change the dice generator.

Forum rules
Please DO NOT POST BUGS on this forum. Please report (and vote) bugs on : https://boardgamearena.com/bugs
User avatar
Jellby
Posts: 3545
Joined: 31 December 2013, 12:22

Re: change the dice generator.

Post by Jellby »

King Derf wrote: 09 June 2025, 21:11 Having one in 200000 events occur every game is not right.
But it is. A single sequence of 10 rolls has a probability of at most 1 in 60 million, and yet, every time you roll 10 times, you get one of those. Literally. 100%. Amazing!

You define your "unlikely event" after you've seen it. That's one of the most common pitfalls.
User avatar
ermocito
Posts: 9
Joined: 30 March 2020, 18:57

Re: change the dice generator.

Post by ermocito »

I believe the designers when they say their system is actually random.

I don't need proof of a flaw in the current dice system to make this simple request.

The official Catan app has a setting option where you can assure that every 36 dice rolls create the expected distribution, and another where you can experience this effect minus up to 3 of the possible rolls.

I request the game designers add such an option. Some players prefer the truly random system, but many players would prefer forced-distribution variation. If it isn't unacceptable in the OFFICIAL app, it should be acceptable here. don't need proof of a flaw in the current dice system to make this simple request.

And yes, I could and do play in the app. But this site has many more players and most games have several optional settings.
User avatar
cboers18
Posts: 3
Joined: 15 August 2020, 01:20

Re: change the dice generator.

Post by cboers18 »

I'm a mathematician and the randomizer feels wonky to me. I agree with the person who posted that the Inventors tend to come together and the Spies tend to come together (meaning that often when one player pulls a spy card, another player will pull the same card in the same round, and likewise for the inventor).

I wouldn't be surprised to have a game of all 5s if I play long enough. I expect unlikely events. But when one player settles 8, 10 and 3, and all three numbers are over-rolled, I start to wonder. The randomized also sometimes appears to have knowledge about the state of the game. For example, after building Aqueduct, my opponents each rolled 7 for the next 3 rounds. This dice behavior is familiar to me from other implementations of Catan, where the dice block the current winner. It does not happen uniformly on this site, but often enough that I pause and wonder.

I'm a mathematician with time on my hands. I would love to come up with an experiment to validate some of my claims. But I'm not a statistician. If anyone can design a good experiment for me to test the randomizer, I would love to try it out. (Something super easy to check from the end-game stats, like that the player with the lowest elo at the table tends to roll more 7s)
User avatar
cboers18
Posts: 3
Joined: 15 August 2020, 01:20

Re: change the dice generator.

Post by cboers18 »

https://boardgamearena.com/13/catan?table=690705614

Great example where one player settled only 3, 4, 8, 10, 11 and still managed to get an average of 1.5 cards per roll. My mathematician senses tingle when 3, 4, 8, 10 and 11 all happen to be over-rolled and 5 and 6 (my settlements) are severely underrolled. It's not weird to see a game with too many 8s. It IS weird to see a game with too many of all of one player's settlements.
User avatar
Jellby
Posts: 3545
Joined: 31 December 2013, 12:22

Re: change the dice generator.

Post by Jellby »

cboers18 wrote: 21 June 2025, 02:53I'm a mathematician and the randomizer feels wonky to me.
Then you should know that feelings are not the way to evaluate randomness.
It does not happen uniformly on this site
I guess it happens randomly?
cboers18 wrote: 21 June 2025, 03:41 https://boardgamearena.com/13/catan?table=690705614

3, 4, 8, 10 and 11 all happen to be over-rolled
There were 60 rolls (counting re-rolls due to easy start)
3 was expected to be rolled 3.33 times, it was rolled 4 times. Wow.
4 was expected to be rolled 5 times, it was rolled 4 times. I don't call this over-rolled
8 was expected to be rolled 8.33 times, it was rolled 13 times. I grant you that one
10 was expected to be rolled 5 times, it was rolled 5 times. I don't call this over-rolled either
11 was expected to be rolled 3.33 times, it was rolled 4 times. Wow again.

As usual, I see only selective (and false) reporting
DivergentRalph
Posts: 7
Joined: 04 May 2025, 11:42

Re: change the dice generator.

Post by DivergentRalph »

gmrsmrt wrote: 07 June 2025, 11:25
DivergentRalph wrote: 07 June 2025, 11:07 Hi, I believe this topic needs investigation and a lot of data in order to come to a conclusion. Otherwise it will be people pointing out bizarre things and others saying it's just chance.

Right, so gather more data...

In the example you spotted you should add the probability of having any other number (2 to 12) to exceed the expected frequency by
a probability smaller or equal to the one you observed. And also the probability for each number to underscore the expected frequency by
a probability smaller or equal to the one you observed.
Once you do this you you get a probability 10 20 times the one you quote.
A relatively small number which occurs tenth or hundredth of times on BGA each day.
I will not because colonist.io has much better interface for catan.

also your claim that the dice are correct are unfalsifiable from a philosophic pov - nothing will count against it.
User avatar
Romain672
Posts: 1348
Joined: 05 April 2016, 13:53

Re: change the dice generator.

Post by Romain672 »

DivergentRalph wrote: 21 June 2025, 13:14also your claim that the dice are correct are unfalsifiable from a philosophic pov - nothing will count against it.
In practice, that's wrong. My calculations take a lot of things into account.
(which takes into the calculation: if the player already talked by the past in the forum, how weird the claim is (if you say '6 too much rolled' it's fine. If you said that the even days there is more 9 than 7 and odds days that the 10th roll is usually above 10, that will not be fine), your number of games since your last shared game, i divide by 11 if that can be done with any number, divide by 2 if the claim can be high and low, and the probability of the event)
That's a lot, and that's the point, many things can happen during one game, and while one game should not be enough, since I'm now used to the claim in the forum, I know what you can except from a game.
If you find after those calculations probabilities of 1/10k, then congratulations, you find an extremely weird event happening. Find me more of that, and something has a chance to not work properly.
If you find probabilities of 1/1M (even from a single game), then it's the weird game of the whole website in the last few years. Find a second, and something have an high chance to be wrong.
Find probabilities of 1/1000M, and there should be an investigation. Maybe that single game was bugged, maybe it was rare occurence, maybe someone find something to abuse. But that shouldn't be happening.

Alternatively, like I often said, you could do a claim, something which can be repeated and which can be checked by stats. Then in your futures games, you show us game where that event happen.
Thing is, I ask you something specific, and if you like saying lots of differents things happening, then you can't do that claim, since there would be way too many claims (I tried to count them once and I find around ~1000 claims possible, if you don't do the usual stuff).


In my last post, my calculations find that the game shared here got something which should happen every 4 and 26 games.
That's so far from my 1/10k, 1/1M, or 1/1000M, and that's the point: that work well, so finding a game which should happen once every 10000 profiles is extremely hard.




And people don't like my posts and answer to them for in my opinion two reasons: 1- they are harder to follow and 2- that find high probabilities. People like saying that this event was a 1 out of 200k, but when you look further, you can find that in their last 100 games, that event should be happening one time 1 of 12. While 1/12 is still very unlikely, if you got multiples people who could be saying that, or multiple series of 100 games, then that's not unlikely anymore, that's expected. But so they will stick with their 1/200k since in their mind, it's something which shouldn't happen.
Then other people read them, and that confirm what they were thinking, and are now convinced the system is rigged.

It's simple to share wrong information.
It's extremely hard to get back to the truth, since that require a lot more arguments and explanations.
People will stick with a post of 2 sentences, and will maybe not read and not take into account longer posts, or posts which require thinking, or posts which contradict what they were already convinced for.
User avatar
Romain672
Posts: 1348
Joined: 05 April 2016, 13:53

Re: change the dice generator.

Post by Romain672 »

I will add that if you want, i can code something, then take your last 100 games, and check how much times there was something weird about the developements cards.
The problem is that players could be drawing multiples of them, so I'm not sure I could do anything with the results.
Except by taking all pov (which would be too long), I could see which player draw which card.

And people are just taking developments cards, waiting either to have 3 horses (don't remember the name sorry) to play them once per turn. Or wait for the black pawn (don't remember again :( ) to goes into their territory to play their cards.

If all players do that, then everyone will keep their horses, and then someone will get the black pawn in their territory, and will play one of their horses. As a response, another player will play one of their horses, and move the pawn to another place, and so on.
You finish by having many horses played in a row. But that's the strategy.

And that look an easy way to be biaised about that.
User avatar
euklid314
Posts: 679
Joined: 06 April 2020, 22:56

Re: change the dice generator.

Post by euklid314 »

DivergentRalph wrote: 21 June 2025, 13:14 also your claim that the dice are correct are unfalsifiable from a philosophic pov - nothing will count against it.
No, it is in fact very easy.

Anybody may make a claim about what is wrong with the dice and post their claim in the forum. Then they check it within their next 20-50 future(!) games (the number is a rough estimate - it depends on how likely the claimed event occurs). If they will find something that is outside a 99% confidence interval the attention of the forum will absolutely be theirs!

Once there was a bug in Cant Stop and one player did exploit it. He could not influence the dice but he did know in advance when he would fall. Thus he produced absolutely crazy sequences for a long time. This was noticed and it is was so far outside of the possible that it was soon clear that something fishy was going on...
User avatar
AmongMajesticRuin
Posts: 20
Joined: 18 August 2023, 14:05

Re: change the dice generator.

Post by AmongMajesticRuin »

NOBODY slams the CATAN dice generators more than I do. The one for STEAM is even worse IMHO. HOWEVER, I played a game of CATAN IRL last weekend where the dice rolled 3's, 11's & 12's more than 6's & 8's. I was gobsmacked and thought I was in a weird online simulation.

I am now WAY LESS vocal about the dice.
Post Reply

Return to “Catan”