The thing is you won't remember when your dice have favored you in incredible ways. Only when you lose is when the whining instinct comes to surface. Buckle up. Playing backgammon online is a choice.

Since the last post on this was two months ago, I guess the controversy has died down. (I'm sure it hasn't)

I think I've seen all of these people complaining about rigged dice on the poker sites I play on, complaining about bad beats.

It's really annoying hearing people complain when playing live. At least on BGG you can just ignore the comments and you don't have to listen to them.

Playing for money, it's usually the people that complain that lose the most. So, we put up with it.

The best players in the world play on BBG. IF anyone on the list of 32 Giants think BGG is rigged, I'd like to hear from them, here.

FWIW, I've never met a good player who thinks the dice on Galaxy are rigged.

I’d like to see BG Galaxy be transparent about the algorithm used for the dice rolls. Try Backgammon NJ HD if you want to see transparency about the dice algorithm from the site creator. Much better site to learn the game as well.

yeah whenever i think of transparency and things on the up and up i definitely think of things associated with new jersey

Now, after playing around 1300 matches (usually of 7 points or more), I still stand by what I wrote earlier. Again, I've played backgammon for a long time, live and on-line. I've placed in a state tourney (money and trophy).

It's frustrating playing on Backgammon Galaxy because the dice are so ridiculous. And don't tell me not to play here. I've already explained why I play here and on other sites. So I will reiterate.

I'm literally getting tired of the rate my opponents roll doubles when I have them on the bar (sometimes multiple checkers). I constantly hear myself screaming in my mind that "They rolled doubles off the bar and the exact doubles that they needed". I tracked this early on and wrote about the results in my other post. Yep, small sample size, but now, after 1300 matches and thousands of games, you'd think that the dice would begin to gravitate back to normal dice odds, but it doesn't happen. Doubles off the bar (or any doubles) should be rolled about 16% of the time. But off the bar (which is more noticeable than other game doubles), they are constantly coming in at a incredible rate (as in not credible), upward of 70% of the time sometimes (but even when it's 35-50% of the time, that's still way too much...more than double what real dice odds are).

Now, when I'm on the bar, I constantly (the majority of the time) have a difficult time coming in against 2 and 3 point boards and many times in critical moments. With all the rolls that would get me in, this shouldn't happen with this kind of frequency. Enough said.

Now about the error rate and the luck factor. I've played matches where my opponent was more than incredibly lucky the whole match and the the analysis showed that the luck factor was either around even or that I had more luck which was absolutely false, unless they count bad luck as luck. Again, no way to determine how the luck factor is figured.
The error rating is questionable too. Do they rate you down (error) if you fail to cube on the last few rolls with few checkers on the board and you're going to win the game but are just trying not to be rude??

Another anomaly happened to me on Backgammon Galaxy. I won a match and when the analysis screen showed up, my icon was glowing (pulsating) yellow. Then I saw that my rating went down 100 points. Again, this was after I just won the match. I wrote to support, but crickets...no reply. Has this happened to anyone else here?

I play anyone regardless of rating, so when I lose to a lower rated player (sometimes several hundred rating points below me), and if, by the "analysis" their error rate was better, I lose a ton of rating points. Some people posting here are talking about all the highly rated players. I'm sure they are very good...maybe even world-class, so I'll give them credit for that. But what I'm seeing in my experience is that, as a general rule, those highly rated players won't play players rated well below them, so they are hedging their bets and limiting themselves against those big negative rating swings. Nothing wrong with that and this isn't a blanket statement. I'm sure some of them play much lower rated players, but in my experience, most won't.

No need to reply to this. If you think the dice are "normal" here, that's fine. I won't change your mind and you certainly won't change mine. I play on-line because it's easier to play a lot. I just take the on-line games for what they are...a poor substitute for the real thing.

Post some matches

Seems like you should stop playing on Galaxy and play on some other site that you like better.

Disclaimer: I shouldn't post that because it will give the rigged camp some ammunitions. I am not at all in that camp.

Nevertheless, I was playing a few days ago and I was on the bar against a 2-point board (6th and 4th). I then rolled 66, 64, 64, 64 before entering back.

I genuinely laughed at my (lack of) luck, shrugged it off and went on to win the game.

I had a funny game a few years back where I "supergammoned" XG bot

It had 3 on the bar while I was bearing off, and the bot kept failing to enter. Finally, I rolled a big double, leaving 4 pieces on 1.

Bot proceeded to roll 1-1 and then I rolled another double.

A human opponent would've probably quit me for good right there, but the bot somehow gathered itself and very quickly was up many points on the session

Instead of berating and telling me not to play here, why not try to prove me wrong. I would accept that.

The program can calculate the best moves out of thousands of possibilities in a split second. How hard could it be then, to do simple math calculations to show in each match, for example, the amount (percentage) of doubles rolled off the bar in the rolls it takes for each player compared to real dice odds? Also, how many percentages of doubles in general in the game for each player for the amount of rolls compared to real dice odds? Yes, it will vary, and in any one match it's not a tell-tale sign, but if it continually comes in at a much higher rate than normal dice odds over the long haul, then maybe something is wrong with the dice generator. If it doesn't, you will prove me wrong.

It should also be easy to calculate the percentage for each player of coming in from the bar against various boards during the match (2, 3, 4, or 5 points covered) compared to real dice odds. For example (we'll assume one checker on the bar) , against a 5 point board, only one particular number gets you off the bar, so 11 in 36 or about 31% of the time. Do that calculation for all the various boards in a match and compare that to real dice odds over the long run.

The end of game analysis could show: COMING IN FROM THE BAR AGAINST (one checker):
5 Point board: REAL: 31 % YOU: ???
4 Point board: REAL 55 % YOU ???
3 Point board: REAL: 75% YOU ???
2 Point board: REAL: 89% YOU ???
1 Point board: REAL: 97% YOU ???

Casinos rely on odds to make their profits and they never lose in the long run because the odds that they rely on always gravitate back to the real odds which are always in their favor. So if the dice generator truly puts out random rolls for two die, then the odds should gravitate closer to real dice odds in the long run, not consistently substantially different than that, correct?

And how about some transparency on how the luck factor is figured? Just say how it's calculated. What does the program look at? Again, are you rated down as an error, if technically you should cube at the end of a game that you're going to win in roll or two, but are just trying not to be rude? I would like to know because there have been matches where my opponent was not just lucky, but incredibly lucky the whole match, and yet sometimes it shows the luck to be about even or even the opposite which in the cases I mention would be 100% incorrect because it wasn't even close to that. I'm aware when I have lucky matches and when my luck is better than my opponent's. I also know when it's not as in the cases I've mentioned.

I haven't seen robot players on BG, but on other sites, people would swear up and down that the bots don't cheat, but I've proven unequivocally that, in fact, they do cheat. It's not my opinion. I've shown with several undisputable factual examples of them cheating (see my other post).

So instead of just showing contempt for anyone who questions the true randomness of the dice on Backgammon Galaxy, just prove them to be wrong with real calculations. If it shows the dice to be close to normal random odds, I'll acquiesce. If it doesn't, will you?

for anyone interested I'll (100% seriously) offer you the same action Marc offered earlier in the thread:

Are you prepared to lay that bet with those odds on the Android version of AI Factory Backgammon (current version 3.42)?

Make sure you talk to Jeff Rollason before you do.

I can call the die in certain situations 100% of the time. He's tried using white noise (i.e only call the cheaty bit of code if an RNG call between 0-10 is greater than 7) but he's got loose bits of code in there that cheat anyway.

You *will* lose if you take the bet.

Be careful. Let me know, I'll be happy to bankrupt you - funds into an Escrow account beforehand to stop you bailing out when you're about to lose.

You can see this for yourself, very easily.

If you weren't aware of it, XG, in addition to analyzing the strength of your moves, can also simply analyze all of your dice rolls for you and those rolled by your opponents. It will display, in a graphical format, the exact answers to your request.

Here's how to do it:

First, download as many of your Galaxy matches as you can. To do so, simply click the Galaxy ANALYSIS link and then click the DOWNLOAD icon in the far right margin. The match will be saved as a text file to your hard drive, most likely in your default download folder, wherever that may be.

Immediately after clicking a download icon, simply click the next match download icon in the next row. It only takes a second or two to click each icon. Eventually you will have to scroll down to see more icons. But that's it. It's just click - click - click, etc, scroll... click - click - click, etc, scroll... You can download your last 100 matches, for example, in probably not more than a couple of minutes. (Alas, I know of no way of downloading matches from Galaxy in batch mode. For now I think downloading is limited to one match at a time, as described.)

I suggest then saving all of these match text files you download to a new folder called, for example, Galaxy_Matches.

Start up eXtreme Gammon. Click FILE - IMPORT - BATCH IMPORT. Here you can import as many matches as desired, all at once. To do so, click CHOOSE and then navigate in your own folder structure to the match folder (Galaxy_Matches). Select all of the files in that folder and then click OPEN.

(Alternately, you can also drag and drop matches into the input box.)

Next, find and then select your Galaxy username among the list of players.

If you're just interested in analyzing the dice rolls, I suggest setting the Analyze Level to None or Very Quick. This will save a bit of time. Obviously, if you're interested in analyzing your performance, you might wish to use a stronger Analyze Level, such as World Class.

Click the Create Profile button and check the Override Previous Analysis for this player.

That should be it. Click the START button to begin. The program will then analyze each match, one after another. Go make yourself a cup of coffee. It should only take several seconds, at most, to analyze each match, although the exact time will depend, of course, on the Analyze Level you choose and the length of each match. The program might be finished before you return to your desk.

When the analysis is complete, from the top menu click PLAYERS - SEE PROFILE RESULTS. Select the DICE tab.

Here you can see your total number of dice rolls, and that of your opponents, from the matches. You can see the percentage of those rolls that were Doubles, the percentage of times you entered from the bar and the Expected Percentage from the bar. You can see the Doubles percentage once the game was considered a race, etc.

Click the DETAILS tab for even more dice-related data. Here, for example, you can see exactly the number of times (or percentage) you rolled 6-6. You can see the number of times you rolled doubles in a row. You can see the number of times you entered from the bar against a 1-point board, against a 2-point board, against a 3-point board, etc., and what the expected number should be.

You can do all of this for your dice, your opponent's dice, or for both dice!

With the graphical format displayed, it is very easy to see how all of these numbers (or percentages) compare against what the Average or Expected number would be. (Notice the Average line that is running down the column.)

Clicking the ANALYZE tab will give you the results of your dice against the Chi-Square text.

Of course, if you don't have XG, I'd be very happy to do all of this for you. It honestly wouldn't take me more than a minute or two. (I would, of course, need you to e-mail me your Galaxy match text files.)

This dice analysis will give you actual data to support your claim that the dice on Galaxy are the absolute worst. So far you've provided no actual data at all.

In a prior thread you said you tracked the number of doubles from the bar from 12 of your matches on Galaxy, and how your results did not seem all all valid or proper. (You admitted it was a very small sample size... and you also admitted that you didn't bother tracking it after that.)

I'm really curious if you've taken the time to actually import a decent size number of your matches into XG and generate actual dice stats as described above. (And by "decent size number," I'm just talking about 100 matches or so but if you have 1,300 matches to import, so much the better.)

"No need to reply to this. If you think the dice are "normal" here, that's fine. I won't change your mind and you certainly won't change mine."

-Sgt_Pepper 2023

To Edward Collins:

Okay, thanks. I don't have XG. I copied and pasted your analysis instructions to a word document so I can look at it another time. I wouldn't mind doing that. I would have to read your post a few times over. It seemed complicated to me so I hope I understand how to do that. And...that would be great if it could show all the stats that I mentioned. Not sure how it would be displayed but looks like you've done it before.

My e-mail is: Stonecub@comcast.net

I don't come to this forum very often, so e-mail is the best way to contact me. Then I'll have yours so I can send you that (if I figure out how to do it...lol).

Thanks again. We'll see what happens.

Pepper, I'm genuinely curious.

What is the raw data for your opponents rolling doubles off the bar? How many rolls total, and how many doubles, for your entire sample of games?

To Neil S.:

Well...I don't know exactly. I wrote to Edward Collins and gave him my e-mail. I don't have XP so he said he would analyze it for me. I tried his instructions and I only got so far on BG because the "download" button he mentioned doesn't show up on my screen. Maybe I'm doing something wrong but he never e-mailed me back.

Also, I was way off in my calculations of my opponents rolling doubles off the bar. It's far worse that I thought. I was comparing the rate they roll doubles to the actual rate to roll any doubles on 2 dice (6 in 36 or about 16 %). But...actually when they come in off the bar I almost always have at least 2 to 5 pips blocked, so the real percentage would be much lower than 16% (which would only apply to an open inner board with no points blocked...almost never happens). So, they'd have to roll not only doubles, but exact doubles to get out with some pips blocked as not all doubles would work. So, for example, if I have 5 pips blocked in my inner board (only one particular double would get them off the bar with a double) the true percentage to come off the bar with doubles is NOT 16%, but less than 3%. The percentage would increase a little with each less pip blocked. Even with only 3 pips blocked, coming in with doubles is only 8% chance. I can tell you, my opponents don't roll 92 times before they hit their particular doubles that they need. As I tracked, more than 30% of the time is common and up to 70% of the time is not unusual, so something is way off with these dice.

If I ever hear back from Edward Collins and can do the analysis, I'll let you know the results (I pretty much know them already without exact numbers). Again, I would be more than happy to acquiesce if I'm proven wrong, but so far no such proof has been forthcoming from anyone. Cheers.

Sgt Pepper,

Sorry, just now, today (3/25) I'm seeing your reply on 3/15, when you replied to my comment on 2/27.

(Like you, I also don't visit that forum regularly!)


I just e-mailed you, at the e-mail address you gave, and included a screen shot to help you locate the Download button on Galaxy.


-Ed

By the way, here's a sample output of the eXtreme Gammon analysis of my dice rolls, from 100 of my Galaxy matches. (I really should have included this graphic with my instructions that I posted about a month ago on how to generate this graphic. Duh. Not sure why I didn't think to do so back then.)

The graphic shows the detailed results of my 2,368 dice rolls in these 100 match files. It shows the distribution, the number of doubles in a row, the times I entered from the bar against a 1 point board, a five point board, etc.

Below you can change the display to show your opponent's dice or both dice, as well as the Confidence Interval.

In this display I chose to show the output as percentages, but you can change it to occurrences if desired.

As shown, nothing here from these 100 matches is outside of the expected norm.

I have been playing Adikus backgammon for some time. I can quests most of my players dice rolls because so many of the games are repeats of others I have played seemingly over and over.
First complaint is not being able to start the game. Very rarely do I start. Second is the amount of doubles all my opponent’s get. Some one after another. Hard to win a game with those odds but game after game this happing is unbelievable. Another repeat feature is I am always bumped. Always! And the I will sit on the bar for multiple rolls with my opponent having men open in his home court.
Something is not right at Adikus. They claim their rolls are random but they are not. Way to many perfect rolls for all my opponents and high rolls were as mine most common rolls are 1/2 or 1/3. Over and over again. I have had players watch my games and no one can believe the luck my opponents have. Unbelievable is what they say.
ADIKUS stinks in fairness. I’m sure these winners think they are so good but anyone would win starting then getting 5 or 6 rolls of doubles, never sitting on the bar for rolls, etc. They are full of crap.

I don't know anything about Adikus backgammon (I never even heard of it) but instead of just claiming the rolls at that site are not random, it would be much better if you posted the actual scores of your games there, to prove this. (Ideally, you would make the games available on some server or website, for all to review.) Ideally, of course, you would post a large sample of games, containing at least a few thousand dice rolls.

Only then will your claim have any merit.

Any idea why you are on the bad end of the rolls and not your opponent? Why does the site favor them?

You say you've been playing for some time. Not sure why you continue to play there if you believe the dice are suspect.

I'm a backgammon newcomer, but aren't dicerolls kinda irrelevant given that BG gives half the credit to the better scored player?

Agree 1000% :-)

I don't repeat what I posted on reddit, just some links.

The Official Backgammon Software Complaint Form: https://www.bkgm.com/rgb/rgb.cgi?view+546

Here https://www.bkgm.com/rgb/rgb.cgi?menu+computerdice you'll find a lot more articles about a lot of SW (it seems not to be maintained anymore, therefore most entries are old) and even "BG by George" was suspect to cheating and it plays really weak.

The only way to analyze random dice is is to collect and publish data. Everything else is irrelevant.