Dice Algorithm inBackgammon Seriously Flawed

[dsyring]

Just completed two games of backgammon in which between the two of us there were more than 20 doubles, including one run of 6 doubles rolls in a row. This happens often in playing, and I’ve conversed with many players as we roll that doubles seem to happen far more often than is statistically likely. This undermines the experience of playing...becomes much less about the skill of the players and more about the silliness of the doubles. Is there any way you might consider revisiting the programming on that dice algorithm to make it appear more like the randomness of actual dice?

[Jest Phulin]

Please check out the backgammon-specific forum.
https://boardgamearena.com/forum/viewto ... 71&t=14562 (among several other threads)

This current forum, "Discuss about board games", is in the "Non-BGA topics" category. A flaw in the random number generator being used is definitely a BGA topic.

That being said, the short answer is that anyone who has ever done a statistical analysis of the dice rolls has never found a problem, and everyone who has claimed a problem has not been able to provide statistics (only "it seems like" anecdotes) showing a problem. In Backgammon, Can't Stop, Stone Age, or any other game where dice have a strong influence on the game.

And, 6 doubles in a row happens statistically 0.00214% of the time, or once every 46,656 rolls. The average game takes around 50 rolls. So we're talking more often than once every 1,000 games. As I'm writing this post, there are 854 games in progress. So, about an 85% chance that one of them will match that feat of 6 doubles in a row.

Yes, the event is uncommon. But it has a lot of chances to happen. Those two factors will tend to result in it happening.

[dsyring]

Thanks for your thoughts, and apologies for posting in the wrong place. I guess it’s a function of the thousands of games played on the site, but in any given game my experience of playing is much more likely to be “spoiled” by excessive doubles than when I play in person with actual dice. It makes me glad that the real “reward” for playing on BGA is connecting with people all over the world. It’s less about playing the game and more about the people.

[Een]

It makes me glad that the real “reward” for playing on BGA is connecting with people all over the world. It’s less about playing the game and more about the people.

That's a really great point of view

I guess it’s a function of the thousands of games played on the site, but in any given game my experience of playing is much more likely to be “spoiled” by excessive doubles than when I play in person with actual dice.

A player posted in another thread a link to a video that I found fascinating: viewtopic.php?f=245&t=16243&p=62389#p62389

It explains how bad humans are at "feeling" randomness. Which probably means that as admins, we are condemned to see this topic come up again and again. Which actually means that random functions in use are good at being random, because if they matched human expectations, they wouldn't be

[Theacratis]

doubles happen consistently when blots are on the bar waiting to come back into play. i would be rich if i could bet on the number of times this has happened. I shouldn't be able to bet on so called randomness and win consistently ! I Haven't checked the statistical accuracy of this situation but it seems way beyond coincidence. Could the dice programme and algorithm be broken. i would have thought a simple random dice programme should be easy to implement. I guess its the same for all but that's not the point, something feels fundamentally flawed in this programme.

[CaractacusPots]

Ah yet another thread about the rather obvious lack of randomness or game skewing in relation to Backgammon.

Seems there are a lot of people who according to the site fan boys are just imagining things ! lols

I've only thus far played 38 games of BG here and won 19 of them

In that short space of time and as a very long standing BG player of many years who has studied the game, studied books and strategies of early opening moves, mid-game moves, back game strategies, blitzes and bearing off optimisation I can already tell that the system is just a farce. The dice throws may well be randomly generated but how those generated numbers are used, exactly when in the game they are given to you is likely the point at which randomness goes out the window.

In relation to the doubles OP here is talking about and blots I can only support him. The primary issue is not the specific dice rolls, it is the dice rolls happening at just the precise moment in the game needed to turn the game around and re-balance the game for the losing underdog/poor player. It has happened with monotonous regularity to the point that I just resign now when it's obvious that the system has chosen to make the other guy win. This isn't sour grapes. I understand dice and I have years of games under my belt.

In just 38 games I have already witnessed at least 4 games, possibly more, where having been knocked off and opponent only having the 6 point covered I am dealt double 6 so that I can not return. This is in dice terms merely a 1 in 36 chance roll but the fact that it happens with such frequency at the precise time you have a blot and opponent only has the 6pt covered is the problem. In one of the games my opponent made his 1pt on the roll after my 6-6 meaning I could then not get back on the board if I threw 6-6, 1-1 or 1-6 and sure enough the system gave me 1-6. So a 6-6 followed by 1-6 at precisely the point in the game where it was needed to keep me off the board and turn the whole game round to the poor player who at that point was way way behind.

I see similar squiffiness in the game Can't Stop, another dice game. You win by completing 3 columns but the game is always balanced so that it is a close thing. I have never yet completed 3 columns with my opponent not completing a single column and the moment you get 1 or 2 columns ahead of opponent your ability to string lots of dice throws together seems to go out of the window.

Analysis of dice throws will never show any anomalies on sites that balance outcomes because the truth is that the dice throws ARE random. But on many sites the throws are generated BEFORE the games actually start or in card games like poker the cards are all generated before the hand begins. That means systems can then allocate those random dice throws or cards at any point in the game as required. If you have a genuine 6-6 throw in your set of randomly generated throws you could be given it in the first move (for a great solid start) or at the point you are knocked off and opponent only has the 6pt blocked.

Had this debate on numerous threads now and it's a pointless one because the sites and employees will never give you the truth if they are indeed tweaking outcomes and you can never be confident that any source code they provide is the full and actual source code being used.

In the end it is what it is. Every gaming site imo balances outcomes to prevent strong players dominating newbies and weaker players too much. Very especially in online gambling sites like poker and so on. Google Play store is absolutely replete with Backgammon apps and every one of them has player reviews all complaining of the same rigged re-balancing that goes on to try and keep everyone happy. There are no exceptions.
There simply does not exist, to the best of my knowledge, any totally genuine online backgammon or online poker site that is not in some way tweaking outcomes. It's a real shame because I love the game so much that I would probably pay to join any site that was completely genuine.

In the end your best bet is simply to download the latest copy of Jellyfish or GNU which are free and play a strong game of BG.

[Tisaac]

What employees are you talking about ?
You are really going to give statistical analysis over your 36 games ???? 36 ???
Come on, play at least 100 gamesif you want to starting.building data evidence. (Or find some piece on code on the bga studio that's in favor of your claim )

[aesche]

Tl;dr

[euklid314]

CaractatusPots: You are exactly falling in the trap mentioned in the video given above (https://www.youtube.com/watch?v=tP-Ipsat90c).
You are searching for patterns (we humans are very good at that). Always if you find something that is very rare, you are convinced that something is fishy. Whenever you post it here on the forum and you are told that your observation is nothing special, then you find something else that you find improbable.

In this thread you have postulated something that is easily verifyable, namely entering from the bar against a 1point board. You are claiming that you fail to enter more often than you should whenever your opponent has blocked a single point. We agree that this should happen (on the long run) every 36th time.

Of course you have played far too few games to do a statistical analysis, but you could take the time and nevertheless crunch the numbers for your 38 games.

You will have very few occurrences in your games but nevertheless I am interested on how often you entered and how often you failed.

We humans tend to forget the many instances where something happens that we expect. We only are surprised if something happens that we do not expect. The problem is that intuitively you are expecting a fan on a one-point point not to happen at all - perhaps once a month if you play a lot. But its probability is approx 3%!

[Romain672]

I see similar squiffiness in the game Can't Stop, another dice game. You win by completing 3 columns but the game is always balanced so that it is a close thing. I have never yet completed 3 columns with my opponent not completing a single column

In my imaginary world, I often played some game of rock/paper/scissors in best of 21. The first one who won 11 rounds win the game.
And yet, I never win or loss 11-0.
My imaginary world is fishy for sure.