Auf jeden Fall, mir hat man mal beigebracht , das Backgammon eine Wahrscheinlichkeitsrechnung ist . Bei BackgammonGalaxy ist meine Erfahrung je niedriger die Chance ist um so höher die Möglichkeit das der Algorhytmus oder wer immer entscheidet was gewürfelt wird , eben genau den Zug bzw. Wurf bringt. Man kann Wetten darauf abschließen. Also Finger weg und wer nicht ohne auskommt auf keinen Fall echtes Geld in diese Seite investieren.

Translation, anyone?

Translation:

Anyway, I have been teached that backgammon is a probability calculation. My experience with Backgammon Galaxy is: The lower the probality of an event is the higher are the chances that the algorithm or whatever lets this event happen. One could bet on this. Hands off! Anybody who can't live without it should never invest real money in this site.

O Tannenbaum, o Tannenbaum,

Wie treu sind deine Blätter!

How ignorant and disrespectful it is to post in his native language on an English forum? But taken into account the content... well at least the ignorant part matches....

I won't b**ch about doubles however this game's error rating system is terrible for those who blitz. I tend to float right around 1600 rating wise and 30 games ago I switched to only playing players rated 1850 and above. My win ratio (about .700) did not change one bit however, I rarely win the rating game. I just looked at my past 10 games and I won 9 out of 10 yet of those 9 wins I only won a single game's error rating. Generally, people are much better at playing by the book than me but when I go through the analysis I wouldn't change 80% of what is being called errors and blunders as the moves led to success. Seems to be a significant fallacy in the ELO system here to me.

... when I go through the analysis I wouldn't change 80% of what is being called errors and blunders ...

There's your problem. Good players get to be good by learning from their mistakes rather than insisting that their mistakes must in fact be correct.

There is a word that describes the phenomenon where you commit lots of blunders and still win 70% of your games. It especially applies when you have only played a small number of games (such as 30). That word is luck.

You need to remember that in games like backgammon and poker that have significant luck components that winning and playing well often have very little relationship over the short term. You can easily play well and lose, and you can play badly and win. Your 21-9 record proves absolutely nothing. You probably should listen when the ‘bot that is the current state of the art in bg tells you that you have blundered. It is almost certain that you made bad plays in those instances and got bailed out by luck.

Who stubbornly refuses to accept that he made an error is condemned to repeat this error again and again and again ...

The example is not 30 games, it is several hundred. The 30 games only applied to taking on 30 higher-rated players. If you read my post my point was changing to higher-rated players made no difference, you are way off my point. Prior to those 30, I took on evenly-rated and higher-rated players. Your "luck" explanation is BS.

But you're wrong. In those cases a blitz approach and a non-norm approach won. Just not the error rating. I don't think this is sinking into your head. nvm

Several hundred games is still a small sample. It is luck. You do know, right, that you are also arguing with Robertie, our moderator. He just so happens to also be a world-class backgammon player who has won major tournaments. If he says you are blundering with the moves you make, it would probably be a good idea to stop arguing and take it seriously. He certainly knows what he is talking about when it comes to backgammon.

As an alternative, suppose it is not just luck. You say you have won 70% of the time with your blunders. What percentage of the time might you have won if you had made better moves instead? Maybe you would have won 75% instead. Just as in poker, a strategy having a positive EV does not make it the best strategy, winning 70% of the time with your moves does not imply that you could not have done better.

EnderEZ posts is so full of self ego. I don’t believe he wins 70% of his matches. The analysis is there to help you make equity plus decisions.
While I do agree blitzing generates more error cause the opponents on the bar and has no moves to make errors.
Blitzing and racing are generally done when one’s ahead in Pip count. Priming and contact plays are +EV when behind If you are generally blitzing then it’s too one dimension play

I kept track of 3,000 bring ins and ran about 10% under the odds on each pip. Small sample size? Maybe, but if that was the issue it wouldn't have been every pip. Games and matches are frequently insanely one sided. I have no idea why they have done this but either their RNG is ****ed up beyond belief or they are intentionally running an algo to make play more "interesting".

I'd like to comment on the dice on Backgammon Galaxy. First a little background. I've played live for many years. I've played in live tourneys in Vegas, Chicago, and Denver. I've placed in a state tourney (money and trophy). I've played a lot on sites like Gridgammon (old Gamesgrid), Gammonsite, Pogo, and now Backgammon Galaxy, so I have a lot of dice generators to compare.

First what I like about Backgammon Galaxy. I like the rating system in that you don't lose points (even if you lose the match) if you're error % is lower than your opponent (although there's no explanation of how the error rate is determined). For example, does it rate you down if you don't cube at the very end just because you're being nice and don't want to appear rude? The graphics are nice and neat too.

Before I comment on the dice on BG, I'd like to state one thing: First, all computer generated dice are pseudo-random at best. They are NOT random (the dice on BG are no exception). No computer generated / algorithm-based dice are random.

This has been proven by a computer geek who beat the Keno game in Atlantic City (He worked for the gaming commission in Vegas before he went to jail for this). He unequivocally demonstrated that no dice, number or card generator can duplicate true randomness. There will always be a pattern. And if there's a pattern, it's a certainty that someone can figure out how to electronically manipulate it and exploit it.

I haven't seen any bots on Backgammon Galaxy, so not saying anything about them. Just making a point on the bots on other sites. People there say the dice are random and the bots don't cheat. I have undeniable and definitive proof that the bots cheat. Don't believe me?

Here are some examples from Gammonsite:

1. I'm ahead in a match to 9 points (5-4) with one of them. During our next game it cubes...I take. Right then a
window pops up saying there is a sychronization error and we must exit. When I come back, the bot is ahead in
the match 8-5 and goes on to win the match. The game we were in was far from over and not only did it CHEAT to
take the game, but it gave itself a gammon and 4 points !!!

2. April 5th, 2022: Match with a bot. Score was 0-0. Pip count was about even. The bot cubes...I take. Once again, a
window pops up saying there is a sychronization error and we must exit. When I come back to the game, the bot is
now ahead 4-0. This is the second time it gave itself a gammon win with the cube on 2 before the game was even
played out. Totally corrupt.

3. I'm in a back game and the bot is getting ready to bear off. It rolls a 5-3 and by the rules has to take the whole roll
if possible. It can't move the 5 first because I'm blocking that point. It would have to move the 3 first and then
the 5 to bear off which would leave a blot for me to hit. Instead the bot acts like it can't move thereby not leaving a
blot. What a CHEATER !!!

4. And what about when the bots start rolling so fast, you can't even see the dice roll...and then inevitably, they fill in
their inner board before the dice slow down to normal. They should try to make their CHEATING less obvious.

These are only the undeniable proof that they cheat. There are many more "anecdotal" instances where their cheating which would be "arguable, but they still don't pass the smell test...like when a bot rolled seven double 6's in a row. Yes, possible, but c'mon. I've never seen that happen ever in all the years I've played.

So let's put aside the notion that the dice (on any site) are random and that bots don't cheat. Simply not true.

Now for Backgammon Galaxy Dice. OMG...of all the sites I've ever played on (and all of them have suspect dice), the dice on BG are the absolute worst of them all. I've read the comments from Bill Robertie. He's a great player. I even have one of his books. But to say the dice on BG are normal...what planet do you live on Bill?

Any good player knows that in a match of any length (7 or above) a player rated hundreds of points below you really has no chance. On Backgammon Galaxy, I play anyone regardless of rating (some do not). I've been shut out in 7 to 9 point matches with some who are 800 rating points below me because they dice were one-sided the whole match, miraculous rolls, etc. etc. This wasn't just an anomaly. It didn't just happen once but several times. I can hear the arguments...I must not have played well, but that simply doesn't hold water. Any good player will admit that this would never happen in real life. Yes, some luck is involved and in short matches, it would be possible, but never in a long match and never so many times. And how is it that sometimes when your error rate is much better (lesser) than your opponent and it shows you also had more "luck", but you STILL lose the match? This has happened there too.

Now for doubles off the bar. Arggg. I started noticing that when I put my opponent's checker(s) on the bar, the rate that they get off with doubles is astounding. Just to be sure, I tracked 12 matches. With any roll, bar or not, doubles should be rolled about 16% of the time (6 out of 36). Of those twelve matches, only two matches came around that normal percentage (14.2 % and 12.5 %). Here are my opponent's doubles off the bar percentages (usually the exact ones they needed) from lowest to highest in those other 10 matches: 33%, 38%, 40%, 44%, 45%, 50%, 50%, 60%, 60%, 67% . I know...small sample size. But it keeps recurring...I just didn't track it after that but it's very noticeable.

And please don't tell me not to play there if I don't like the dice. I play there and on other on-line sites because it's easier to find games and play a lot compared to live competition (which I still go to). I would just say to anyone to have fun with the on-line servers and take them for what they are...a poor substitute for the real thing.

I think you’re way off on the idea that a difference of a couple hundred Galaxy rating points means the better player should always win. A couple hundred rating points is only a few PR points, which (depending on the match length and exact PR difference) would only make the better player something like a 60-80% favorite. This might be the source of some of your frustration.

Well, I didn't say a couple hundred rating points difference. There were several matches where the difference in rating was over 800. If the ratings are true, I'll take a bet that I beat a player rated 800+ below me 99% of the time in a match of 7 points or more (probably 5 points as well). They certainly wouldn't win repeatedly and wouldn't shut me out as on Backgammon Galaxy.

Again, I think you’re way off here, and if you’re serious about wanting to make this bet feel free to propose a way to test this and I would be happy to give you action.

Anecdotally, the top players on Galaxy happen to be about 800 points above me, and I certainly win more than 1% of 7-pt matches against players in this range who have an average PR of around 3. (Heck I also win way more than 1% of 7-pt matches against XG which has a PR of 0).

You're way off base here. A couple of weeks ago I lost a match to a player rated 1300 points below me. The score was 0-11. In backgammon, such things happen. (And they also happen in live play as well, although there you have to estimate the skill difference.)

Sgt_Pepper,


There is no evidence at al the dice on Galaxy are "suspect" or "rigged" or "designed" and your comments certainly did not offer any evidence.

You said that luck is never involved in a long (9-point) match and that's not true at all.

Yes, it's very possible, and frequent, that you can have a better error rate AND more "luck" and still lose the game/match. (A simple example. You roll a fortunate dice roll or two, to turn your highly unfavorable position into what might now be a 50-50 game. You will be awarded "luck" for those rolls, for that turn of events... it was exactly what you needed to get you out of a jam... but now this game is, in this example, still a 50-50 game, that you end up losing. You played better, had better luck... but still lost. Happens all the time.)

Your post is one I've seen a hundred times before, on dozens of different websites and forums and servers over the past 20 years, going back to the days of FIBS.

You "feel" the dice are wrong.. but you have no actual dice rolls nor any actual evidence to back up your feeling.
You "feel" the dice are suspect... without providing anything for anyone to analyze or evaluate.
It is very likely you are remembering all of the times your opponent "got lucky..." while conveniently forgetting all of the times when you yourself were on the favorable side of a series of rolls.

Your sample size (as you mentioned yourself), is very very small.

You offer no explanation as to

1) what advantage(s) would accrue to Galaxy from doing so?
2) how (the exact method by which it would be achieved)?
3) which player would be favored in any given circumstances... and why!

Why is it that NONE of the top backgammon players who play on Galaxy, many of which are the best in the world, support your claim of suspicious dice?

The truth is, Galaxy has absolutely NO incentive to create a site with cheating/suspicious/non-normal dice. In fact, the opposite is what is true.

The time to believe in something, anything at all, is after there is sufficient evidence and justification to warrant that belief. Not before. Your belief in bad dice is completely unwarranted and unfounded, and based upon a combination of ignorance, small sample sizes, or otherwise bad data.

Show us some actual evidence to support your claim! Please play a sufficient number of matches, gather a proper sample size, and run a proper dice distribution analysis. When you do that I'm confident you will discover the dice distributions has converged towards expected, and you can see the Chi-square test for randomness will show no indication that the dice are nothing but random.

The real problem is that most people have no clue what random dice should look like. I once had a stats class where the professor assigned the class the task of flipping a coin 10000 times and recording the results. Inevitably, some in the class would not bother to actually flip the coin, but instead would just write a “random” sequence of “H” and “T”. Inevitably those students got busted. How? Well, most would never think to include a sequence of 8 to 10 heads or tails in a row, but in the true random data such sequences are almost inevitable. Someone who thinks a bg site has rigged dice would likely look at someone who actually generated the random data and say “That coin is rigged!”

Real random data contains all kinds of non random looking sequences. Humans evolved the ability to look at seemingly random data and find hidden patterns that exist— that is of great survival advantage. We also unfortunately have the converse ability, the ability to look at actually random data and find patterns that do not really exist. I suspect that this ability is the cause of most, if not all the “This site is rigged!” complaints.

One of my favorites anecdotes re: randomness is that when the iPod shuffle was introduced it did in fact randomly (pseudo-randomly, if you want to be precise) shuffle through your entire library. People complained that it was not "random" because they would get stretches of the same artist and that they felt the randomness was broken because of that. To assuage these people, Apple changed it so that the same artist wouldn't come up consecutively. In other rules, they broke the randomness and introduced rules for the express purpose of SEEMING MORE RANDOM.

To Edward Collins:

First you write that I said there is no luck in matches of 9 points or more. There is some luck in every match. I never said that. I was just making a point (that you would probably agree with) that the luck factor is diminished in longer matches when playing an opponent rated several hundred point below you. And I spoke sometimes in general terms about the dice from all dice generators on the various sites. Regarding that, I absolutely proved with facts that the bots on some sites cheat (not on BG...I haven't seen any bots there).

But I stand by the fact that no dice, number, or card generator can produce true randomness. This wasn't proven by me, but by a computer geek in Vegas who conspired with a friend of his to beat the Keno game. He got caught, and later confessed how he did it...because he was able to detect a pattern. Much more to it than that to his proof, but that's the Reader's Digest version.

I play BG on several sites. If they all were the same as far as dice randomness (which again, is not possible), then I shouldn't notice any major difference with the dice on Backgammon Galaxy and other sites. But I absolutely do. Again, hundreds of matches and the amount of doubles rolled when I have my opponent on the bar is no less than astounding compared to real dice odds. Yes, I tracked a small sample size, but have noticed the same thing in match after match. I wouldn't notice the doubles off the bar if they came in around 16% of the time as they should (yes, I know, that's not absolute, but in the long term it should play out that way and certainly not the other way around).

Anyway, we each have our opinions and that's fine. I provided some facts that cannot be disputed (those bots cheating). People can say the dice generators are random. I say they're Pseudo-random at best and I think that was proven by that guy who beat the Keno game in Vegas.

Anyway, just wanted to correct you when you wrote that I said there was no luck in longer matches. Never said that and never will.

This thread has the most intelligent arguments I've seen about the dice on BGG and that's why I signed up.
A few points I wanted to add.
A. Timing.
Examining the distribution of numbers on a large scale (thousands of rolls) is significant to understand if these are generated by "random" dice. What people miss is the possibility to provide a distribution very close to the "real life" roll- and cheat comprehensively.

Cube numbers are like words. Each of the numbers has value but the timing and order in which they are "said" is by far more influential.
In any series of rolling dice that is large enough, you can "take care" of a distribution to looks random but create a total different meaning for each of the players. i.e- provide a high ratio of 6-6 and 6 that will appear in most roles when player A is on the bar and still "normalize" the distribution between the two players.
This is part of what happens in Backgammon Galaxy in most of my games.
B. "But why?!"
Like Instagram, Broadway shows, gambling and porn sites, Facebook, water park, stand-up shows etc- everything is part of a human activity called entertainment.
The logic in the algorithm or managing the user experience in the entertainment businesses, assumes that he knows he doesn't really need it, it's just about killing time. The more time the average user spends on the site, the more the company's value increases.

The thing that makes people love entertainment is strong feelings like anticipation, tension, disappointment, emotional explosions and relief etc. This is why the cube on BGG generates so many users who thinks (or sure ) that it's a rigged cube. It's meant to provide entertainment- not mathematic accuracy.

The main argument of the players who are sure that there is no cheating in these systems simply disconnected from business reality. They don't know how to explain why they set up the system for hundreds of thousands of $ (or more). Anyone who thinks that the owners of Backgammon Galaxy invest money just to supply playing grounds for the players and the business plan was based on selling backgammon boards should be examined by a psychiatrist. The goal is based on the same entertainment logics of all great entertainment based businesses.

C. Transparency
The BGG user interface is the best I've seen. The principle of a point system that measures game quality in addition to the result is an innovation, but there is really no transparency. The log files are arranged in a way that makes it difficult for a player to download them all for analysis with external tools and there is no explanation of the rating/ quality score. It looks like transparency but it's not. Again, the data exists and the option to give the users a zipped txt files of existing information is a joke because of the low storage and development costs. it's in the D.B all ready.
I still play there, same username but know it's fake dice and I've learned to predict dice rolls pretty well.

-
I have played in the last 20 years on more than 10 different platforms of backgammon games. In all of them, without exception, the dice roll isn't near a a realistic game dice. I have nothing against BGG in particular. I just hate cheating and misleading in general.

mlafeflon

I do not know whether or not BGG has random dice. I strongly suspect that they are random, or at least as close to random as one can get. Your entertainment factor can be achieved by actual random dice just as well as by rigging them. The advantage of random vs rigged dice is that non random dice will inevitably be detected and players will quit playing. Also you do not need to pay a programmer to develop a random dice generator like you would to develop a rigged one.

I have heard the timing argument before. What I have not heard from anyone proposing it is a testable, specific hypothesis of how the sequences are non random while maintaining an overall random distribution. All I ever hear is generalizations about how it is possible that a non random order can still generate the right overall distribution. While this is true, such a non random order would still be detectable. For example maybe 66 is rolled at double its normal probability after a 31 is rolled and at half its normal probability after a 43. That would be a specific testable hypothesis about how a non random sequence could occur in an otherwise random distribution. It is also easily tested - look at a hundred thousand 31 rolls and a hundred thousand 43 rolls and see how many times the next roll was 66 for each.

That is just a made up example. If you want anyone to take your claims seriously then propose your own hypothesis, look at lots of data and see if it actually pans out. I have never seen anyone claiming rigged dice actually do so.