I agree with Alan that Hyper is different enough in the game itself to be excluded from the average from certain aspects (eg, that it doesn't have the structural game play of the other variants) but for completeness and ease I think it should be included in any group measurement. Otherwise we'd have to examine each game to see whether it belonged with its fellows or was an outsider within its own clan. Which chess variants are not quite truly chess? The only exceptions I'd have are the anti- games which I think are almost an abomination under the eyes of Great Bunny.

Hauling out the old stats that I posted a while back...

Game ........... #1 ........... #20 ....... Average ....... Range ...... Players
Hyper ........ 2188 .......... 2120 ..........1960 ..........160 ........... 536
Back .......... 2431 ..........2210 ..........2002 ..........210 ........... 882
Nack ......... 2011 ...........1881 ..........1696 ..........185 ........... 257
Crowded .... 2081 .......... 1853 ..........1690 ..........163 ........... 202
Race .......... 2026 ..........1911 ..........1688 ..........223 ........... 234

You can see from the ratings of the #20 and the average rating that Nack, Crowded and Race are somwhat similar but not in terms of the range. Backgammon and Hyper are in their own camps. any average which was only an average would be skewed by the Backgammon rating. The only meaningful ways to produce a representative value are the ones that normalise the raw ratings before combining them, for example calculating the standard deviations of a player's ratings, averaging them and then multiplying by some factor to make it look like a rating again. I don't know the maths well enough to know whether that can be done incrementally after each match. It could be difficult keeping them up to date.


Just out of interest, here are the same figures now that we're two weeks into the new BKRs. The range for Hypergammon has shrunk by 7.5% and those for Crowded and Race expanded by about 6%, and Nack has expanded by over 8%.

Hyper ........ 2313 .......... 2122 ......... 1974 ......... 148 ........... 532
Back .......... 2441 ......... 2216 ......... 2008 ......... 208 ........... 898
Nack ......... 2108 .......... 1893 ......... 1693 ......... 200 ........... 252
Crowded .... 2086 .......... 1865 ......... 1693 ......... 172 ........... 198
Race .......... 2043 ......... 1932 ......... 1696 ......... 236 ........... 233

Hrqls: Lolol.

Well the pages are there. All it takes is a bit of detective work (ie of the shoe leather variety).... I look forward to seeing your results. ;o)))

BB: Alan only ranks #422 in Hypergammon so that one's not a real gammon and doesn't count.

alanback: I read this somewhere: "In addition, if your self-esteem depends upon achieving a high rating on this server, you really need to reconsider your goals."

Hrqls: It's easy. Go to the game and count the number of times the cube doubled, unless the game ended with a dropped cube in which case ignore that one. Then check to see whether the opponent lost a single, gammon or backgammon (applauding their foolishness if they did the latter by resigning prematurely). Then do the arithmetic thing.

grenv: Greek backgammon tradition eschews the cube. Given the chance they'd eschew it and spit it out.

WhiteTower: Why would you break the block and run with one of the men?

1) Because you're not that far behind in the race and it's worth running and hoping for a tasty double to get the other guy moving too.
2) Because you've run out of timing and running a single man is necessary to prevent your home table from collapsing.
3) Because the odds of a hit and return hit are higher than the odds of catching a man by having a block.

Do you have any positions in your games that we could be specific about?

grenv: With the cube I do sometimes double prematurely against someone who I think will drop. But that's a different mistake to the wrongful resignation one. I didn't do a thing, I was just bearing off as normal and hoping to win a gammon. Getting the match by mistake is a useful win but it's no victory.

Pedro: Aye. Don't *twinkle* BOP Do *twinkle* BIFF It *twinkle* THUD Again *twinkle* POW!!


In fairness to ZEROZERO: Take a look at your resignation pages in the German
and French versions. It's the same old hidden black message. I see it as simply more proof that this warning needs to be in red for everyone.

[Those links are specific to Pedro. Anyone else who wants to see should click any resign link in one of their own cube games and then change the /en/ (if your language is English) in the address bar to /de/ or /fr/.]

Pedro: It has? With the red warning and all?! Who did it?? Let's all tell 'em off.

Twinkle? Nice one. Good idea that man!

grenv: "Anybody who understands the game at all would have tried avouding the gammon (or backgammon) before resigning."

Actually, with cubed backgammon being so new and resigning having always been a straightforward "This game (only) is lost", it's reasonable to expect some inappropriate resignations to be made by mistake. Call it foolishness, being a turkey or human nature as you like.

I prefer the latter but with this new red warning, if it ever happens again I'll join you in pointing the finger of accusation. LOL ;-)

WhiteTower: (LOLOL)

Pedro: [[Sure you can write a message all in red. You just have to know how to use BrainKing properly. And a genius like yourself surely knows everything about BrainKing?]]

Fencer: Thanks, that's all it needed. (Though I don't think "Are you crazy??" would be over the top when it's resigning a Backgammon. )

Pedro: I'm glad you're such a superior being, better by far than us "stupid turkeys". Those turkeys include Tayfun, myself and alanback so obviously we're talking about real idiots here.

Pedro Martínez: It's happened to numerous people and it's "only their fault". Each time. Time and again.

The point is that it's preventable. Trivially preventable. Good design doesn't encourage people to make mistakes. Good design takes human perception into account. This is generally the case here at BrainKing but here is one instance where it's clearly not the case. You may not have a problem with it but many have and many more will. Ask those who have failed to notice the warning whether they would have made the mistake if the text was in red.

It's preventable.

Fencer:

In this match my opponent was ahead 3 : 2. I was winning the 4th game comfortably and my opponent decided to resign. He didn't see the warning about points loss.

This was our conversation:

To: TC
Date and time: 6. November 2005, 16:16:45
Subject: Re: Your opponent resigned the game (Backgammon)

Thanks. :-)
I don't understand why you resigned though, especially a backgammon with a 2-cube! You could have run to save the gammon. Another three rolls for me to get off and only two needed to save yourself.

You are reading a reply to this message:
You got it, congrats!
___________________
From: TC
Date and time: 6. November 2005, 16:30:56
Subject: Re: Your opponent resigned the game (Backgammon)

I thought with my resign only the game and not the match! This is my failure and may be also bug for such type game matches.
___________________
To: TC
Date and time: 6. November 2005, 16:34:12
Subject: Re: Resignation

You're mostly correct. Resigning itself only pertains to the current game. But you resigned a backgammon when the cube was at 2 and so you gave me 6 points. It was that which ended the match!

Did you not see the warning message telling you how many points would be lost?
___________________
From: TC
Date and time: 6. November 2005, 16:45:04
Subject: Re: Resignation

No I did not see any message!
___________________


I really don't like winning matches this way. It's even worse than winning through a timeout.

Fencer, please make the change to the warning message. How many people are going to have to lose significant amounts of points and even matches because this trivial change is left undone? I really fail to understand your reluctance. Andersp has made the change in the Swedich version and I am certain that there will have been no players mistaking a bright red warning notice for just another line of black text in a block of lines of black text.

Important stuuff should stand out. It's basic design!

The same principle applies to the match score. I know you won't move it where up it can be seen but perhaps you could see your way to making it blue instead of black?

grenv: The figures are from the list of established players. Median would be a better word than average though I doubt the average would be much different. I'm not sure what effect the total number of games would have other than making for a smoother bell curve distribution.

I was also surprised when I saw how high the averages were. Compared to any other site these numbers are silly. At FIBS, for instance, average players are around 1500 and top of the range is 2000. At Vog its 1600 and 2050.

The following order is that given by Pedro and Hrqls:


Game ........... #1 ........... #20 ....... Average ....... Range ...... Players
Hyper ........ 2188 .......... 2120 ..........1960 ..........160 ........... 536
Back .......... 2431 ..........2210 ..........2002 ..........210 ........... 882
Nack ......... 2011 ...........1881 ..........1696 ..........185 ........... 257
Crowded .... 2081 .......... 1853 ..........1690 ..........163 ........... 202
Race .......... 2026 ..........1911 ..........1688 ..........223 ........... 234

The Range value is that between the average and the #20, on the assumption that the top 20 are unrepresentative. Given the sample size, that's open to debate. (If you fel strongly enough to argue, just post your own version)

It's interesting to note that the average ratings are in the same order, though it's very close with the last three.

What any of this really means though ... ?? ;-)

grenv: Lol. That's exactly why I made the poll. I'm interested in seeing how opinions correlate with the ratings and whether they both add anything to Hrqls' enquiry.

As a side issue in this topic, here's an opinion poll...

What, in your judgement, is the best order for the following in terms of how much luck is involved? (Highest luck factor at the top).

   • Backgammon
   • Backgammon Race
   • Crowded Backgammon
   • Hypergammon
   • Nackgammon

LionsLair: You should have moved out of your opponent's home anyway! ;-) At that stage of the game your own home table is your only asset. Breaking your home table should be done with the utmost reluctance. You drastically reduced its effectiveness by moving 6/1 4/1. A 3-point table is next to useless at holding an opponent. Better would have been to run, 22/14, and get the man home, hopefully to make a block on 6. That would give you a 5-point table and might save the game. The danger of your two remaining blots being hit is certainly a worry but keeping those back men safe is pointless if, when they get a chance to take action, it's ineffective.

Andersp: "It's already in red in the swedish/norweigian translation"

Way to go Sweden/Norway!

Are you gonna bump up the font size of your match scores and put them in blue? If you do then I shall go green.

grenv: Oh, I quite agree but it ain't gonna be moved. I reckon that's total pie in the sky. Adding colour is a suggested alternative.

grenv: It's clear once you know it's there but several people have missed it. All it would take is a splash of paint...

Talking of colour... If the match score isn't going to be moved up beside or above the board then putting it in colour would be a helpful alternative. At least the eyes will spot it immediately when scrolling down to find the score.

Hrqls: Lol. You missed it because it didn't say it or because it's easily missable?!!!

All hail Fencer to highlight that warning!!

Woolsey's Law

To double or not to double, that is the question. What should our thought process be?

First of all, we (mentally) move to the other side of the table and ask ourselves the following question: If we were doubled in this position, would we take? There are three (not two) possible answers to this question:

1) Yes, I'm absolutely sure it is a take.
2) No, I'm absolutely sure it is a pass.
3) I'm not 100% sure.

If the answer is (1 - Take), then it still might be correct to double. This would be the case if there were a significant chance that on the next exchange (we roll, he rolls) he will now have a big pass. If the position is that volatile, it is proper to double even if the take is easy. Otherwise it is correct to wait.

If the answer is (2 - Drop), then it still might be correct to double. This would be the case if either the gammon chances were small or if there were a significant chance that after the next exchange (we roll, he rolls) he will now have a big take. Otherwise it is correct to wait and play for the gammon.

If the answer is (3 - Hmmmm), then it is always correct to double. I have written this elsewhere calling it Woolsey's Law. It is valid in virtually all situations. The only possible exception may occur when you are well ahead in the match and the double would put you out or nearly out. In this situation it may pay to be conservative. Otherwise, follow Woolsey's Law. You won't go far wrong.

Source: Backgammon Galore

Fencer: Lol. It would be very interesting to implement it, but that's a game that only one can play - the programmer!

I'd vote for perfection in the existing games myself before effort went into new games but Grasshopper does look like it would pose some interesting new challenges.

Wow. I just made 10 moves in a row at Gold Token using the you know what facility.

It was GREAT!!

Andersp: Lol. You took the doubled ofered and redoubled in roder to end the game quickly? Resigning is a tad quicker than that, methinks. Or did you mean the match because the cube at 16 certainly ended the match.

That aside, I'm very curious about why you wanted to end the match??

Andersp: "translated to english : "i know the rules i also know that there are cheatingprograms "

Whinging about cheats, eh Anders? Can't accept that you lost?

Seriously though. If a player like frolind doubles, you should never redouble on your next go unless he's rolled an awful, terrible, stinking number.

What frolind said about "gifts" is very true. Your second redouble was immediately after accepting a double that should have been dropped as if it were a hot rivet. In each case you redoubled when frolind's post-double roll had sent a man of yours to the bar. In the second case that meant two on the bar and with a 4-point table to get back into and builders at the ready!

WhiteTower: In the short term yes but not in the long term. So what is it that they need to make sure of?

redsales: Lol. Good side step :-)) I don't understand why the machines would need such things as payoff and payout adjustors. That's the first thing I'd like explained because my thinking is that randomness alone is sufficient to ensure the percentages - over the long term (which is what they claim in the article).

WhiteTower: Slot machines? Surely not!

:op)

Hrqls: I know as much as you.

... one interesting browse later...

Random.org offers true random numbers to anyone on the internet. If you want to know how the numbers are made and what it is that makes them true, read the introduction to randomness and random numbers.

Now I know a little more. ;-)

alanback: Lol. Yup. That's why I specified an infinite series. ;-)

If you like we can bring statistical significance into the mix. That'll allow us to talk about finite series. Not iron-clad as guarantees go but good enough.

What I'm waiting to hear is how these machines are programmed to effect whatever guarantees redsales is talking about. What are they deciding on? How do they know when the randomness so far isn't good enough? When is it time to increase or decrease the percentages?

redsales: "pure randomness cannot guarantee a certain %age, because there are no guarantees in a purely random system."

That is soooo wrong! If it's perfectly random then an infinite series is absolutely guaranteed to have whatever percentages. It's the not-quite-random ones that are dodgy.

You'll agree that perfectly made dice are perfectly random? And that the 1/6 possibility of a 6 occuring is not a random number but is "manipulated" by the dice designer (through the choice of having 6 sides)? Are the dice "skewed" to give 1/6 of each number in the long term or is that just how it comes out in the natural course of events?

So how are the machines any different? Why would they need to be?

redsales: Why would it need to?

Dice are "programmed" to give a 6 in 1/6 rolls but the pattern of 6s in the stream is entirely random. Over the long term, though, the 1/6 can be verified.

The slot machines have their own percentages which can be verified over the long term as well, but at any one moment the machine may be ahead or behind its quota. On what basis would it make it more or less likely to roll a winner given that it "knows" that all will come out in the end if it leaves well enough alone?

On the other hand is this the "looseness" and "tightness" that the article was talking about, where the machine ups and downs the odds for some reasons (that I don't understand yet)?

Edit: Ah, I've reread the article; I'll answer my own question. The looseness and tightness is an observation and perception of the winning pattern and not something generated knowingly by the machine.

alanback, Vikings, playBunny: Aye, you're all correct. When I refreshed the page skipinnz's game situation message was off screen. I only saw Vikings' in isolation and it thus seemed too general.

Vikings: That's not quite true. It's the first time that a 4:X score is reached, whether 4:0 or 4:3.

skipinnz: One point or more, depending on the match score.. Dead cube (and other terms).

I believe the cube was introduced to make gambling more exciting.

grenv: Not for ajtgirl. She went to bed firm with her belief. I'd still like to explore that.

Yep, "Roll" -> "Pass" sounds good, and as manual as usual.

Vikings: Lol. I hope so! I never write stuff that doesn't make sense to me - at the time. Other people, and me at a later time, might have different opinions, though!


I wonder, does the possiblity for misinterpretation make sense as I explained it? It made sense to me!

Vikings: Everyone knows about using your left hand!

It was this bit: "its what you are wanting, when someone can't move, you can go ahead and go again". I know what it says now but it took some effort. Part of the confusion is that "you" would be ajtgirl who has been talking about when she can't move. Your sentence switched viewpoints and "you can go ahead and go again" didn't make sense as "you" is stuck.

ajtgirl: Vikings tried to explain but I didn't understand what he said either, lol - have to read it a few times.

Automatic passing (Auto-pass) means that you don't have to roll the dice when you're stuck on the bar against a closed table or in any position where there's no point rolling the dice. You skip your turn and the opponent goes again. I reckon that idea sounds familiar to you, eh?

Many, many people have asked for this. Many more would like it but haven't asked for it. Beyond them there are many more who haven't thought about it but who would say "That's a good idea" if they were told. There are probably a few going the other way; players who like rolling the dice when they can't move, but aren't there always, lol.

As I understand it, it would require huge changes in the code that plays games, so there's one person who really, really doesn't like the idea of auto-pass.

ajtgirl: If you're rolling real dice, do you believe that the odds change every time you roll them? If you roll a 6-6 with real dice does that reduce the odds of rolling 6-6 on the next go?

grenv: 'Drop' makes loads of sense and it's well established as part of Backgammon terminology, but I take your point. Once you've got used to the words, whatever they are, it's second nature and you can't see what the fuss is about. I'm still cautious with these BK buttons, don't want to go the wrong way, but I'll get there, as will Rose and everyone else... I suspect that I'll eventually not go to the word but will learn that left is take/continue and right is drop.

At Vog the words "Take" and "Drop" are used. perhaps because they are short and fit onto the board in the area that's provided. When playing GnuBg I don't read the words (Accept and Decline) as they're are rather small on my monitor, but hit the green tick or the red cross as appropriate. I think such icons would go nicely with the indistinguishable grey buttons here.

Why not change "check" to "I can take your king next turn"?
They could at least spell it properly - cheque.

Rose: I think the wording should be changed, with "Drop" instead of "Reject". It's much harder to misunderstand as it works for both the cube and the game. Drop one you've dropped the other. And "Continue" would perhaps be better than "Accept" (although the usual term is "Take").

alanback: "there's no bluffing in backgammon (it's a "perfect information" game)."

It may be perfect information in the game theory sense but bluffing is most certainly an option because backgammon is rarely a perfect opponent game. Grenv's example below illustrates this. An opponent who accepts a cube offered in order to terminate the game and who then redoubles? That was not a bluff but it shows how a cube can be taken under the wrong circumstances. The converse is also possible. The weaker player knows they're up against someone much stronger and drops the cube as soon as it appears even though it was offered too early (ie, a bluff).

Another form of bluff, perhaps, is the offering of a blot when you know that hitting it will likely cause the opponent more trouble than letting it go. I've been on all four sides of that one, lol.

Walter: It's called a dead cube.

alanback: "Ratings are an approximation and are no better than the data" ... and formula ... "on which they are based."
;-)