Dice rolls seem very off in Catan

[TheJeneral]

I've been playing in the Catan arena, and because it is a game that relies on dice roll likelihood, I pay attention a lot to the spread. The last couple of days, the dice rolls seem way off. I looked back over my last 5-6 games and not a single game falls within expected patterns for dice rolls, but it didn't used to be this way. Anyone else noticing this? Should I report it as a potential bug? It's throwing off my selections and I'm thinking I should just choose based on tiles and ignore the numbers (except 12 and 2 which do seem to be mostly uncommon as expected).

Would love to hear others' thoughts!

[buu441]

Please show us the data you collected and the statistical tests you performed to arrive at this conclusion. If you findings are indeed right and there is a statistically significant deviation from the expected behavior, this is a game breaking bug that needs to be addressed urgently, many lives could be on the line.

[RobertBr]

I pay attention a lot to the spread.

Its amazing how people say the exact opposite of what they mean.

You don't understand probabilities and because you are using your lack of understanding as a basis for looking at a problem are verging on joining a small group of conspiracy theorists who regularly turn up on the forums.

[JohnnyCyclops]

Mods - please reassign this to the dedicated Catan forum where the OP will find a number of similar posts dating back quite a while, plus a whole host of polite rebuttals asking those players for actual analytic data.

[RICK7425]

dice are way out of wack. just played a game with over 90 rolls and not a single 10 roll. what the heck!???? Game #632313035

[Romain672]

dice are way out of wack. just played a game with over 90 rolls and not a single 10 roll. what the heck!???? Game #632313035

49 rolls. It's 1.4% (remember you could do that claim for any of the 11 numbers). Far from unlikely.

[bitguru]

I played a game an in the last five hours or so where 5 never rolled, which is even less likely. (5 has probability 4/36, while 10 has probability 3/36.) I don't know how many turns that game had, but if it were the same 49, then that would happen only 0.3% of the time.

But that's not what caught my eye. In the last few games there have been a lot more 2 and 12 rolls than I would have expected. Maybe twice as many. That caught my eye.

Also I think there was once three consecutive 11 rolls. (There was also four consecutive 8 rolls, but that's not quite as unlikely. Four consecutive 8 is 1 in 2700. Three consecutive 11 is 1 in 5900.) These also caught my eye.

But I'm not saying there's anything wrong with the dice-rolling algorithm. Weird things can and do happen, and I've only been on BGA for two days.

I might want to look at the code (if it's even available for looking) to see how they handle meshing powers of 2 with powers of 6, but just because I'm curious. (What I mean is if you ask for three random bits, you can assign 6 of the possible bit combinations to different faces of a die, but then what do you do with the other 2 bit combinations? Simplest is to throw them out and ask for a fresh set of three random bits, but is that the best way? This is usually taken care of by the randomness library, not the app, though.)

[Jellby]

(I know you don't mean these statements as "proofs", but let my replies be for those who think they are.)

then that would happen only 0.3% of the time.

So what? Do you imply that things that should happen 0.3% of the time should never happen? I see a contradiction there

there have been a lot more 2 and 12 rolls than I would have expected.

So what? Expectations are just that, expectations, not sure predictions

Also I think there was once three consecutive 11 rolls.

So what? Nothing forbids it.

(What I mean is if you ask for three random bits, you can assign 6 of the possible bit combinations to different faces of a die, but then what do you do with the other 2 bit combinations? Simplest is to throw them out and ask for a fresh set of three random bits, but is that the best way? This is usually taken care of by the randomness library, not the app, though.)

I don't know this game, but typically a game will just ask for a random integer between 1 and 6, and then another one, and that's it. Not a lot of room for mistakes.

However, there are known cases where the developer has apparently done the "wrong" thing (https://boardgamearena.com/bug?id=130393).

[euklid314]

I might want to look at the code (if it's even available for looking) to see how they handle meshing powers of 2 with powers of 6, but just because I'm curious. (What I mean is if you ask for three random bits, you can assign 6 of the possible bit combinations to different faces of a die, but then what do you do with the other 2 bit combinations? Simplest is to throw them out and ask for a fresh set of three random bits, but is that the best way? This is usually taken care of by the randomness library, not the app, though.)

The game developer once posted the relevant code fragment. It was
die1 = bga_rand(1,6)
die2 = bga_rand(1,6)

If you read BGA FAQ you will learn that bga_rand calls the cryptographically safe PHP function random_int(). This RNG function is of the highest mathematical standard and is public domain. If you find a problem in this PHP function you will get worldwide media attention.

As you see via the simplicity of the source code quoted above that it is highly unlikely that one can possibly mess it up on BGA side. Since nobody has ever shown data to suggest any problems with the RNG it is safe to assume that it works as it should.

[bitguru]

So what? Do you imply that things that should happen 0.3% of the time should never happen? I see a contradiction there

No, you completely misunderstand me. I am saying that these things caught my eye, yet I don't doubt that the dice-rolling algorithm is correct. I'm not sure how I gave you the wrong impression.