Complaint Regarding Dice Probability Distribution in Settlers of Catan

[parzi42]

Random number generation is pretty straight forward and every language i ever worked with had an implementation in their respective libs. I am pretty sure BGA did not reinvent the wheel, so to say. Sorry, but this is most likely just "bad luck" on your side.

[Jellby]

While there is no real issue with such artificial boosts when playing against bots, it is unfair and frustrating in multiplayer games against real people.

Honestly, that's the part I don't understand. What difference does it make whether your opponents are human or not? You don't mind losing as much as you mind someone else winning? Perhaps it would be best for you to assume that you live in a simulation and everything and everyone around you is a bot, you're the only real person. Now you can enjoy the game.

The rest is the standard "I have no idea what random means" rant.

[nik592]

The probability of a given number coming up on a d6 roll is 1 in 6, but that doesn't mean that if you roll a dice 6 times, you're guaranteed to get one of each number. There's a set number of distributions for any given number of dice rolls, and each of those is equally likely to come up.

It's like if you flip a coin 5 times and get all heads, that doesn't make tails any more likely on your next flip. In a set of six coin flips, the distribution of six heads is just as likely as a distribution of 3 heads and 3 tails.

But yeah, feel free to add your evidence to the bug report over there, if it makes you feel better. Or as robinzig says, stop playing Catan, and then random distribution won't bother you. Probably. Guess it depends what you replace it with. I'm not a fan either btw, and while luck has something to do with it, I also find player trading particularly frustrating.

3 heads in a row followed by 3 tails in a row, before anyone gets confused.

Not only 3 heads in a row followed by 3 tails in a row. Any 3 of each in a specified order. The order could be 3H then 3T, or HHTTHT, or HTHTHT. Those specific sequences are all as equally likely as 6H. What confuses people is that they don't understand how any one of those can be equally likely as 6H, and yet it's also true that an equal number of H & T in any arrangement is 20 times as likely as 6H.

Ed.: even -> equal, because the example was three each and three isn't even and that may sound confusing

I think Ceaseless's point is that if you're looking for ANY distribution of 3H and 3T, then that is more likely than 6H, because there are more combos that include 3H and 3T. Which is true. What I was saying was if you flip a coin 6 times, the distribution that you got (including the order) is as likely as any other distribution - whether that results in 6H or 3H and 3T. Much like the distribution of dice rolls over a game of Catan is as likely as any other distribution (including order) of the same number of dice rolls. Yes, there are fewer sets where 20 of the 30 rolls are all 4s, but that kind of distribution will still happen at some point.

I guess that's possibly the part where people misunderstand. Yes, there are more distributions where the numbers rolled follow a rough bell curve peaking on 7 (for 2 d6), so people believe "most" of their games should follow that distribution. When they get a few that don't, well, that's clearly evidence of RNG gone wrong.

[Ceaseless]

The probability of a given number coming up on a d6 roll is 1 in 6, but that doesn't mean that if you roll a dice 6 times, you're guaranteed to get one of each number. There's a set number of distributions for any given number of dice rolls, and each of those is equally likely to come up.

It's like if you flip a coin 5 times and get all heads, that doesn't make tails any more likely on your next flip. In a set of six coin flips, the distribution of six heads is just as likely as a distribution of 3 heads and 3 tails.

But yeah, feel free to add your evidence to the bug report over there, if it makes you feel better. Or as robinzig says, stop playing Catan, and then random distribution won't bother you. Probably. Guess it depends what you replace it with. I'm not a fan either btw, and while luck has something to do with it, I also find player trading particularly frustrating.

3 heads in a row followed by 3 tails in a row, before anyone gets confused.

Not only 3 heads in a row followed by 3 tails in a row. Any 3 of each in a specified order. The order could be 3H then 3T, or HHTTHT, or HTHTHT. Those specific sequences are all as equally likely as 6H. What confuses people is that they don't understand how any one of those can be equally likely as 6H, and yet it's also true that an equal number of H & T in any arrangement is 20 times as likely as 6H.

Ed.: even -> equal, because the example was three each and three isn't even and that may sound confusing

Yeah, any combination of a single set of flips has an equal chance of happening. My statement for 3 heads and 3 tails was since they already laid it out that way I wanted to sidestep people misreading it as any combination of 3 heads and 3 tails so named a specific one by choosing the order already written. Otherwise, the full generic group of 3 heads 3 tails put us right in the middle of the 7th row of Pascal's Triangle for the 20 possibilities you pointed out, AKA all the ways you can rearrange the 6 coins flips divided by both the ways to redundantly rearrange 3 heads and 3 tails.

In hindsight, I think it would have been better if I had further clarified and said any one of those orders had an equal chance of happening rather than only naming 1, which still would separate it from the idea of all of the combinations of 3 heads and 3 tails being grouped together.

[EuSuntGiGi]

Hello everybody and happy new year.

Devs please change the dice algorithm in catan, 6-8 dice have the lowwest priority in comparison to 9,5 ,4 , 11 and 2-12. And also randomize the dice , i've seens games with same dice 3-4 turns one after another. Its frustrating to see repetead dice

[Kayvon]

Game developer here (but not for Catan). When a random event occurs in a game, like rolling dice, the code asks BGA for a random number. BGA doesn't know how the number will be used in any way. It doesn't know what a "good" or "bad" number is. It doesn't even know which player(s) the number affects. All BGA knows is the code is requesting a random number.

What your reporting is simple bias in how you've viewed the numbers. Here's an analysis of half a million random numbers generated by BGA. The distribution is exactly what you'd expect.
https://boardgamearena.com/forum/viewto ... and#p59122

[Jellby]

Devs please change the dice algorithm in catan

Done. It should be fixed now. Thanks for reporting.

[robinzig]

Devs please change the dice algorithm in catan

Done. It should be fixed now. Thanks for reporting.

LOL

[Ceaseless]

Hello everybody and happy new year.

Devs please change the dice algorithm in catan, 6-8 dice have the lowwest priority in comparison to 9,5 ,4 , 11 and 2-12. And also randomize the dice , i've seens games with same dice 3-4 turns one after another. Its frustrating to see repetead dice

It's New Years, not April Fools, save that joke for 3 months.

[FrankJones]

No, the OP is correct. I have noticed something odd; whenever two dice are rolled, a total of 7 comes up more often than any other total. This should be fixed immediately.