enough with the current way at the moment, though I'm concerned for those players who get caught out and lose points needlessly.

BBW I've slightly changed what I said in that post so it may not read as you originally saw, but the bit about avoiding backgammon by resigning a gammon wouldn't be the case. (Like you, I did think it might but it's more about losing more points than you should rather than fewer.)

LionsLair: I've already asked Fencer about that. The cube games are not separate variations and so they "share" the same ratings scale. There are arguments for having the games as separate (another rating scale to climb) and for not (it keeps things simpler, not least of which is the length of those dropdown boxes!). Fencer's all for simplicity at this point!

He did mention, though, that when he unveils BrainKing's Ladders (in due course, lol) we'll have separate ones for cube and non-cube backgammons, so we can use those to demostrate our respective chequer and cube skills. :-))

There's a major departure from how we understand backgammon to work.

Resignation has not been implemented in the standard way. Fencer didn't realise that resignation is a multiple-choice request that the "winner" accepts or rejects. Instead, we have a Brainking version that automatically decides whether it's a normal, gammon or backgammon loss depending on the postions of the men. If there's a man on the bar or in the winner's home table then it's a backgammon - regardless of how the winner's men are positioned. If the loser hasn't yet born off a men then it's a gammon. Otherwise it's a normal loss. There's no choice for the loser and hence no requirement for the winner to accept or reject. An example of how strange this is - if you resign when it's your first move then you will lose by a backgammon!

The result of this is that resigning is not an option unless you are going to lose by a gammon anyway. To have a normal single-point loss means having to play the game out until you've taken a man off the board. Only then can you resign a single.

I've spoken to Fencer and my feeling is that he's put a lot of effort into "The Bug" and into getting the BG+cube together and that adding a request-resign--accept/deny dialogue is one complication too far. As resignation is only a shortcut, it's not as serious as The Bug, ie it won't mean the wrong player losing, but it will extend games longer than otherwise or cause unwary players give away points that they needn't. Because of that last possibility, I've requested that the warning be highlighted in red or something.

(I'll add the short version of this issue into the list but thought it deserved a separate post)

BIG BAD WOLF: Indeed! There's a rule called the Jacoby rule which states that gammons and backgammons don't count unless the cube has been turned at least once (when, if you accept, you should be taking into account the game position and risk). It's a very reasonable rule for Hypergammon given that gammon is so common.

grenv: I agree, though it's nice if it looks correct on the board for the same reason that BBW's idea about having the scores near the board is good - it saves having to hunt around the page for the information.

lovelysharon: Lol. What if he never conmes back? The never ending match.....................

lovelysharon: Is Diogenes going to resign or are you going give him a reprieve?

rod03801: "Will the cube be usuable in the backgammon games in those? Will there be an additional step in the creation asking if you want the cube for the backgammon portions?"

Interesting questions. The Backgammon-with-cube has the same Game Id as plain Backgammon so no existing tournaments will be effected. With new ones I imagine that you'd either have the cube with your gammons or not but not both. (Maybe there'll be a single Use doubling switch covering all games or maybe there'll be both versions of every backgammon name.)

LionsLair: Lol. I think the bug fix was the bone. This new development is a free pass to the butchers store room! (For you lions, that is. For bunnies like me it's a most juicy field to nibble in, full of luscious grasses and tasty flowers)

WhiteTower: I reckon I'm still missing something then because you wouldn't play the 1s to bearoff but would shift other pieces down. Unless we're talking complete Duh? players.

Ah, perhaps if the turtle had men trapped on the hare's 2 and 3 points, for example, and hadn't been forced to relinquish them ... ?

WhiteTower: How so? The man trapped on the ace point can't be released until bearoff is finished? Or am I thinking of something of the wrong game (I know it as Tapa)?

Fencer: Coool. Count me in.

Fencer: Who's LongJohn?

Hrqls: Black (wayney) resigned because he was a cheat. He was kicked out of BrainKing for rating manipulation.

BrainKing04's involvement is suspicious, especially with regard to his games with bigboss. Five games, all resigned by bigboss before the first move.

Reversi 8x8
10th Oct
29th July a
29th July b
23rd July
21st July

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.

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);
    }

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....

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!

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.

Hrqls: You are required to use the dice to tha maximum but not necessarily the pips, so what you did is perfectly legitimate and the best move.

alanback: The seed posts.

Bry: 21 days time control?! Has the first tourney of that series finished yet? Lolol. ;-)