דיווח • Board Game Arena

2022 14 Jun 18:37 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

sorry, dont have the time (right now) to go through the games - maybe just check how the 2 random values are generated ?
since its pseudo random, it might, in combination, generate a pattern.....

PS: I try to post more games like that later

2022 16 Jun 09:19 • darhf • הצעה זו לא נותחה על ידי היזמים עדיין:

Report short description
"Dice weirdness (since last update?)"
changed to
"Are dice results really random ?"

2022 16 Jun 09:26 • darhf • הצעה זו לא נותחה על ידי היזמים עדיין:

I'll keep this one open to let people post their weird dice results.

But I can assure you that the dice randomness is indeed random.

2022 16 Jun 19:31 • tabicat • הצעה זו לא נותחה על ידי היזמים עדיין:

This happens in real Catan games as well. Certain dice rolls just happen to occur far less often than their statistical probability suggests. In other words, it's not usual for a roll like "4" or "9" to almost never come up, even though statistically is should happen about 10% of the time. I've seen games where a "6" has come up only a few times.

2022 17 Jun 02:44 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

@darhf ok, cool that you checked - But I will still try to incorporate that in my meta, and see how it ll work out :P ( the hypothesis is that you can, on the first clustering, see which numbers (a.i. 2-12) will occur more often next (and prob. in clusters, again))

boardgamearena.com/7/catan?table=276157622

Dice result ratio: 2 (~3%) 0Dice result ratio: 3 (~6%) 2Dice result ratio: 4 (~8%) 11Dice result ratio: 5 (~11%) 15Dice result ratio: 6 (~14%) 21Dice result ratio: 7 (~17%) 17Dice result ratio: 8 (~14%) 14Dice result ratio: 9 (~11%) 9Dice result ratio: 10 (~8%) 11Dice result ratio

15 x 5, 21 x 6, 11x 10

seems not to special in the end stats ... but the first rolls seemed , lets call it "weird" again - and that it worked out from there as assumed, doesnt proof anything since it's just one game ( I know )

2022 17 Jun 02:53 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

@tabicat I know, that is not why I made this post; It's not because it seemed a little off^^; it would be weird if its not that in most games :D (but also got games where numbers[2-12] got rolled x3-x4 as much (relatively) as one would think; for example the 3 8/36 instead of 2/36, which also should happen sometime)
I was talking about patterns, because it's pseudo random - which can be greatly made, but might generate patterns, even if both dice individually are close enough to real random(a.i. atom decay)

2022 19 Jun 11:42 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=276714567 10s and 6s ftw

2022 19 Jun 12:17 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=276730675 10s and 6s ftw ...

2022 19 Jun 12:50 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=276756731 broken and clustered 4 and 9 - and clustered 5 at one point

2022 20 Jun 21:58 • XDannyOrz • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=277112701 11 rolled 0

boardgamearena.com/table?table=277046660 11 rolled 0

2022 21 Jun 05:52 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=277211706

10 and 9 clusters in the beginning - (most notable in similar combinations) - also in mid game 11 clusterings happen to appear - 9 and 11 clusters reapper (with mostly same red and yellow dices)

Dice result ratio: 2 (~3%) 1; Dice result ratio: 3 (~6%) 5; Dice result ratio: 4 (~8%) 7; Dice result ratio: 5 (~11%) 6; Dice result ratio: 6 (~14%) 11; Dice result ratio: 7 (~17%) 20; Dice result ratio: 8 (~14%) 11; Dice result ratio: 9 (~11%) 16; Dice result ratio: 10 (~8%) 10; Dice result ratio: 11 (~6%) 9; Dice result ratio: 12 (~3%) 4

again as averages, they dont seem too weird - its that they(the clustering numbers- here 9 11 and 10 at start) are so often in quick succession to themselves, with often identical dice values for the red and yellow dice ...

2022 21 Jun 10:32 • XDannyOrz • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=277257715 the dice 9 only rolled 2 times
but it showed 4 in the statistic.

this game is totally bugged with the dice

2022 21 Jun 18:41 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

@XDannyOrz its 4% ....-.-

2022 24 Jun 16:58 • joel2v32 • הצעה זו לא נותחה על ידי היזמים עדיין:

I've played 2 Games including this one - boardgamearena.com/5/catan?table=278116377

I found in both games the 8's and 10's are very under rolled. It seems the 6 is over rolled in comparison.

Just a gut feeling.

2022 25 Jun 05:58 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=278264526

little clusters of 4s 5s and 6s ... later 9s and 12s (and 5s continueing)

2022 25 Jun 06:01 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

also, since it seems people dont read past title - and "weird dice", was too vague and clickbaity (to force people reading) could you change it to "Dice result clusters - Are dice results really random ?" at least ??

2022 3 Jul 18:37 • Jimbobbery • הצעה זו לא נותחה על ידי היזמים עדיין:

Each game contains an incredibly high recurrence of a single roll which is statistically highly unlikely – this happening in even 75% of games would be explainable but 90%+ of games containing an anomyly suggests there is something that can be improved in the random number generation.
For example in the last game I played there were more 5s rolled than 6s and 8s combined. Over a large sample size (100+ rolls) this can absolutely happen – but it happens game after game so there is likely to be a code issue here.
Are the dice rolls being calculated in real time at the point of rolling or is there an array of random numbers being created when the game is loading and worked through during the game – the game has the feel of the later, which in a game for which the core mechanic is a bell curve distribution of resources, makes playing the game farcical and the outcome determined after 2 or 3 rounds.
At best it feels like something very static is being used to seed the random number variation. Consider generating these in real team and seeding with something truly random, like the last player’s turn in milliseconds.

2022 6 Jul 04:13 • Texans80 • הצעה זו לא נותחה על ידי היזמים עדיין:

twenty one 6s and two 8s rolled in my game

2022 7 Jul 23:02 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=281535628

6s and 9s - again not suprising it is rolled often - but it was obv. from the start when there were the first , lets call it clusters again ...

2022 8 Jul 02:50 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

also this 2 onslaught (also 4s and 11s) - seems noteworthy (but maybe not cluster specific - just things that happen)

boardgamearena.com/7/catan?table=281576357

2022 8 Jul 03:07 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

@Jimbobbery

what would be the difference between creating a table beforehand and generating it live, based on something like milliseconds ?
it is still determined either way, what the result will be, with the difference, that one would have the feeling to have control of it, if one knew how its generated [or believe to know should be enough :P] - if it's a good algorithm which creates the pseudo random results (low predictability in trends and next results), it should be just fine, shouldn't it ? (if there is no way for players to get that table, that is :P )

2022 13 Jul 18:35 • Dendarii • הצעה זו לא נותחה על ידי היזמים עדיין:

I haven't run any official statistics analysis, but the dice really do seem off on here. Sure, it's possible for the same number to roll 3 or 4 or 5 times in a row, but for that to happen multiple times per game, nearly EVERY game, often with the same number, and having 1 number basically EVERY GAME that rolls more than any other (usually a 9, 10, 11, 3, 4, or 5) to the point that it basically determines the outcome every time? Very strange.

And yes, 7 is the most frequently rolled number, but to the point where more often then not it's the first roll in almost every game I've played, and if not the first, then ALWAYS in the first round, and then again, rolling a seven sometimes 3 or 4 or 5 times in a row being almost EXPECTED at this point from how frequently it happens that again the robber alone can basically determine the outcome of the every game? Really seems off.

In terms of a pseudo-random algorithm, it really seems like there might be something wrong in the coding.

2022 16 Jul 19:23 • Jimbobbery • הצעה זו לא נותחה על ידי היזמים עדיין:

@silentProtest - Thanks for the reply.
Few things -
For Settlers, being able to predict the next dice roll isn't the most important thing as most of the decision making made on placement is made at the start of the game and is done so based on a bellcurve of 2D6 rolling which will peak at 7.
For every game I've played the abnormally highly occurring number has been on the 2nd from bottom row, on the hex 2nd from the left - importantly, players seem to be catching on to this - the highest ranked players are going straight for this hex and the hex with the corresponding number. Might be a conincidence but doesn't feel like it.
Populating the array at the start of the game will help in the performance of the game but removes the opportunity of adding genuinely random input into the psudo randow generator - the generator probably is producing a bell curve of numbers just now - but they're not piviting on the number 7 which a genuinely random input will give you over time.
The seed range from current time in Miliseconds (MS) is very small and relies on the system being called to return the time updating its response accurately every MS. For example I've seen reports of time related functions on Windows only changing their return value every few milliseconds so the same seed value will come back repeatidly.
Hope this is helpful - again, as players seem to be catching on to this and flooding where the high occurance is popping up.
Cheers

2022 19 Jul 18:34 • LordMouton • הצעה זו לא נותחה על ידי היזמים עדיין:

The more I play, the more I think there is an issue with the dice. Everytime, numbers under 7 are over balanced.

#284621586

4 : 20x
5 : 15x
6 : 19x
7 : 7x
8 : 14x
9 : 7x
10 : 7x

4 was actually the number on the hex described by Jimbobbery above : 2nd row from bottom, 2nd hex from the left

2022 19 Jul 23:44 • Just Keep Climbing • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=284707125
7s rolled 3 times in a row and not the first time I've seen it.

2022 24 Jul 20:28 • Romain672 • הצעה זו לא נותחה על ידי היזמים עדיין:

Please try to give something specific. So something like:
"if two succesful players roll two 3s on a row, then the probability to roll 3s in the whole game would be higher than normal"
That way, we can test that hypothesis, and see if it's true or not.

Atm, that post look just a collection of randoms games.

2022 25 Jul 01:26 • silentProtest • הצעה זו לא נותחה על ידי היזמים עדיין:

since the description on top isnt clear enough I copy paste my example from the forum, to illustrate better what I mean:
.... what i mean by cluster is some thing like that (lets use [1-6] with 1/6 each): 123561235644441235612356 ; this would be a 4 cluster - and if this happens once in a game this would happen prob. again later with same number ( not rarely with same combination of red and yellow dice number) - THAT is what I find curious, and way to often that it would be just random (obv. Id need a larger data set too proof that - super unlikely sequences can just happen - but it would be nice to have a number to put on, how unlikely that consistency is... )

2022 10 Aug 23:04 • Just Keep Climbing • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=290567566

9s were spun in clusters throughout the game and I'm fairly certain that they were spun more frequently than sixes and eights COMBINED. I'm not a premium member so I cannot see dice statistics, so if someone could double-check this game I would appreciate it.

2022 18 Sep 23:46 • Romain672 • הצעה זו לא נותחה על ידי היזמים עדיין:

@plshavemercy:
Here is the 2s/3s/4s/5s/6s-7s-8s/9s/10s/11s/12s rolls done
0/3/2/9/8-13-10/18/8/3/2

18/76 rolls were a 9.

It has a probability of occuring of 0.14%, 1 out of 701.

Since you can do the same claim for every number, I will 'multiply properly' that probability per 11.
That now give 1.56%. Something unlikely like that or worst happen one game out of 64,2.

So you are slightly better than the game shared on the forum, but still far from finding a really strange game: I'm still waiting for a probability under 1/10k :D

And please note it was not the claim done on that bug report. Rolling the same dice accross one game =/= having many time a specific dice at one moment.

2022 20 Sep 05:01 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I just played a game where 8 was rolled 20%, 10 was rolled 15%, 3 was rolled 3% and 5 was rolled 6%. This is highly unlikely for truly random dice, especially considering one player had both 8 and 10, while another player had 3 and 5. The dice seemed to be clearly giving an advantage to one player over another.

2022 23 Sep 07:26 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another game with unexpected dice outcomes. 10 rolled 3 times in 4 rolls (probability of that happening is 1 in 500). End game stats:
Percentage of 2's rolled: 7%
P(3)= 4%
P(4)=11%
P(5)=10%
P(6)=12%
P(7)=18%
P(8)=10%
P(9)=9%
P(10)=13%
P(11)=4%
P(12)=2%

2022 24 Sep 00:55 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another unlikely game:
P(2)=0%
P(3)=0%
P(4)=13%
P(5)=11%
P(6)=11%
P(7)=13%
P(8)=26%
P(9)=18%
P(10)=5%
P(11)=0%
P(12)=3%

2022 24 Sep 04:59 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

21% 8's in my last game. More importantly, I knew 8's would be over-rolled after the first two rounds. The bias is evident from the beginning and doesn't change throughout the game.

2022 24 Sep 06:49 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just finished a game where 10 was rolled twice as many times as its expected value. More interestingly, I observed in the comment board that 10 would be over-rolled after the first few rolls of the dice.

2022 24 Sep 07:40 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another game where dice were wildly in my favor, right from the beginning. I easily won with 10 points, while both other players had 6 points, even though they were better players with better initial configurations.
P(2)=5
P(3)=11 (this is why I won)
P(4)= 11
P(5)= 13
P(6)= 6
P(7)=22
P(8)=13
P(9)= 3
P(10)=8
P(11)=3
P(12)=5
Predicted the dice would be in my favor after the first 6 rolls of the dice. Other players thought I was a bot. Really need to look at that RNG

2022 24 Sep 08:39 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I got a 1 in 50,000 event - this game opened with three 12's in a row.
P(2)=3
P(3)=7
P(4)=8
P(5)=8
P(6)=7 (less 6's rolled than 12's!)
P(7)=20
P(8)=20
P(9)=8
P(10)=5
P(11)=7
P(12)=8

2022 26 Sep 03:00 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another game where 52% of the outcomes were 5 and below. The roll of 5 occured 31% of the time. The probability of this happening is less than 1 in 10K.

2022 26 Sep 06:34 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Every game I play has extreme outcomes.
P(2)=7
P(3)=7
P(4)=5
P(5)=21 (21 or more 5's in a game is 1 in 1000)
P(6)=9
P(7)=14
P(8)=7
P(9)=12
P(10)=9
P(11)=2
P(12)=7

2022 26 Sep 07:35 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Now I got an event that's 1 in 100K (no nines rolled the entire game)
P(2)=8
P(3)=10
P(4)=19 (1 in 2K chance of 19% or more 4's)
P(5)=11
P(6)=13
P(7)=14
P(8)=16
P(9)=0
P(10)=5
P(11)=3
P(12)=2

2022 26 Sep 19:12 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Next game also extreme with 66% of all rolls being 6 and under. 8% twos.

2022 27 Sep 02:50 • Mathew5000 • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers18 Please post the table number of these games so that others can verify what you are saying.

2022 27 Sep 19:31 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

A game with 6% 8's (9 in 1000 chance) and 18% nines (2% chance)

2022 27 Sep 19:33 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Mathew5000 Thanks for the tip. I'll do so from now on.

2022 27 Sep 19:59 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

This game ended early, but shows the sort of clustering that can happen. 8 nines out of 30 rolls (1 in 100 chance)
boardgamearena.com/table?table=303601034

2022 27 Sep 20:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=303612490

25 eights in 90 rolls (1 in 10K chance)

The only numbers I didn't have in the initial setup were 2, 8, 11, 12. This is significant and affects the dice bias.

2022 27 Sep 22:30 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=302620573

Less than 1% chance of so few 6's

2022 27 Sep 22:35 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/gamereview?table=302633200

1 in 50K chance of starting the game with three 12's in a row

2022 27 Sep 22:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=303184521
1 in 100K chance of no nines rolled in the game

2022 28 Sep 02:23 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=303694100

This is the nearest to "normal" (pun intended) game I've had in a while, and still has several 1 in 100 events

Just six 6's out of 100 rolls

2022 28 Sep 07:07 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=303726873

Longest game ever, and quite "normal" in the dice distribution. BUT, 9 was over-rolled (there were 15% nines, which is not too unlikely (13% chance). 3 was under-rolled. I predicted this would be the case in the chat box within the first five minutes of the game.

2022 28 Sep 07:26 • Mathew5000 • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers18 Some of your probability calculations are wrong.

First consider boardgamearena.com/table?table=302620573

There were 63 dice rolls in the game, of which a 6 came up 4 times (moves 40, 61, 63, and 170).

Let X be the number of times you get a 6, when you roll a pair of dice 63 times. Then the probability that X is 0, 1, 2, 3, or 4 is:

0.000081 + 0.000824 + 0.004118 + 0.013504 + 0.032671 = 0.0512, or about 1 in 19.5.

You said, "Less than 1% chance of so few 6's" but in fact the probability of 4 or fewer sixes is greater than 5%.

Now consider boardgamearena.com/table?table=303184521

This game also had 63 dice rolls, of which none were 9's.

Let Y be the number of times you get a 9, when you roll a pair of dice 63 times. The probability that Y is 0 is:

(1 - 4/36)^63 = 0.00059895, or 1 in 1669.586.

You said, "1 in 100K chance of no nines rolled in the game" but in fact the probability is about 1 in 1,700, not 1 in 100,000.

2022 28 Sep 08:00 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Mathew5000 Thank you for the correction. I couldn't find the total number of dice rolls. The sum of the table at the end is not always 100 (probably due to rounding), so I sometimes assumed that was total number of rolls, while it's probably probabilities. Many calculations I was then assuming about 100 rolls of the dice. I will correct it going forward.
I still think 1 in 1,700 is still too low of a chance to be random and that the dice are biased.

2022 29 Sep 09:04 • Mathew5000 • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers18, I do think that the end-of-game dice statistics are a little confusing, as to whether they are expressed as a count or as a percentage. I made a suggestion for improvement, if you agree please vote for it: boardgamearena.com/bug?id=72332

As to your main point (and that of some of the other posters in this thread), I believe it's an instance of the "rare events fallacy". Have a look at this article by mathematician David Hand: slate.com/technology/2014/02/the-improbability-principle-rare-events-and-coincidences-happen-all-the-time.html

It's understandable to ask, "In that game I just played, the probability of seeing zero 9's would be only 1 in 1,700 if these are fair dice. I haven't played anywhere close to 1,700 games yet so isn't it suspicious that I'm already encountering this 1-in-1700 event?"

The reason why it's *not* suspicious is, there are so many different things that could happen, any one of which looks unusual, and the chance of one particular thing happening is minuscule, but the chance that *any* of them happens is much much larger.

As a thought-experiment, what if you sat down before playing a game of Catan, and wrote out a list of specifically what results in the dice statistics you would consider unusual (assume each game has 60 to 70 throws). So the list would include, having two or fewer 5's show up, two or fewer 6's, etc. You might say it would be unusual if we get more than eight 2's rolled, or more than ten 3's. You might say it would be unusual if there's any sequence of six throws in which four of them are 3's. And so on. You could write a list, specifying exactly what you would consider unusual, on the understanding that any event which happens that is not on your list, you don't have the right to call it unusual after the fact. (It's important that you make this list before playing the game, otherwise the results are tainted by selection bias.)

Then you've got a list of events each of which is considered unusual, and you ask, what is the probability that *at least one* of the events on your list occurs in the game you're about to play, if the dice are fair. That is something that can be calculated exactly (because the list is expressed unambiguously).

What you'll find is, if you have exhaustively listed (before the game) *every* event that you would consider unusual, and then carefully calculate the probability that at least one of these events on the list occurs, you'll reach a probability that's pretty high, maybe over 0.5, of at least one of these events occurring.

Then if that's the case, it's actually not suspcious at all, that in most games of Catan you play on this site, something "unusual" occurs. In fact you would expect it to!

2022 1 Oct 00:13 • Chatfie • הצעה זו לא נותחה על ידי היזמים עדיין:

It is most definitely NOT RANDOM.

2022 1 Oct 05:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Mathew5000 I was just chatting with my husband about the rare event fallacy. He's a PhD software engineer and I'm a Master's Math professor. I was at the same time pondering if you could find any finite strings of digits in the decimal expansion of pi.

But, wouldn't you say that if I could predict the rare event within the first five minutes of the game, that there is a problem with the implementation of the RNG? I expect rare events. I should not be able to predict them.

Here's another game with 3 11's in 4 rolls (about 8 in 10K). This kind of occurrence is expected, but it's weird if it happens at the point in the game where 1 player has settled on every number except 11. This game also had 20% 8's and 20% 9's, leading to a quick victory for me, since I was depending on those rolls.
boardgamearena.com/6/catan?table=304559931

2022 1 Oct 06:38 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=304554996
10 rolled more than either 6 or 8 (predicted this within the first few rolls of the game)

2022 2 Oct 05:43 • A Thirsty Rogue • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=304824944

Predicted 8's rolling more often after 10 rolls.

Two 6's for entire game

2022 4 Oct 04:39 • Dave C Toronto • הצעה זו לא נותחה על ידי היזמים עדיין:

Hello your majesties, the developers

Hope this message finds you well.

I feel like a couple of strange random distributions in rolling today

boardgamearena.com/table?table=305268260
99 rolls - 20 x 4's (20.2%) - expected is 6.94%

boardgamearena.com/table?table=305268260
92 rolls - 29 x 8's (31.5%) - expected is 13.9%

Maybe I need a larger sample of dice rolls for this to make sense, but it kinda put me off the game today. It made me perceive that the random number generator was not within expected parameters.

Really just need somewhere to vent - it was a frustrating experience. Did I get a couple of anomalies in a row? Please reassure me the dice rolling subroutine is working as expected.

Thanks for running a great site I am happy to contribute to and keep me happy for hours :)

Dave from Toronto

2022 4 Oct 18:11 • ganzbaff12 • הצעה זו לא נותחה על ידי היזמים עדיין:

Totally agree

on table boardgamearena.com/table?table=305534566
there was only 4 times 6 at all!

2022 5 Oct 02:50 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Like Dave said, we love you, game developers. Great site!

Here's a game with 20% 8's (not all that low probability, but really weighed the game against the player with the highest elo). I'm curious if anyone can comment on how the "average number of tiles around settlements" is computed I the end-game stats. In this game, you'll notice how the player with the highest average number of tiles produced the least number of resources by a significant margin.

boardgamearena.com/8/catan?table=305666720

By the way, I've noticed this non-randomness in other games on this site as well, particularly in Viticulture. It's subtle, but noticible.

2022 5 Oct 03:24 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=305664566

1% chance of so many 2's and 5's

2022 6 Oct 00:43 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 11's and 12's than 8's and 4's

boardgamearena.com/8/catan?table=305916568

2022 6 Oct 03:18 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

3% chance of so many 4's

boardgamearena.com/5/catan?table=305948392

2022 6 Oct 07:48 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

2% chance of so many 4's

boardgamearena.com/8/catan?table=305979928

2022 6 Oct 08:48 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

No 4's in this game

boardgamearena.com/6/catan?table=305985648

2022 11 Oct 05:24 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

About 7 in 1000 chance of so few 9's. So much clustering of repeated 3's and 11's.

boardgamearena.com/6/catan?table=307368873

2022 11 Oct 11:46 • Mathew5000 • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers: I have a few questions and comments.

[1] You said: "But, wouldn't you say that if I could predict the rare event within the first five minutes of the game, that there is a problem with the implementation of the RNG? I expect rare events. I should not be able to predict them." My question is, how do you predict within the first five minutes of the game exactly which rare event is going to have occured when the final stats are in. For example: here's the most recent Catan game I played, lasting 41 minutes: boardgamearena.com/6/catan?table=307408908 In 54 throws, there were quite a lot of 9's and 11's, with relatively few 6's, 7's, and 8's. But how would you have predicted that within the first 5 minutes? The first 7 dice rolls of this game (approximately 5 minutes) were: 11, 10, 5, 6, 4, 6, 7. As the first five minutes of the game had a couple of 6's but no 9's at all, how could you predict from that, that by the end of the game there will be *fewer* 6's rolled than 9's?

Or if this particular game is not a good example, could you choose another Catan game I played within the last week or two (boardgamearena.com/gamestats?player=88974824&game_id=1539) and explain how I could have predicted, within the first five minutes of the game, a rare event that shows up in the final dice statistics of that game.

[2] What do you think is a possible mechanism for the issue you perceive, with the dice not being random? Do you think that the code is written so that for each game of Catan, the code looks at which dice rolls will benefit a certain player in the game, and then produces those rolls at a higher frequency? How would the code decide which player to benefit in a particular game?

[3] I'm not asking these questions rhetorically. In one of the early responses in this thread, Darhf (one of the developers of the BGA implementation of Catan) said, "I can assure you that the dice randomness is indeed random." I find that entirely plausible. But I am keeping an open mind and I want to try to fully understand your position, which is the reason I asked [1] and [2] above.

[4] One of the reasons I find it plausible that the RNG is properly written, is that it's pretty easy to write code that simulates the sum of two dice. Other random processes, such as shuffling a deck of 52 cards, are much trickier to simulate, because the number of permutations of the deck is more than 8 * 10^67, and you need to make sure that each one of them is equally likely to occur. But a die has only 6 faces, so the code is straightforward.

2022 12 Oct 05:06 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Table #307651606

2% chance of so few 4's. The player with the top elo had almost no cards at the beginning of game in spite of having high probability spots and not being particularly picked on by the robber.

2022 12 Oct 05:27 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Mathew5000, are you a malayalee by chance? I'm a Canadian but I'm married to a malayalee.

Here's my analysis of table #307408908.

Based on initial setup, Nafnaf had 5,6,8,10,11
Jonas had 3,4,5,6,8
Mathew had 5,6,9,10

First rolls were 11 (only Nafnaf), 10 (poor Jonas), 5 (all get something), 6 (all get something), 4 (only Jonas), 6 (all get), 7 (Nafnaf gets to move robber), 12 (nobody)

6's at the beginning do not hurt or help anyone (everyone gets cards, so doesn't give much info about dice bias)

Based on this first set of rolls, Nafnaf has a very slight advantage. Even his unlikely numbers got rolled. At the end of game, 11 was rolled 13 times (an insane advantage for Nafnaf) while 3 was never rolled (hurting Jonas). I maybe should have written down a few more rolls to see that the dice are slightly hurting Jonas (game ending with few 6 and 8) and helping you and Nafnaf (lots of 11 and 9). I would be interested to know the average elo of each player at game start.

2022 12 Oct 05:47 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Mathew5000 If we're getting technical, then actually a function like Randint(1,6) is at best pseudorandom.

But I do believe that there might be a bias coded in to encourage players with low elo to keep playing and to make things slightly more challenging for better players. It's easy to code a function of elo that prefers rolling 7's for players with low elo or slightly avoids the initial placement numbers of players with high elo. I can only guess based on observations. It's not uniform since occasionally I've received the advantage in spite of having the highest elo among my competitors. The dice here just remind me too much of the dice of a Catan version I used to play against AI that would slow down rolling the numbers of the current winning player. In those games I would settle on every number and then the game would only roll 7's. On BGA I also tend to have better outcomes settling on all the numbers rather than the high probability spots. It's not as in-your-face as the online Catan I mentioned earlier, but the bias seems evident to me.

Could you post a link to that second game you were talking about? The stats link didn't work for me

2022 12 Oct 07:22 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=307678615

1% chance of so many 7's
4 and 10 rolled 150% more often than 5 and 9 (predicted after a few rounds that 3,5 and 9 would be under-rolled)

2022 12 Oct 09:01 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=307693001

Out of 49 rolls, only two 6's. Forty percent of the rolls were 8 or 9 (1% chance of that happening)

2022 12 Oct 19:05 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=307820064

2% chance of so few 8's

2022 13 Oct 03:23 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=307945036

6 in 1000 chance of so few 10 and 11.

2022 13 Oct 04:02 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

One of those rare times where dice were wildly in my favor, even though I had the highest elo at the table

boardgamearena.com/1/catan?table=307950692

No 9's in 60 rolls (about 1 in 1000 chance)

2022 14 Oct 05:27 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=308253956

1 in 1000 chance of so many 7's

2022 15 Oct 18:54 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=308710813

1 in 500 chance of so few 8's and 9's. More 12's than 9's in this game

2022 15 Oct 19:11 • RaxoGG99 • הצעה זו לא נותחה על ידי היזמים עדיין:

Nah developers want to make this game bad

2022 15 Oct 19:46 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=308729494

No 3's in 79 rolls (about 7 in 1000 chance)

2022 15 Oct 22:22 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/6/catan?table=308729494

So many 9's (incidentally 9 was the only number that the highest elo player hadn't settled on)

2022 17 Oct 00:30 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=309114990

To me, this game really shows evidence of the dice bias as a function of elo. I (with the highest elo) was the only player settled on 10, but it did not roll until the very end of the game. If you watch the game, roll by roll, you will see the evidence. I'll post the probability of so few tens below. I expect games with no 10's. But I don't expect them to coincide with incidences where the top elo player is the only one settled on 10.

2022 17 Oct 00:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=309114990

1.8% chance of no 10's in the first 46 rolls (by coincidence the player with the highest elo was the only player settled on 10)

2022 17 Oct 05:04 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Here's a very-near-normal distribution. Except the low occurrence of 8's. Incidentally, for most of the game, I was the only player settled on 8 and I had the highest elo.

boardgamearena.com/5/catan?table=309150219

2022 20 Oct 17:46 • Alison MacPrickle • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm finding weird dice results more often than not, I'm certain there is a problem. I also play a lot of Backgammon and it has nothing like the 'sticking' problem that is happening in Catan. I've just played a game which had roughly equal sixes and eights (about 16 and 18), but 30 (thirty!) number fives and not a single 4. If that was just one game, then fair enough, but almost every game I play is spoiled by weird dice. I've noticed many more cases of players quitting because they get frustrated.

2022 20 Oct 17:55 • Alison MacPrickle • הצעה זו לא נותחה על ידי היזמים עדיין:

This is the game table where we had 30 number 5s and not a single number 4 boardgamearena.com/table?table=310173666

2022 22 Oct 16:58 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=310173666

The probability of so many 5's in Alison's game (13 5's in 44 rolls) is 7 in 10K. It is not a coincidence that only one player settled on 5 in the initial placement.

2022 22 Oct 17:01 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Above, I should have said that it's no coincidence that only one player settled on 4 (there were no 4's rolled) and that same player was the only player not settled on 5 (5 rolled 30% of the time)

2022 23 Oct 07:01 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=310933860

7 in 10,000 chance of so few 4 and 8-12

2022 25 Oct 04:23 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 11's than 7's in this game, as well as 20% 9's (giving an advantage to the player with the medium elo)

boardgamearena.com/8/catan?table=311510699

2022 25 Oct 06:33 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=311546602

12 8's rolled but only 3 6's.

I had the highest elo at the table and had settled on all numbers (3-11) except 4, 8, and 11. Is it a coincidence that these numbers were over-rolled?

2022 25 Oct 07:31 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just as many 2's as 4's and 10's

More 5's and 9's than 6's and 8's

boardgamearena.com/8/catan?table=311543385

2022 25 Oct 08:07 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Before the game started, I predicted 5 would be over-rolled (I didn't have a 5) and 3 would be under-rolled (I started with two settlements on 3's). I had the highest elo at the table. Sure enough, my predictions were correct (more 5's than 6's or 8's and half as many 3's as the expected value).

boardgamearena.com/1/catan?table=311557303

2022 26 Oct 03:09 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

7 in 1000 chance of no 11 or 12 (but I wasn't able to predict this one. I thought 3 would be rolled less than it should (it was) and 4,8,11 would be rolled more (they weren't)). Dice favored orange strongly.

boardgamearena.com/4/catan?table=311807666

2022 26 Oct 03:51 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

1 in 10K chance of 15 9's in 62 rolls.

boardgamearena.com/1/catan?table=311811493

2022 29 Oct 05:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Finally a game with truly definitive evidence that the dice are not coded well.

1 in 10K chance of 5 elevens and twelves in the first 8 rolls.

The player with the lowest elo only settled on two 8, 5, 10 locations at the beginning of the game. In my experience, this encouraged a truly exaggerated error in the dice. The dice gave a slight advantage to the lowest elo player, but this was evidenced by an extreme amount of 5's, 8's and 10's, especially at the beginning of the game. It was a long, long time before we saw a six or a nine rolled. I'll look at the game log and try to see if I can calculate the probability of so few 6's and 9's, as it was quite extreme. The initial placements of the players and their relative elo encouraged the dice to be more extreme.

boardgamearena.com/5/catan?table=312704062

2022 29 Oct 05:53 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=312704062

In this same game, there was a second occurrence of three 12's in seven rolls (7 in 10K chance)

There were no nines in the first 50 rolls (1 in 500 chance)

2022 29 Oct 06:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=312704062

Same game as the above two comments.

38 of the first 40 rolls were 5, 8, 10 (where the lowest elo player had settled) or 7,11,12 (where no one got cards). The probability of rolling 5,8,10,11,12,7 is 21/36.

The chance of rolling 38 of these numbers in the first 40 rolls is 1 in 5,000,000 (1 in 5 million chance).

This game is my best evidence so far that the dice are rigged.

2022 29 Oct 06:55 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=312712883

No 4's in this game (6 in 1000 chance). Is it a coincidence that the player with the lowest elo did not settle on 4?

2022 2 Nov 02:16 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=313865704

More 2's than 8's

2022 3 Nov 02:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=314168344

The lowest elo player settled on every number except 6. There was only 3 sixes rolled in 62 rolls (probability of that happening is 1%). Also 1% chance of so many 10's.

2022 9 Nov 03:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

An anomaly: the dice were clearly to my advantage, even though I had the highest elo. The only place I didn't settle was 8 and it was rolled less than 11, where I was the only player settled. More 11's rolled than 8's and tied with 6's. Still only 3-4% chance of this occurrence, so not that rare, but definitely gave me an advantage

boardgamearena.com/4/catan?table=315991890

2022 12 Nov 05:36 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Game #316891747

1 in a million chance of no 2,3,8, or 10 rolled in the first 31 rolls (the numbers which the highest elo player settled on)

2022 13 Nov 06:35 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=317190597

6 in 10,000 chance of so few 2,5,9,10,11

There were more 12's rolled than 5's

2022 14 Nov 19:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=317654708

Only 1 eight in fifty rolls (4 in 1000 chance)

5 in 100K chance of so few 7's, 8's, 9's and 12's

2022 16 Nov 04:05 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/gamereview?table=318071492

Just as many 11's as 9's and 10's combined.

The dice clearly favored one player from the beginning. 24 of the first 30 rolls were 2,5,6,7, or 11 (1 in 1000 chance). The favored player had settled on 5,6,11

2022 16 Nov 08:10 • ufm • הצעה זו לא נותחה על ידי היזמים עדיין:

en.wikipedia.org/wiki/Birthday_problem

2022 16 Nov 19:15 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=318244849

More 11's than 8's. The high prevalence of 9, 11 and 4 benefitted the player with the lowest elo.

2022 16 Nov 19:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just a note that, even if the distribution was perfectly "normal" at the end of each game, it is still possible to code a bias into the dice, for example, by
1. Ensuring one player's numbers come up mostly at the beginning of the game, while the other player's numbers are rolled mostly at the end (too little too late when the first player has already established cities)
2. Giving most of the sevens to one player
...

2022 24 Nov 01:20 • Just Keep Climbing • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=320436060

8s were not rolled at all until after move #100, and they weren't rolled all too many times throughout the game in general. Both 8s were brick, (and the other brick was 12,) so that made it very frustrating for us, especially for my two opponents when I blocked their 8 while my 8 was open :). So this is an unbiased report (as I won the game) that I believe something is wrong with the dice rolls on BGA.

2022 28 Nov 20:34 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The lowest elo player was the only player settled on 3. The only number they didn't settle on by mid-game was 5.

There were 16% 3's rolled (more than 5's and 6's combined!) There was only one 5 rolled the whole game. Coincidence? I think not.

boardgamearena.com/7/catan?table=321798773

2022 28 Nov 20:39 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Also, I think I'm seeing a dice bias in which player rolls 7 the most often. In this game, the lowest elo player rolled 6 7's, the next player had 3 7's and I just rolled one 7.

boardgamearena.com/7/catan?table=321798773

2022 28 Nov 21:55 • Romain672 • הצעה זו לא נותחה על ידי היזמים עדיין:

One quick thing that you can do is to take your probability that i will call p and do the following:
1-((1-p)^11)
So like if you said that your rolled X sixies, and this or more sixies has a probability of 1%, you apply my formula which give 10.5% (>10%), and instead of saying
"on that game, the probability of rolling X sixies or more was 1%"
you can said
"on that game, the probability of rolling X sixies or more, or rolling any other number with an even more unlikely event was 10.5%"

Most of your events will be way more likely. But that's normal since you choose to check a specific number after seeing which number was the most unlikely.

And if you want, you can count your 100 next games, and write all your most unlikely event. That should give a second good sample of what is going on. (but start at any moment like just after you read it, not at a moment you choose)

2022 29 Nov 05:16 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Thanks for the tip, Romaine672. As a mathematics professor, I keep your calculation in the back of my head, but like to post the specific event probability (partly because it becomes quite complicated to do your calculation when analyzing the probability of so many 3's, 6's, 8's and 10's, for instance). I have a hypothesis that there is a bias across all the games on BGA, which encourages players with low ELO to keep playing and makes it more of a challenge for players with high ELO. It is not a uniform advantage for low ELO players, since occasionally the highest elo player gets the advantage (like this game boardgamearena.com/7/catan?table=321903512, where there were as many 3's rolled as 7's). I have encountered this bias in Viticulture, where I will sometimes draw only white vine cards, or get all the worst visitors, or get two starting papa's that both have the tasting room. This has happened often enough to catch my attention and make me go hmmm... With Viticulture it would be hard to prove the bias. With Catan it's easier because of the mathematical nature of the dice rolls and the large number of rolls in the average game.

If my suspicion is correct, then the highest elo player at the table should settle on as many different numbers as possible, while maintaining spots with high probability of rolling, while the lowest elo player should settle on as few different numbers as possible (but high probability spots). In my experience, incorporating these two strategies encourages the dice bias and the likelihood of rare events.

2022 30 Nov 01:06 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Check out this crazy distribution! Only 7% fives and sixes combined. 52% fours, nines and tens combined (chance of that is 1 in 100,000). Really clobbered the player with the highest elo.

boardgamearena.com/4/catan?table=322151166

2022 1 Dec 17:36 • Roaks • הצעה זו לא נותחה על ידי היזמים עדיין:

Table #322629096 had this same type of problem. We had 3 3s in a row followed by 7 7s in 8s rolls. So in 11 rolls, we got a string of 3 3s and 7/8 7s. That is very statistically unlikely, even in a simulation. The dice need to be fixed in CATAN so that this game is not favoring one set of players over another.

2022 2 Dec 20:42 • Jevanyn • הצעה זו לא נותחה על ידי היזמים עדיין:

I've seen bad dice randomness in the box set as well, to the point where I threw away the original dice from the set and put four new pairs of dice in the box, one for each player.

2022 5 Dec 00:42 • TedinTO • הצעה זו לא נותחה על ידי היזמים עדיין:

I definitely agree that the dice do not follow expected probabilities. In my experience this has worsened to the point that the winner of most games of experienced players is (in effect) the dice, not the "winning" player.

Although I try to counter this reality by placing my initial settlements close to as many different numbers as possible, almost invariably the dice decide to fixate on the numbers to which I am not adjacent (i.e. 5s and 9s - or worse - when I am near both 6s and 8s.)

Although I really like the game concept, this flaw has led me to quit one or two games in extreme frustration and consider giving up playing the game entirely.

P.S. I agree with one poster about the box set. Correcting with different sets of dice is an innovative solution that I must try to remember.

2022 7 Dec 11:07 • Fantomat • הצעה זו לא נותחה על ידי היזמים עדיין:

None of what you are doing in this thread follows proper scientific methodology. en.wikipedia.org/wiki/Scientific_method

Please use science (math, cs) to prove your assumptions not your feelings.
What you are doing is called cherry picking : en.wikipedia.org/wiki/Cherry_picking
Check this also : en.wikipedia.org/wiki/Fundamental_attribution_error (low elo having better luck)
en.wikipedia.org/wiki/Confirmation_bias

Dice do not have a will, randomness does not have biases, only humans have. We, as humans, will always have difficulties to correctly understand randomness, because our brain is made this way. We have all sort of biases and feelings that are corrupting our judgement relative to random events. It's ok that you'll are concerned about this topic and if dice were not random it would be indeed a big problem. But nothing here objectively tells me that it is the case.

The day you are using science to prove your point, tell me I'll be happy to check the work and to admit you are right and that dies are rigged.

One last piece of advice : in probabilities your personal experience is not relevant. One game is not relevant. A few games is not relevant. You need data, a lot of data. ~ 1,000,000 random dice throws could be a starting point. Taken from all of the catan games ever played at random. Then you can start working on something.

2022 7 Dec 20:38 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Fantomat, I cannot predict which numbers will be over-rolled or under-rolled (more or less than their expected values) in a game with table dice. I can predict it in BGA Catan (after the first six rolls). That's the difference.

I am hypothesizing that the developers have coded in an advantage for players with low ELO on all the games on BGA.

2022 7 Dec 20:49 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another hypothesis:
1. Have the lowest elo at the table
2. Settle on two initial spots with the same three numbers (two 8,3,4 spots for example)

I predict that the distribution of dice rolls will be far from "normal" (especially if you plot the dice rolls for the first half of the game - I have noticed that occasionally the dice try to correct the skew somewhat at the end of the game). I have not been able to test this hypothesis often because I find it hard to be the lowest player at the table, but I encourage others to try.

2022 7 Dec 20:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Fantomat I can also scientifically graph thousands of sevens rolled and compare how many go to the highest elo player and how many go to other players. How skewed do you think this metric should be to show evidence of bias coded into the dice?

2022 7 Dec 22:05 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Fantomat, design an experiment to test the hypothesis that the dice give an advantage to one player at the table, as a function of the relative elo of the players and their initial placements. Each individual game absolutely does matter. I'm not saying that the dice are not "normal" over the set of all games.

(I don't mean this to brag, but just so you know, I have a graduate degree in science, my thesis was the awarded the best in the faculty of science, and I've published in journals with people at NASA and Harvard)

2022 7 Dec 22:11 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@darhf Thank you for assuring us that "the dice randomness is indeed random". How can you prove that to be the case? Even the best random number generators are not truly random, and that non-randomness can always be taken advantage of by smart players, even if it is very slight.

2022 8 Dec 00:10 • darhf • הצעה זו לא נותחה על ידי היזמים עדיין:

Nothing is truly pure random, even physical die; But the BGA random function is random enough to play Catan.

It just asks for two numbers between 1 and 6. That's all.

There is no link anywhere in the code betweeen a random call and whatever it will be used for. The Random function doesn't "know" it's used for a dice roll, a card draw, who is playing right now or anything, that is not how game adaptations work.

2022 8 Dec 04:32 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The developers are keeping this thread open to "let others post their weird dice results". In that vein, here is another game for study. More 12's than 9's. Just as many 12's as 6's. No coincidence that the highest elo player was depending on 6 and 9.

boardgamearena.com/7/catan?table=324605871

2022 9 Dec 00:15 • LDZ777 • הצעה זו לא נותחה על ידי היזמים עדיין:

youre dice randomness program is cheap..... you need to get a better one!

2022 11 Dec 05:40 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=325502804

The lowest elo player had the worst starting position (only 18/36 chance of getting a card on any roll)

The highest elo player had the best starting position

But 5 and 9 were only rolled 8% of the time instead of the expected 22% (the chance of this is 3 in 1000), giving a strong advantage to the player with the lowest elo and the worst starting position (who easily won the game)

2022 14 Dec 23:56 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Try to guess which numbers the highest elo player did not settle on at the beginning of the game based on the dice distribution:
2 rolled 0%
3 rolled 3%
4 rolled 15%
5 rolled 17%
6 rolled 15%
7 rolled 7%
8 rolled 19%
9 rolled 5%
10 rolled 12%
11 rolled 3%
12 rolled 3%

5 and 10 are the answers. 9 is one of the numbers the lowest elo player did not settle on.

46% of the rolls were 4, 5 and 10 (chance of this happening is 1 in 1000)

boardgamearena.com/5/catan?table=326667525

2022 17 Dec 19:19 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=327462968

Double as many 3's rolled as sixes. The only numbers I didn't settle on were 3 and 4. 3,4,5 rolled 44% of all rolls (probability of that is 1 in 1000). Didn't expect this because there was another player at the table with higher elo, but it became obvious the dice would be biased against me after I didn't get any cards in the first five rolls of the dice.

2022 19 Dec 04:59 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/gamereview?table=327888718

As the highest elo player, I settled on every number except 11. Then 3 elevens were rolled in a row (1 in 6000 chance). Really painful loss (double as many 11's rolled as 5's)

2022 21 Dec 23:51 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

As the highest elo player, I once again settled on every number except 11. Result: twice as many 11's as 8's were rolled and I ended up with the least number of cards per turn at the table.

boardgamearena.com/table?table=328738647

2022 22 Dec 00:28 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I was the highest elo player. The only number I didn't settle on was 9. Then 9 rolled way more than expected. Meanwhile the lowest elo player did not settle on 5. There was only 1 five rolled the entire game. Just making observations of correlations between the numbers that are rolled more or less than their expected values and the relative elos of the players at the table.

boardgamearena.com/gamereview?table=328752182

2022 23 Dec 05:09 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/9/catan?table=329106473

7 in 10,000 chance of so many 5's, 6's and 10's. This time the dice were hurting the player with the lowest elo.

2022 23 Dec 05:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just as many 2's rolled as 8's. Dice favored me, the middle elo player, avoiding the only two numbers I didn't settle on (5 and 8) and going overboard on the numbers where I was the only one settled (2).

boardgamearena.com/9/catan?table=329111691

2022 23 Dec 08:25 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 2's than 8's in this game. One player settled on both 8's and both 10's and got clobbered. 6,8,10 were only rolled 20% of the time (1 in 1000 chance of that happening). 3 and 11 were rolled more than 6 and 8.

boardgamearena.com/5/catan?table=329135601

2022 23 Dec 20:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=329281366

8 rolled more than 6 and 9 combined. The numbers that were rolled more than expected corresponded to the settlements of the lowest elo player.

There were 3 ore hexes: 3,9, and 11. I settled 9 and 11. The lowest elo player settled 3. Then 3 was rolled more than 9 and 11 combined.

The elos of the players were 300, 308 and 102. The player with the elo of 308 ended up with one of the strongest positions I've ever seen in the game, with the highest sum of probabilities of getting cards, along with a 3-for-1 port. Yet he got last place and the lowest elo player won, with the weakest positioning.

2022 26 Dec 23:13 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/gamereview?table=330168361

This game rolled as if the dice had information about the game state. I (the highest elo player) settled on every number except 5,9,11. At one point in the game there was a string of 16 rolls where 14 of them were 5,7,9,11 (4 in 10,000 chance). That was enough to put me in last place. The moment I became last place, there was a string of 3's rolled in sequence (a number which only I was settled on).

The winner of the game was the lowest elo player, who never had a settlement on 6 or 8, and did not settle on ore until late in the game. The weird timing of the dice rolls caused his win.

2022 30 Dec 21:43 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I was happy to not be the highest elo player at this table. Fastest win ever in only 36 rolls! The numbers which would have disadvantaged me: 3,7,8,10,11 were only rolled 7 times in 36 rolls (probability 1 in 10K)

boardgamearena.com/4/catan?table=331412953

2022 30 Dec 22:15 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I was the highest elo player at the table, so had a disadvantage. The middle elo player had a strong dice advantage with 41% of all rolls being 8 or 4 (and his first settlements were both 8's and both 4's). There is a 1% chance of rolling so many 8' and 4's. There were double as many 2's rolled as 9's.

boardgamearena.com/7/catan?table=331425862a

2022 30 Dec 23:05 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Another game decided by the random number generator. It's rare for me to get last place among such poor players. It was all decided by the random number generator. More 11's rolled than 6's (the lowest elo player was the only player settled on 11. I settled on every number except 11.

boardgamearena.com/7/catan?table=331431482

2022 31 Dec 04:32 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 12's rolled than 6's (the lowest elo player was the only one not settled on 6). 1 in 1000 chance of so many 12's. Lowest elo player easily won (others had 3 and 4 points)

boardgamearena.com/7/catan?table=331503198

2022 31 Dec 21:18 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

There were as many 12's rolled (2 in 58 rolls) as 5's and 8's. 3 in 10K chance of so few 5's and 8's

boardgamearena.com/8/catan?table=331713624

2023 3 Jan 19:40 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=332644879

More 10's rolled than any other number. Is it a coincidence that the highest elo player settled on every number except 10?

2023 3 Jan 20:15 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=332650646

Here's is an example of a rare game that doesn't fit into my usual observations. As the highest elo player, I settled on every number except 4 and 8. So I expected 4 and 8 to be rolled more than their expected probabilities. 8 was over-rolled. But 4 was rolled much less than it should. And 9 was over-rolled (the middle elo player did not settle on 9). Usually I find the over- and under-rolled numbers end up hurting or helping one particular player, but in this case, I don't see any of the usual correlations between extra rolls and elo.

2023 12 Jan 04:10 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/9/catan?table=335461569

Dice strongly supported the lowest elo player, who didn't settle on 6 and had two cities on 9.
6 rolled 6% (was less frequently rolled than 3)
9 rolled 19%
4 rolled 17%
(5 in 1000 chance of so many 4's and 9's)

2023 13 Jan 20:59 • Grodar • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=336032168

Dice roll: rate of 2 (~3%) 2
Dice roll: rate of 3 (~6%) 6
Dice roll: rate of 4 (~8%) 19
Dice roll: rate of 5 (~11%) 8
Dice roll: rate of 6 (~14%) 2
Dice roll: rate of 7 (~17%) 21
Dice roll: rate of 8 (~14%) 19
Dice roll: rate of 9 (~11%) 8
Dice roll: rate of 10 (~8%) 10
Dice roll: rate of 11 (~6%) 4
Dice roll: rate of 12 (~3%) 0

I had similar distribution in my previous games. It seems that this is not random Gaussian distribution.

2023 18 Jan 03:08 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Thanks @Grodar for posting this game - it's an excellent case study in the errors of the random number generator.

Notice a few things: in 48 rolls, one player ended with 11 points and one player only had 2 points. This was not because of robber play, but because of the random number generator. Both players started with the same probability of having their numbers rolled (19 out of 36 chance of getting a card on any single roll). Yet the middle elo player was doubly settled on 4 and 8 (which were rolled way out of proportion). The middle elo player's numbers (4,8,10) were rolled 50% of the time, when it should be about 30%. The highest elo player was doubly settled on 6 (which only rolled once in 48 rolls - a 3 in 500 chance). In the end, the middle elo player received THREE TIMES as many cards over the course of the game as the next player in spite of having the same starting position probability.

Must have been an incredibly frustrating game to play.

2023 18 Jan 03:49 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/9/catan?table=337549710

Half as many 9's rolled as 12's

2023 19 Jan 20:22 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=338092642

3 in 1000 chance of so many 8's rolled

2023 19 Jan 21:24 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Highest elo player settled every number except 6 and 11. Then 6 and 11 were way over-rolled, especially early game (when it matters most). Falls into the typical pattern I've observed.

21% sixes when it should be 14%

boardgamearena.com/4/catan?table=338106225

2023 23 Jan 04:16 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

4 in 1000 chance of so many sixes and nines

Just as many 8's rolled as 11's

boardgamearena.com/5/catan?table=339274034

2023 24 Jan 07:29 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Lovely example of typical dice bias. Lowest elo player settled only 5,9,10. 5 and 10 over-rolled. 7 also over-rolled, with vast majority of 7's rolled by lowest elo player. Two highest elo players settled 4 and 6. They only rolled 13% of the time when it they should be a combined 22%.

boardgamearena.com/4/catan?table=339656315

2023 31 Jan 10:07 • iAreku • הצעה זו לא נותחה על ידי היזמים עדיין:

This game dice is FAR from RANDOM. Our table 342109558 had this beginning

7
7
7
7
12
7
7
4
4
4
4
7
...

And then a mix of 7s all over the game....

2023 31 Jan 21:36 • chobo21 • הצעה זו לא נותחה על ידי היזמים עדיין:

Table ID #342287246 Bug, cheating or just strange luck?

Winning player rolled 7 seven out of their eighteen rolls. Myself and third player each rolled two 7's apiece. I'm salty because I thought I objectively had the best starting position and probability to win.

2023 1 Feb 05:14 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=342413921

Lowest elo player was the only one to settle 4 and didn't settle 6.

Result: 24% fours!!! :o (1 in 10,000 chance) and only 1 six rolled in 45 total rolls.

2023 3 Feb 16:45 • daysleeper8 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=343224552

Some matches have predictable dice rolls, like in our previous match.
I had the best odds (had all numbers except 6 and 2) but I ended with 0.6 resources per turn.

Dice roll: rate of 2 (~3%) 2
Dice roll: rate of 3 (~6%) 8
Dice roll: rate of 4 (~8%) 6
Dice roll: rate of 5 (~11%) 9
Dice roll: rate of 6 (~14%) 21
Dice roll: rate of 7 (~17%) 11
Dice roll: rate of 8 (~14%) 20
Dice roll: rate of 9 (~11%) 6
Dice roll: rate of 10 (~8%) 9
Dice roll: rate of 11 (~6%) 6
Dice roll: rate of 12 (~3%) 3

2023 4 Feb 05:02 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Normally I find I'm the highest elo player at the table and the dice are against me. This time I was the second highest elo and the dice favored me strongly (I was the only one settled on 12 and it rolled more than 7)

boardgamearena.com/1/catan?table=343431035

2023 4 Feb 21:16 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Once again I was not the highest elo at the table and the dice favored me strongly. Almost half the rolls were 3,4, 10 (chance of this is 1 in 1000). There were 8 times as many 3's rolled as 7's (24% threes and just 3% sevens). The highest elo player did not settle on 3. I ended up with an average of 1.7 cards per roll. The random number generator won this game, not my skill.

boardgamearena.com/4/catan?table=343672786

2023 4 Feb 21:53 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

2 in 10,000 chance of so many 9's

boardgamearena.com/8/catan?table=343682695

2023 10 Feb 01:24 • Xeen_Dread • הצעה זו לא נותחה על ידי היזמים עדיין:

The dice drop in clusters and then normalize in the 2nd half of the game, by which point it doesnt matter that you finally get your numbers because whoever started on the good numbers is now settled on yours too.. I've noted many times where weird stuff happens. Also seems that the dice favor high karma and premium members.

2023 10 Feb 13:11 • 13Hattrick • הצעה זו לא נותחה על ידי היזמים עדיין:

It's unbelievable how many games become so bad and annoying because of unrealistic statistics. It can't be that so often numbers like 3 4 or 11 are rolled more often than 6 or 8. Do you developers even know what "Normal distribution" (Gaussian distribution) is? And also as many have reported, players with a low elo are very often "helped", do you do this on purpose to make the game more entertaining? it makes the game worse!!! Just let the numbers be rolled realistically.

2023 11 Feb 17:05 • NachoLibre • הצעה זו לא נותחה על ידי היזמים עדיין:

Table ID 346005727, more 10(6), 11(6), 12(5) and 3(5) than 6(4).... more 5(11!!) than any other number...
It's just one of many examples, I really can't believe this anymore, you really need to do something, otherwise a premium version of this website is really pointless. Why should I pay to play when such statistics just take the fun away?

2023 23 Feb 00:33 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=349901704

Imagine this: at one point in the game
I settled 3 (twice), 4, 5 (twice), 6 (twice), 8 (twice), 9, 10 (twice), 11 (twice), and 12

I realize the only way to get past the dice bias when I'm the highest elo player is to settle on as many numbers as possible.

At the same point in the game, the winner had
4, 6, 9 (twice), and 10

They ended up with just as many cards as me. This was all because of dice timing (my numbers rolled when the robber was there. 9 rolled 21% of the game 😲)

The other opponent had
4, 5, 6 and 8 (twice)

They both took a port on their initial placement. Probabilistically, I should have had waaaaay more cards than them. But I didn't. I ended with just 5 points.

2023 23 Feb 17:44 • GTSchemer • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers18 since you are a math professor, I wanted you to have concrete information. I am a computer engineer with multiple degrees and decades of experience in software, so I know what it is like to be knowledgeable in a field. I also develop BGA games so I have seen and write the source code.

WHAT IS POSSIBLE:

It is conceivable that a game implementation can use a pseudorandom number generator poorly, producing bias. For example, see en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle#Potential_sources_of_bias and how implementation errors can cause it (this is not related to Catan, but is a general observation).

It is conceivable that a pseudorandom number generator can behave poorly, producing an uneven distribution over its expected range, or clustering numbers in a statistically invalid way over large numbers of samples.

WHAT IS NOT POSSIBLE:

Your elo is irrelevant. BGA implementations use a call like this, which is incredibly simple:
$new_value = bga_rand(1, 6);

This produces a new dice value that is either 1, 2, 3, 4, 5, or 6.

As a math professor you're well aware of function inputs and outputs. Note in this case, the API call has zero knowledge of elo, zero knowledge of which values might be good or bad, and zero knowledge of anything except for the minimum and maximum requested integer.

I tell you this to urge you, with respect, not to spread misinformation about PRNG values being related to rank. The source code is incredibly simple and easily verified.

What you COULD do, if you wish to do statistical analysis, is make a spreadsheet over a large number of games, tracking the dice results in each game WITHOUT considering elo, and determining the overall distribution of die results after (for example) 50 games. If such a large number of games still has an abnormal distribution, then this might be evidence to examine the bga_rand() API call more closely.

2023 6 Mar 20:07 • yfimc • הצעה זו לא נותחה על ידי היזמים עדיין:

I am also seeing super weird dice clustering like some games with tons of 11, 12 and 2 and almost no 8, 9. maybe 11 will roll 5 times in a row. i know these are not impossible but the dice rolls don't seem to follow the bell curve distribution i would expect, even over large samples!

2023 10 Mar 22:58 • nocheatspls • הצעה זו לא נותחה על ידי היזמים עדיין:

czy kości są losowo rzucane?

2023 10 Mar 23:15 • nocheatspls • הצעה זו לא נותחה על ידי היזמים עדיין:

KTO DO CHOLERY NALICZA TE PUNKTY??????????????

I KTO RZUCA KOŚĆMI ŻE TYLKO JEDEN GRACZ DOSTAJE 7 ? W KOŚĆIACH?

2023 3 Apr 06:55 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The dice are just weird. Here's an example. One fourth of the rolls were 9. No threes rolled in 63 rolls.

boardgamearena.com/1/catan?table=363134109

2023 3 Apr 07:30 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

When I blocked the 3 at the beginning of the game, my opponents were confused. Why block a 3 when you could block a 6 or an 8?

Turns out 19% of the rolls were 3 and 7% of the rolls were 8. How could I possibly know this at the beginning of the game?

boardgamearena.com/4/catan?table=363140553

2023 3 Apr 21:49 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

This game is interesting to analyze, since it fits some patterns I've been observing.

At first glance the final game statistics look pretty "normal", but some things jump out
Blue settled only 5,9,10 at the initial setup. 5,9,10 then rolled much more than its expected value
White settled only 8,3,4, whose combined total rolls were just as expected.
I needed 11 to be rolled to win, but it only rolled 4%, while 3 rolled 8%
I've also been noticing that some players roll significantly more 7's in some games than other players. In this game, blue rolled 9 sevens, while white rolled only 1 seven.

As a note for myself, blue had the lowest elo at the table. These are observations. Observations and hypotheses can never, by definition, be "misinformation"

boardgamearena.com/1/catan?table=363340432

2023 3 Apr 22:24 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

At the beginning of this game, I commented that I expected the dice to help me. I was right. 4 rolled 18% of the time when it should be 8%. My position was weaker than my competitors. Probabilistically, I should not have won.

boardgamearena.com/4/catan?table=363369076

2023 6 Apr 06:47 • firstpn • הצעה זו לא נותחה על ידי היזמים עדיין:

if you can't fix this problem . I'll stop playing permanently

2023 19 Apr 22:25 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

2 in 100,000 chance of so few 3's and 5's

boardgamearena.com/1/catan?table=368797125

2023 24 Apr 00:17 • Gussto • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm not sure if it's a bug or if users are able to manipulate the game. The other players were hard on 4's and 8's half way through the game. Once positioned around said 8's, 90% of the rolls were 8's 4's and 7's. I've never seen anything like it.

Table #370186736

Dice roll: rate of 2 (~3%)
4
Dice roll: rate of 3 (~6%)
4
Dice roll: rate of 4 (~8%)
11
Dice roll: rate of 5 (~11%)
6
Dice roll: rate of 6 (~14%)
8
Dice roll: rate of 7 (~17%)
23
Dice roll: rate of 8 (~14%)
21
Dice roll: rate of 9 (~11%)
8
Dice roll: rate of 10 (~8%)
8
Dice roll: rate of 11 (~6%)
9
Dice roll: rate of 12 (~3%)
0

2023 24 Apr 21:24 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=370473334

More 11's than 8's or 6's

2023 24 Apr 23:00 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=370508173

More than double as many 10's as expected. This benefitted the orange player, who also happened to roll 10 sevens through the game while the other players only rolled 7 three and four times respectively. This pattern (where the person who rolls the most sevens also happens to get assistance from the dice on other numbers) has happened often enough that I would consider it noteworthy enough to examine more closely.

2023 25 Apr 11:22 • ginalucia2 • הצעה זו לא נותחה על ידי היזמים עדיין:

Is it really random when a player gets 7 in all of their turn?

2023 26 Apr 05:12 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Dice were so weird on this game. So many strings of 3's then strings of 5's. No 9's till the end of game

boardgamearena.com/4/catan?table=370928634

2023 26 Apr 05:28 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=370928634

Same game as above
1 in 100,000 chance of a string of 5 threes in 7 rolls
4 in 100,000 chance of a string of 5 sixes in 8 rolls
Less than 1% chance of only two nines in 74 rolls

2023 3 May 14:27 • goldlyn • הצעה זו לא נותחה על ידי היזמים עדיין:

incredible # of 7's rolled. Eg, of my first 4 rolls in recent game, 3 were 7's. It's become common to have 7 rolled 2 times in a row. Agree with previous comments; dice rolls are hinky.

2023 3 May 23:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Really weird stats in this game. 10% 2's and 12's. Only 1 three.

boardgamearena.com/4/catan?table=373547535

2023 4 May 22:07 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

5 rolled 4x as much as 6
boardgamearena.com/table?table=373846133

2023 4 May 22:46 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

12 rolled as much as 5
4 rolled more than any other number

boardgamearena.com/8/catan?table=373850733

2023 5 May 01:12 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

8 in 10,000 chance of so many 8's, 10's and 4's, to the benefit of one player

boardgamearena.com/5/catan?table=373899016

2023 5 May 02:53 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=373910002

I had such a great setup - six was both wheat and ore. I settled both. I also settled the 8 ore. But I did not get as many cities as the lowest elo player who settled the 10 ore (that was their only ore source). Because 10 rolled way more than 6. In fact, 6 rolled just as rarely as 11. The other high elo player also settled both ore and wheat 6, but ended up with no cities. These dice make it impossible to play with normal basic strategy. My only winning strategy is "settle every number except 2 and 12"

2023 5 May 03:32 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=373914988

More 12's than 4's

2023 5 May 22:28 • goldlyn • הצעה זו לא נותחה על ידי היזמים עדיין:

The computer rolled a 7 three times in a row. In the Catan game before this one, my first 3 out of 4 dice rolls were a 7. It's impossible to play a game when no move is possible because the computer keeps turning up 7's.
table #373853551
Move #41
Progression 0%

2023 6 May 18:32 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=374449603

21% fours! Ended in just 34 rolls. Has me wondering what the hypothetical smallest number of rolls is needed to win the game.

2023 7 May 00:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

15% twos!

2023 7 May 00:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

15% twos!
boardgamearena.com/8/catan?table=374530225

2023 7 May 00:13 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Found a way to win in 13 rolls. Out of curiosity, can anyone win in fewer rolls?

2023 10 May 23:42 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=375873902

1 in 500 chance of so few 5, 9 and 11. I settled every number but 10. Then 10 rolled way more than expected. 3 was double its expected value. It just so happened that the player who benefitted from the 3's and 10's also rolled double as many 7's as the rest of us. I've noticed this pattern in other games as well. This time one difference from the norm is that the dice bias went towards the player with the highest elo.

2023 12 May 05:05 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=376254097

Really weird distribution (double as many 11's rolled as 6's)
1 in 10,000 chance of so few 6's and 7's

2023 24 May 05:00 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Chance of so few 5 and 9 is 1 in 1000.

This is another example of a game where it's really obvious that the way that the roll distribution is different from a pure normal distribution is strongly correlated to the players' initial placements (and in this case, penalized the player with the highest elo, by chance perhaps, or for whatever reason)

boardgamearena.com/5/catan?table=380188073

2023 31 May 03:25 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Laughably bad

boardgamearena.com/4/catan?table=382422632

2023 31 May 08:18 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The way the 7's are distributed among the players are not random in general. Here's a prime example. The players with the lower ELO's rolled seven 11 times and 7 times respectively. The player with the highest ELO at the table rolled 7 once.

boardgamearena.com/1/catan?table=382466496

2023 31 May 23:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

20% fives, which strongly helped me, the highest elo player at the table. I also rolled way more 7's than usual.

boardgamearena.com/7/catan?table=382703627

2023 1 Jun 00:51 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

5% eights? Really?

1 in 10,000 chance of so few 3, 8 and 9

boardgamearena.com/8/catan?table=382713595

2023 5 Jun 21:32 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 11's than 5's
24% sixes

The only two numbers the highest elo player did not settle on initially were 6 and 11. Coincidence again??

boardgamearena.com/7/catan?table=384272045

2023 6 Jun 06:19 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 11's and 12's rolled than 6's.

Is it a coincidence once again that the highest elo player was the only player who settled 6?

boardgamearena.com/table?table=384288729

2023 6 Jun 07:35 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

25% 3's and 4's
11% 5's and 6's (and most of those only at the end of the game)

boardgamearena.com/7/catan?table=384402695

2023 6 Jun 08:02 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

These dice are happy on the fringes

More 2's than 9's
More 11's than 7's

boardgamearena.com/table?table=384415171

2023 7 Jun 04:38 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

More 2's than 4's
More 3's than 8's
Actually 6 times as many 3's rolled as 4's

boardgamearena.com/community

2023 8 Jun 08:08 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Four times as many 8's as 6's
Twice as many 10's as 7's
All incidentally helping the lowest elo player at the table

boardgamearena.com/1/catan?table=385062522

2023 8 Jun 08:14 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/1/catan?table=385062522

This is one of the worst skews I've seen in a long time. Less than 1 in 1000 chance of so many 8,5,10 (all benefitting one player) and so few 6,7

A pattern I've noticed is that when one player benefits from the dice skew, they also tend to roll more sevens. This is a good case in point. The winner rolled 8 sevens while the rest of us each rolled one 7.

2023 9 Jun 06:44 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=385370840

As soon as I picked my initial settlements, I commented that my choices pretty much guaranteed that 4 would not roll (I was the highest elo at the table and I settled on two 4's, one of which was the highest source of brick on the board). Sure enough, 12 rolled more than 4 (until the very end, when they ended with the same percentage of rolls). The person who had one settlement on an 11 brick had more brick throughout the game than my two settlements on 4 and one on 3.

Question: how could I possibly know this would happen??

2023 9 Jun 07:27 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

At the beginning of this game, after the initial placements, I commented in the chat box that 3,4, and 11 would roll more than expected. Sure enough, they rolled 33% when it should be close to 19%.

How did I know this would happen?

boardgamearena.com/4/catan?table=385372698

2023 9 Jun 07:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/4/catan?table=385372698

In this game 11 rolled more than 8. The player who settled an 11 ore (and had an elo of 100ish) easily won over the player who settled 8 ore and 8 wheat, as well as 10 brick and 10 wood (and had an elo of 325ish - one of the top 100 players).

One of the patterns I've noticed happened in this game as well - that the 100ish elo player who benefited from a disproportionate number of 3's, 4's and 11's also rolled the majority of the 7's (four 7's compared to one and two 7's for the other players).

2023 9 Jun 08:08 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I have a theory that the dice tend to favor players with lower elo, based on the initial placements.

Based on that theory, what should happen when the highest elo player settles every number except 2 and 12 (a common strategy of mine, given my observations)? We should see a lot more 2 and 12 than usual.

Here, the lowest elo player settled the 2 ore, while the highest elo player settled 10 ore.

2 rolled 7% of the time, while 10 rolled 4% of the time, as predicted.

boardgamearena.com/8/catan?table=385383356

2023 12 Jun 05:59 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I have noticed a correlation between relative elo and dice bias, but it could be that I'm missing a link (seeing elo as the factor, when it's actually based on something related to elo, like a bias against the current leader, or the player who gets the strongest initial settlements, or something else)

In this game, the dice strongly favored me, the middle ELO at the table, and hurt the lowest ELO player. I have never averaged so many cards per roll before (1.5!!).

Six was rolled 36% of the time. The probability of that occuring is 1 in 10,000.

boardgamearena.com/5/catan?table=386335131

2023 14 Jun 22:32 • Just Keep Climbing • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/7/catan?table=387214976

7 was rolled 5 times in a row. At first I was excited for the robber to finally move off my 8 (which I had 2 settlements on and a port for), but after all the dust settled, the robber was back on my 8. 7 did not get rolled after that for the remainder of the game. A lot of 8's were rolled, but I did not use that port once the entire game :(

2023 15 Jun 07:39 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@plshavemercy, some observations of your game
1. You had the highest elo at the table
2. The fact that the dice might be biased against you was evident after you only got one card in the first eight rolls of the dice (and that from an 11), in spite of a strong choice of initial settlements
3. By the end of the game, you had a settlement on every number except 12. And yet, you only averaged 0.5 cards per roll...
4. The string of five 7's in a row happened when you had the monopoly card in your hand (I feel this is not a coincidence as I have seen this before). The chance of five 7's in a row is 1 in 10,000

2023 18 Jun 04:52 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=388239170

Three 11's rolled before the first 6.
All in all, 13% 11's and 8% 6's.

I was the highest ELO player at the table. I settled every number except 8, 11 and 10. I eventually settled every number except 2. My initial settlements gave me a 23/36 probability of getting a card on any given dice roll.

The lowest ELO player settled 2,5,8 and 10 at the beginning of the game. That's it! They chose 2 ports! And one had a 2! Their probability of getting a card on any given dice roll was 13/36 - almost half my chances. Yet they ended up with a fist full of cards while I sat patiently waiting and waiting for the first 6 to roll, as 8 rolled over and over.

By the end of the game, the lowest ELO player received on average 1 card per turn, while I got 0.7 cards per turn - this was all due to the biased dice.

2023 18 Jun 05:38 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The lowest elo player ended the game with a 46/36 chance of getting a card on any given roll.

The highest elo player ended the game with a 56/36 chance of getting a card on any given roll (based on settlements).

Yet the lowest elo player got 0.9 cards per roll and the highest elo player got, on average, 0.6 cards per roll.

boardgamearena.com/7/catan?table=388236876

2023 19 Jun 05:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Highest elo player settled 6 ore
Lowest elo players settled 3 ore and 11 ore

6 ore should roll more than 3 ore and 11 ore combined

And yet... 22% 11's and 3's (should be 11%)
7% sixes (should be 14%)

There were three times fewer sixes than 11's and 3's

boardgamearena.com/10/catan?table=388569894

2023 19 Jun 08:11 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The highest elo player was the only player who settled 3.

Then 3 rolled less than any other number, in fact twice as many 2's rolled as 3's. (Also strange was 43% of all rolls were 6 or 8 when it should be 28%)

boardgamearena.com/5/catan?table=388589607

2023 19 Jun 14:15 • bocram • הצעה זו לא נותחה על ידי היזמים עדיין:

can you check this statistics?
for me it's not plausible only 8 and 4.
table #388655356

Dadi: percentuale di 2 (~3%)
9
Dadi: percentuale di 3 (~6%)
2
Dadi: percentuale di 4 (~8%)
22
Dadi: percentuale di 5 (~11%)
7
Dadi: percentuale di 6 (~14%)
11
Dadi: percentuale di 7 (~17%)
11
Dadi: percentuale di 8 (~14%)
22
Dadi: percentuale di 9 (~11%)
11
Dadi: percentuale di 10 (~8%)
0
Dadi: percentuale di 11 (~6%)
0
Dadi: percentuale di 12 (~3%)
4

2023 20 Jun 22:27 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/8/catan?table=389134544

The randomizer REALLY didn't want me to win (I had the highest elo at the table)

More 11's rolled than 5's or 6's (the only numbers I didn't settle were 3,4, and 11. 4 also rolled more than 5 or 6).

I only needed a knight or a victory point to win (19/25 chance at start of game). Instead I pulled both monopolies, the road builder and year of plenty!

2023 20 Jun 22:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@bocram
Thank you for posting. This game fits a lot of patterns I've been observing.

Notice that there were no 10's or 11's rolled (1in 1000 chance).

You were the highest ELO at the table. You also had the strongest starting settlements, with a 21/36 probability of getting a card on any given roll. Yet you ended the game with only 0.5 cards per roll on average.

The lowest ELO player had the weakest starting position, with a 19/36 probability of getting a card on any given roll. Yet they ended the game with 1.4 cards per roll on average.

This is the sort of initial settlements that encourage dice bias, in my experience: when one of the lower elo players at the table doubles up on a number on their initial choice of settlements (here one player doubled up on both 8 and 4, and then they rolled a combined total of 44% 😳). When I am not the highest elo at the table, I try to double up on a number in the initial settlements - this strategy has been working for me. When I'm the highest elo, I settle every possible number except 2 and 12.

2023 20 Jun 22:44 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

The probability of so many 4's and 8's in @bocram's game is 7 in 10,000

2023 20 Jun 23:20 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

table=388822860 (for exemple)

I apologize to the developper for "Repeatedly reporting the same thing after it being explained that this is not a bug and that the dice work with a very fair algorythm is aggressive towards the developer and causes unnecessary work for them. The dice are fairly balanced".
i did'nt saw your response (i though i will me notified in my profil)

I was very frustrated because i love this game and did'nt thought dreaming.
Very happy to see this bug report... that i just discovered.
Thank You all for that ! I'll sleep better.

One of the best inspiration : If you can't fix this problem . I'll stop playing permanently

2023 21 Jun 00:03 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@GTSchemer

I thank you and the other developers for this incredible site

I have no doubt that Catan calls
$new_value = bga_rand(1,6)
every time it rolls the dice. I've done enough coding myself to realize how much easier it is to program a simple random dice call than something more complicated. I doubt that the bga_rand function is non-random enough to explain all the weirdness we're all observing.

I actually don't blame the catan developers.

I'm wondering if there might be another function at play at a more site-wide meta-level that encourages low ELO players to keep playing by purposely giving them a little extra luck every now and again. I only say this because I have observed similar weirdness in the randomizer (working as a function of relative elos at the table, but not uniformly (i.e. occasionally the highest elo player will get the extra luck)) while playing Viticulture on this platform. I have no idea how such a meta-level function would be coded - it seems extremely complicated. In Catan, it seems to know where the ore hexes are, where the sheep hexes are, at what moments a 7 would be devastating and at what moments it would be beneficial, and sometimes what development card I need next.

In Viticulture, it seemed to know which mamas and papas were best, which set of vines cards would destroy your chances, and which order cards did not synch with which vine cards. It also seemed to know the best visitor cards.

I have not observed this yet in other games on this platform, but it might be especially evident in Catan...

A lot of these observations could simply be the luck of the draw. In any given game, one player will be lucky and the others unlucky, by default. The reason I stop and wonder is because too often the luck in one dimension goes hand in hand with luck in other dimensions, to the point where I can predict how the final distribution will differ from the normal. I should not be able to make such predictions (which are correct about ,80% of the time) if there is no mistake in the randomizer.

2023 21 Jun 03:17 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

At the beginning of this game, after the initial choice of settlements, I said, "this game is going to have a lot of 8's"

As the highest elo player at the table, I settled every number except 8 and 12, after I got my third settlement.

Sure enough, 8 rolled 23% of the time (should be close to 14%)

2023 21 Jun 03:18 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

23% eights, as predicted

boardgamearena.com/5/catan?table=389193189

2023 21 Jun 08:18 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Here's an example where the random number generator seems to know where the sheep, wheat and ore are.

I had the highest elo and settled every number except 11 and 12. I had a wheat 6, an ore 5 and a sheep 3.

Before I was able to settle the 8, it rolled like crazy. As soon as I settled the 8 (meaning I had settled every number except 11 and 12), there were a string of times interesting rolls: 7 (for opponent), 7 (for opponent), 11 (for me), 7, 3 - i.e. in the six rolls after I settled every number, I managed to lose cards rather than gain.

All in all, 5 and 6 rolled 13% of the time (about half as much as the 25% that they should roll) and 3 rolled 13% of the time (double the expected 6%). I was still able to win, but felt like I was playing against a fourth invisible player: the dice.

boardgamearena.com/4/catan?table=389259270

2023 21 Jun 21:26 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

What happens when the highest elo player settles every number except 12?

12 rolls double as much as it should (6% instead of 3%)

boardgamearena.com/4/catan?table=389447053

2023 23 Jun 06:47 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Like there's a nefarious agent conspiring to assist the lowest ELO player every time.

Hmmm... I see the highest ELO player settled both ore and wheat on 5's? I know! Let's make 11 roll double as often as 5 and let's make 12 roll as much as 5! That'll do the trick. And while we're at it, why not let the lowest ELO player roll 7 twelve (!!!) times while the highest ELO player never rolls the robber? Oh look, it worked! The player with an ELO of 130 and a weaker starting position actually beat the players with ELO's in the 200's and 300's. Another underdog victory for the records. Yippee!

boardgamearena.com/7/catan?table=389875137

2023 24 Jun 15:50 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

Table #390296931
This game was a real joke (I was the third Elo level)

Durée de la partie 58 min
Niveau moyen des joueurs 150
Score moyen 6.25
Cartes développement restantes 21
Lancer de dés : taux de 2 (~3%) 3
Lancer de dés : taux de 3 (~6%) 2
Lancer de dés : taux de 4 (~8%) 5
Lancer de dés : taux de 5 (~11%) 15
Lancer de dés : taux de 6 (~14%) 6
Lancer de dés : taux de 7 (~17%) 29
Lancer de dés : taux de 8 (~14%) 19
Lancer de dés : taux de 9 (~11%) 18
Lancer de dés : taux de 10 (~8%) 2
Lancer de dés : taux de 11 (~6%) 2
Lancer de dés : taux de 12 (~3%) 0

with a specific series :

X lance les dés : 4,5
Y lance les dés : 4,5
Z lance les dés : 5,4
W lance les dés : 5,2
X lance les dés : 4,5
Y lance les dés : 5,4

I was nearbe two "6" tiles.

Hope there was no relation due to the -30 Karma for opening many bug reports
Positive message : we had many laught ...
No (other) comment.

2023 27 Jun 04:02 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/10/catan?table=391137430

24% nines
Five times as many 11's as 5's.

I was the highest elo at the table. I settled every number except 11. I knew the dice would be biased against me when two 11's rolled in the first 3 rolls. Lowest elo player easily won.

2023 27 Jun 20:23 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Beautifully fits typical patterns:

1. The lowest ELO player at the table rolled the most 7's
2. The only number the lowest ELO player did not settle was nine. Then nine rolled as infrequently as 2
3. The lowest ELO player doubly settled 4. Then 4 rolled way more than expected.
4. The lowest ELO player had the highest number of cards per roll (1.1) in spite of not having the strongest starting position.

I'm working on quantifying these observations statistically over the set of all 3-player games that I've played (over 500 games). Hopefully that's enough. If anyone knows where I can go to find all 3-player Catan games ever played, I would love the info to improve my stats.

boardgamearena.com/5/catan?table=391345403

2023 27 Jun 21:01 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Here's an example that doesn't fit the pattern in every way

1. The winner did not settle 10. Then 10 rolled half as much as expected
2. The winner rolled the most 7's

This time these advantages went to the highest elo player at the table (it happens occasionally).

boardgamearena.com/4/catan?table=391351548

2023 27 Jun 21:44 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm starting to get bold in the chats. As soon as this game began, I predicted

1. There will be more 4's than 9's
2. 3 and 6 will be quite rare

Sure enough, more 4's than 9's and 3 and 6 both rolled about half as much as expected.

boardgamearena.com/4/catan?table=391369459

2023 28 Jun 20:47 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

another joke : 29 9s in table=391676721
37 min
Niveau moyen des joueurs
180
Score moyen
6.25
Cartes développement restantes 17
Lancer de dés : taux de 2 (~3%) 4
Lancer de dés : taux de 3 (~6%) 2
Lancer de dés : taux de 4 (~8%) 4
Lancer de dés : taux de 5 (~11%) 10
Lancer de dés : taux de 6 (~14%) 13
Lancer de dés : taux de 7 (~17%) 13
Lancer de dés : taux de 8 (~14%) 15
Lancer de dés : taux de 9 (~11%) 29
Lancer de dés : taux de 10 (~8%) 6
Lancer de dés : taux de 11 (~6%) 2
Lancer de dés : taux de 12 (~3%) 2

2023 29 Jun 17:51 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

second time nearby two six since the beginning and coïncidence he came so few ...
more 11s than 6...

47 min
Niveau moyen des joueurs
170
Score moyen 8
Cartes développement restantes 15
Lancer de dés : taux de 2 (~3%) 1
Lancer de dés : taux de 3 (~6%) 1
Lancer de dés : taux de 4 (~8%) 7
Lancer de dés : taux de 5 (~11%) 14
Lancer de dés : taux de 6 (~14%) 9
Lancer de dés : taux de 7 (~17%) 17
Lancer de dés : taux de 8 (~14%) 10
Lancer de dés : taux de 9 (~11%) 16
Lancer de dés : taux de 10 (~8%) 10
Lancer de dés : taux de 11 (~6%) 12
Lancer de dés : taux de 12 (~3%) 1

2023 29 Jun 22:31 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

table=392029042
Each time i choose the position with 6, it's just a calamity ...
very strange "Random" : it's just incredible (see precedent post)
11 in second position ...

Lancer de dés : taux de 2 (~3%) 0
Lancer de dés : taux de 3 (~6%) 2
Lancer de dés : taux de 4 (~8%) 12
Lancer de dés : taux de 5 (~11%) 12
Lancer de dés : taux de 6 (~14%) 6
Lancer de dés : taux de 7 (~17%) 10
Lancer de dés : taux de 8 (~14%) 18
Lancer de dés : taux de 9 (~11%) 12boardgamearena.com/bug?id=66082#
Lancer de dés : taux de 10 (~8%) 12
Lancer de dés : taux de 11 (~6%) 14
Lancer de dés : taux de 12 (~3%) 4

2023 30 Jun 02:34 • redthedead • הצעה זו לא נותחה על ידי היזמים עדיין:

another joke : I had 3 town around the 10 (with ore). I let You guess ...
it's time for me to stop playing this awfull game.
It's just ... insane.

Durée de la partie
52 min
Niveau moyen des joueurs
144
Score moyen
7.5
Cartes développement restantes 8
Lancer de dés : taux de 2 (~3%) 5
Lancer de dés : taux de 3 (~6%) 5
Lancer de dés : taux de 4 (~8%) 8
Lancer de dés : taux de 5 (~11%) 15
Lancer de dés : taux de 6 (~14%) 16
Lancer de dés : taux de 7 (~17%) 14
Lancer de dés : taux de 8 (~14%) 9
Lancer de dés : taux de 9 (~11%) 11
Lancer de dés : taux de 10 (~8%) 4
Lancer de dés : taux de 11 (~6%) 11
Lancer de dés : taux de 12 (~3%) 2

2023 1 Jul 04:56 • CTG89 • הצעה זו לא נותחה על ידי היזמים עדיין:

Table ID: boardgamearena.com/table?table=392415145
Total rolls: 102
Total 9 appeared: 1 time

What is the probability of this happening?

2023 1 Jul 05:31 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@CTG89

The probability of one 9 in 79 rolls is 0.099029% (about 1 in 1000)

The end game stats give the (rounded) percentage of 9's rolled, not the total. In your game there were 79 rolls rather than 102 (which is the sum of the percentage from the rolls table)

2023 1 Jul 05:35 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@CTG89

Notice that you had the highest ELO at the table and settled twice on 9, and they were both ore hexes.

The other two players settled every number on the board between the two of them, except 2, 12 and 9. Is this still a coincidence??

You had the strongest starting position in the game. Probabilistically, you should have won.

2023 1 Jul 16:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just wanted to make a note here that this is likely a site-wide issue. If you search the posts, you'll find that there are complaints about the randomizer for Backgammon, Can't Stop, Forbidden Island, Saboteur, Martian Dice, Piraten Kapern and Tapestry.

Other experts like myself have noticed these issues (example comment pasted below)

I will continue the slow and tedious task of trying to analyze thousands of games and provide concrete evidence that this is the case (but I am a mathematician and not a statistician, so I may need to consult my statistician brother for support, and I also have other things to do with my time, so it may take a while).

"I'm a Statisical expert, Inmho. and others. I concur with others as above.

I've developed Medical software $ 1 Million install +\-, for multiple departments in large hospitals and research ctrs, medical records for diseases difficult to track, advise, schedule, supplies auto ordered, interfaced with other vmajor vendor, at point of contact. multiple advisories, appropriately offer algorithms for Risk Assessment of Death if patient goes to surgery chemo, radiation, etc. For patients with complicated diseases, Heart, Cancer, Diabetes, Orthopedics, High Risk Pregnancy, etc
7 graduate semesters in statistics. Invited speaker to International, National, State, Local agencies, CDC, NIH, NCI, co-laborated & presented to: with FBI, FEMA, LLNL, UCSF, UCB, EMSA, OSHA, NIOSH, CSTE, APHA, DOT, DOL, FDA....
I like game, but clearly NOT random.
I've been doing statistics since 1970s Mainframe, UNIX, PC, Mac, HP, IBM, Sun, SAS, SPSS, BMDP, MUMPS (used by VA and NASA).& currently Sr Epidemiologist III, working on Covid19.
It's a fun game- My thanks for all"

2023 1 Jul 22:45 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Just played two games of backgammon (I have low ELO since I've never played), and what incredible luck I had! 7 sets of doubles for me and only one for my opponent. Second game: 4 sets of doubles for me and one for my opponent. If my theory is correct, my luck should turn when I face an opponent with lower elo than me.

2023 1 Jul 23:00 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Sure enough, played a backgammon game against a player with lower ELO and my luck turned - I lost my first game. Just three games, so not statistically significant, but still....

2023 16 Jul 15:42 • Alison MacPrickle • הצעה זו לא נותחה על ידי היזמים עדיין:

Getting really fed up with crazy dice patterns. Happens all the time. If this is 'random' then please abandon random and write something that restrains these excesses which just ruin the game for everyone. Just played a game with these stats #397393671
2 - 3 rolls
3 - 0 rolls
4 - 16 rolls
5 - 8 rolls
6 - 29 rolls
7 - 8 rolls
8 - 8 rolls
9 - 11 rolls
10 - 0 rolls
11 - 16 rolls
12 - 3 rolls

2023 17 Jul 21:09 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Alison MacPrickle
The chance of so many 4's, 6's and 11's in a game is 3 in 100,000 😯
This greatly benefitted the middle elo player and hurt the lowest elo player
As an aside, the end fame stats show the percentage of rolls, not the total number of rolls (so there was 1 two rolled in a total of 38 rolls for the game)

2023 17 Jul 21:11 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I just finished 100 games of backgammon, and although the dice are weird there, their weirdness is NOT related to relative ELO the way it appears to be in 3-player Catan and Viticulture games.

2023 18 Jul 07:27 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

For whatever reason, this kind of initial set up (where multiple players double up on numbers in their initial settlements in a 3-player game) seems to encourage wonky behavior in the dice.

In this game, 12 rolled more than 4, 9 or 10.

5 rolled 6x more often than 4 when they should be roughly similar in their number of rolls.

Again, these are not shockingly unexpected. What is weird is to be able to predict this dice behavior before the first roll.

boardgamearena.com/4/catan?table=398003901

2023 7 Aug 11:41 • sergut • הצעה זו לא נותחה על ידי היזמים עדיין:

In every online playing website there will be people (with a poor understanding of statistics) that complain that the random number generator is not really random --- and acts against them. Funnily enough, it never happens the other way (people complaining that the dice favours them, that is attributed to luck).

It is not uncommon for surprising rolls to happen even in tabletop Catan with real dice: four rolls of the same number, 11s being rolled more than 6s, four Development Card with victory points being drawn one after the other... all of this can happen on a game of Catan using real dice and paper cards. There is nothing surprising about these things happening too in BGA's Catan.

See also boardgamearena.com/bug?id=64097, where a suggestion about a widget to show dice statistics has degenerated into lots of players complaining that the random number generator is buggy (when it is not).

2023 8 Aug 22:16 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@sergut
Agreed, there are many people who complain about random number generators who don't understand stats.

However in this case, there are multiple complaints by experts in stats (see July 1 post above). I myself have a graduate degree in math.

There are also a lot of people in the forums, myself included, complaining that the dice act strongly in their favor, making the game less fun (see Jun 11 post above).

If I can predict in advance the way in which the dice rolls will differ from a normal distribution, then the random number generator cannot possibly be working properly, right?

2023 22 Aug 02:28 • Baleraphon • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm playing with settlements on 8, 9, and 10. It's been 12 turns since any of those have come up on the dice. This is a repeated problem with these virtual dice.

2023 22 Aug 02:31 • Baleraphon • הצעה זו לא נותחה על ידי היזמים עדיין:

in this same game we went around the table 4-5 times with no 7s. The robber sat on my hex the whole time, and that number was rolled almost every single turn. Once the robber moved, that number was never rolled again.

2023 26 Sep 19:45 • hermilion • הצעה זו לא נותחה על ידי היזמים עדיין:

Dices are difinitly incorrectly implemented and it really makes that game unfriendly to play. Hate it really

2023 27 Sep 03:20 • hermilion • הצעה זו לא נותחה על ידי היזמים עדיין:

Also I bought premium just to play catan and that bug is making me to not ever again waste money on your company

2023 11 Oct 19:03 • LazyBore • הצעה זו לא נותחה על ידי היזמים עדיין:

Dodgy Catan game: boardgamearena.com/1/catan?table=427058192

0 rolls of 6. 23 rolls of 8

I invested heavily in two 6 resources at the beginning, and could not build a single thing!

2023 17 Nov 00:07 • Kieno • הצעה זו לא נותחה על ידי היזמים עדיין:

Table #440192065
9 came up repeatedly to the point one player thought another was hacking.

2023 21 Dec 00:25 • Coisinho • הצעה זו לא נותחה על ידי היזמים עדיין:

how come 4 comes out more times than 6,5 and 9 combined? after over 100 rolls? isn't this supposed to follow some sort of a normal distribution?

2024 5 Jan 09:31 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@LazyBore, I find your game interesting to analyze
boardgamearena.com/1/catan?table=427058192

A couple things to note:
1. The player who won with 11 points started the game with the lowest probability of getting a card on a given roll. (The other two players ended the game with 2 and 3 points).

2. The winner started the game with the lowest elo. The player with the highest elo settled two 6's and 6 never rolled for the whole game

3. The lowest elo player settled two 8's and two 9's which both rolled like crazy. I have noticed in three player Catan games when a player settles two hexes with the same number at the beginning of the game that those numbers roll either in relatively extreme overabundance or underabundance (relative to the expected number of rolls)

2024 17 Jan 08:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Double as many 9's rolled as 8's.

I had 8 ore and 5 wheat and should have had the most cities on the board (probabilistically), but 11 rolled more than 5, and both 9 and 10 rolled more than 8, to the advantage of the lowest elo player at the table.

boardgamearena.com/4/catan?table=463224316

2024 18 Jan 17:00 • lux1101 • הצעה זו לא נותחה על ידי היזמים עדיין:

Please fix the random number generation! I bought premium just to play Catan but this bug makes me going crazy...

2024 20 Feb 02:50 • MichalisRG • הצעה זו לא נותחה על ידי היזמים עדיין:

I just played a game that the dice had results above 7 like 5 times in all game... And i see in many of my games that the dice results are not seeming random at all. We buy premium to support the site and have more quality games. At least you can try and fix some problems that many people report. If the games are unplayable then there is no reason to keep playing here.

2024 29 Mar 18:03 • Coisinho • הצעה זו לא נותחה על ידי היזמים עדיין:

Imagine a game where you are able to secure 5, 6, 8 and 9 from the start (all different resources) and you get humiliated (only 4 points) by someone with 3, 4, 6 and 10 (with 3, 4 and 10 being absurdly hot). 9 only came out twice and 5 came once maximum. 6 and 8 were as common as 11, which was just slightly more than 9 and 5. Considering the odds of this and that something like this happens way to often, it makes one suspect about the randomness of the dice.

2024 29 Mar 18:14 • greenjacket124 • הצעה זו לא נותחה על ידי היזמים עדיין:

Anyone who is actually suspect of the dice just does not actually understand randomness.

2024 17 Apr 23:08 • szypd • הצעה זו לא נותחה על ידי היזמים עדיין:

still buggy.

2024 27 Apr 10:24 • Goodtogo24 • הצעה זו לא נותחה על ידי היזמים עדיין:

soo many complaints of dice please fix it...just played a game 7 rolled 4 times in a row. great board game is bad because of dice rolls...

2024 6 May 00:36 • hermilion • הצעה זו לא נותחה על ידי היזמים עדיין:

2 years and you do nothing about serious bug ... great

2024 6 May 02:17 • greenjacket124 • הצעה זו לא נותחה על ידי היזמים עדיין:

.....You can't fix something that is not broken.....

2024 17 Jun 05:28 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/table?table=526209649

9 rolled as infrequently as 2.

2024 18 Jun 00:45 • Tisaac • הצעה זו לא נותחה על ידי היזמים עדיין:

Catan is very luck-dependent, just look at the max ELO of all time, it's not even 500 !! Compare that to Agricola, Caverna, Santorini, Welcome To, where top players are reaching >700 ELO.
Even very good Catan players can't achieve a high ELO because the result depends a lot on the rolls, just don't play it if you can't handle that. I personnally can't handle that much luck so I don't play it, neither IRL nor online.
Don't look up excuses with randomness or conspiracy theories.

2024 18 Jun 04:46 • Nekojin • הצעה זו לא נותחה על ידי היזמים עדיין:

All of those games that are posted are showing a bunch of people (and one person in particular, over and over again) are falling victim to the Law of Averages fallacy. After you've rolled a 7 three times in a row, the chance of getting a 7 on the next roll is still one in six. Past events do not influence future events. Looking at a short sequence of numbers (and, let's be clear, a single game of Catan is a VERY short sequence) and drawing conclusions from that after the fact is worthless nonsense.

en.wikipedia.org/wiki/Law_of_averages

Picking and choosing games that one feels are "unfair" while ignoring others shows a distinct lack of understanding of probability.

@cboers18: Please, PLEASE stop posting random games as though they prove something. They do not. You are not engaging in rigorous scientific study methods, which makes me severely doubt your claims of expertise and authority. You are engaging in apophenia.

If you want to prove that the randomizer is broken in some way, play ten thousand games, diligently record EVERY roll across those games, and submit that.

2024 18 Jun 07:07 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm working on it, @Nekojin. I'm a published math professor, turned stay-at-home mom, very busy with the kiddos, so getting to ten thousand takes some time 😂 In the meantime, the developers said they'd keep the thread open for people to post their weird results, so I'll keep on.

But, like I've said before, if I can predict within the first three rolls, how the end game stats will differ from a normal distribution, then something has to be up with the randomizer. Designing an experiment to prove it is not straightforward.

2024 19 Jun 09:33 • Romain672 • הצעה זו לא נותחה על ידי היזמים עדיין:

@cboers: if you want to have more weight about people who knows statistic, start by writing an event which is easily checkable if it's true or false.

Like 'The starting 6 numbers of the player with the highest elo will be rolled in average less than usual compared to all others numbers'.
Then in all of your next 100 games, note like the 10 games where something weird happen about this event. Then share it on the forum, here, or in mp.

That will force you to choose a specific event, force the number of game to be definite, which will remove most of the cherry picking.

2024 27 Jun 04:41 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@Romain672 great advice, thank you. I've been working on an easier one to start: the player with the highest elo rolls the least number of 7's in a game, more often than would be expected. I'm working with a set of 1000 games. If you know of any way I can get the end-game stats of games which I didn't play, it would be really helpful.

2024 6 Jul 17:05 • Klosterfrau Melissa • הצעה זו לא נותחה על ידי היזמים עדיין:

In the last game, we had 34 times the number 9 and 22 times the number 6. It was only 18 minutes long and was no fun at all.

Severeal numbers didn't come at all. Not only the 2 and the 12, but also numbers Like the 4 and the 10.

2024 6 Jul 22:33 • greenjacket124 • הצעה זו לא נותחה על ידי היזמים עדיין:

It's amazing to me that people actually think the randomizer is rigged somehow. It uses the same cryptographically secure algorithm as every other game in BGA. There literally is nothing wrong with it. It is just randomness.

2024 7 Jul 00:11 • hermilion • הצעה זו לא נותחה על ידי היזמים עדיין:

534165794 Stilkl not solved. WILL NOT BUY PREMIUM UNTILL IT IS SOLVED

2024 7 Jul 00:12 • hermilion • הצעה זו לא נותחה על ידי היזמים עדיין:

@greenjacket124 No, when you have a series of 20 times 6 and 5 out of all results tyhere is nothing random about it.

2024 7 Jul 01:29 • greenjacket124 • הצעה זו לא נותחה על ידי היזמים עדיין:

@hermillion: No, that just means you don't actually understand randomness and the law of averages. A single game or even a few games is nowhere near a large enough sample size to attempt to show an issue.

2024 31 Jul 10:27 • leosiu430 • הצעה זו לא נותחה על ידי היזמים עדיין:

Un real dice statistics
擲骰：2 的比率 (~3%)
0
擲骰：3 的比率 (~6%)
6
擲骰：4 的比率 (~8%)
11
擲骰：5 的比率 (~11%)
16
擲骰：6 的比率 (~14%)
11
擲骰：7 的比率 (~17%)
23
擲骰：8 的比率 (~14%)
10
擲骰：9 的比率 (~11%)
11
擲骰：10 的比率 (~8%)
3
擲骰：11 的比率 (~6%)
3
擲骰：12 的比率 (~3%)
5

2024 7 Aug 21:20 • Kokita86 • הצעה זו לא נותחה על ידי היזמים עדיין:

I agree, that the dices don't roll random.
There must be something wrong.
It also seems this problem apears since 2022 and no one cares.
That's a pity, because this online version is awsome.

2024 3 Sep 04:17 • djabwana • הצעה זו לא נותחה על ידי היזמים עדיין:

@greenjacket124 it's quite possible for a pseudo-random generator to have bias.

I lost significant skill points due to 2 recent crazy games and am tempted to not play the game any more.
boardgamearena.com/table?table=557463015 15% 12 rolls
boardgamearena.com/table?table=556660512 44% of rolls were 4 or 5

I see the solution as offering the Traders and Barbarians Event Deck (maybe not with the events) so that dice are *shuffled* with a reset instead of *randomly generated*. This would be HUGE for Catan.

2024 16 Sep 23:08 • Coisinho • הצעה זו לא נותחה על ידי היזמים עדיין:

Please watch the game that originated this post. There is no way the numbers favour a single player like that out of sheer randomness. 9 has not come out once after 50 rolls in. The game is all about, 4, 11, 6 and 10 (miraclously after said player made a city there). That player was able to get a 6 point lead in catan expansion. Hilariously, 9 came out after that player performed the number switch to rob me of my 9 for his 4. At some point, it starts defying the odds the other way around and it becomes "impossible" that an event does not happen. I am not making accusations but it started to become to statistically impossible to be true.

2024 19 Sep 16:26 • Pogretskiy • הצעה זו לא נותחה על ידי היזמים עדיין:

Стол #564646870
Ход #209

Кубики: частота 2 (~3%) 2
Кубики: частота 3 (~6%) 5
Кубики: частота 4 (~8%) 7
Кубики: частота 5 (~11%) 8
Кубики: частота 6 (~14%) 13
Кубики: частота 7 (~17%) 8
Кубики: частота 8 (~14%) 11
Кубики: частота 9 (~11%) 5
Кубики: частота 10 (~8%) 18
Кубики: частота 11 (~6%) 10

Кубики: частота 12 (~3%) 13. !!!!!!!!!!!!!!!!!!!!!!

It’s just not the right thing to do. These are systematic mistakes and they’re not logical. Mathematical statistics from a world of pairs. Developers, do something. Respect your users. Get rid of the algorithms. Hire the right people if you can’t do it yourself.

2024 23 Sep 20:43 • mariana2197 • הצעה זו לא נותחה על ידי היזמים עדיין:

In game #566514440 which lasted 40 minutes, the number 5 came up only once, and the number 3 came up many times, which benefited more people. The randomness of the dice is not respected!

2024 2 Oct 01:25 • akarachel • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/5/catan?table=570043430

This game had 4, 6 and 7 over and over again. Makes the game seem unbalanced. I know this can happen, but there does seem to be a pattern of dice sticking on certain numbers. Other games recently had 4 or 10 once or twice each four turns.

2024 16 Oct 12:21 • djimm100 • הצעה זו לא נותחה על ידי היזמים עדיין:

I just played 2 real time games of Catan. Sorry, can't figure out how to get the table numbers.

In the first, I got virtually no resources for the last 2/3rds of the game - maybe 2 - while my opponents racked up score after score. In the second, same thing. Even worse, I had a wheat field that was supposed to produce on 8's. I think only one 8 was rolled for the game! but lots of consecutive 7's for both games.

Something has got to be wrong for your randomizer ... at least when it comes to real time games.

2024 20 Oct 23:56 • tac3706 • הצעה זו לא נותחה על ידי היזמים עדיין:

4
Dice roll: percentage of 3 (~6%)
4
Dice roll: percentage of 4 (~8%)
4
Dice roll: percentage of 5 (~11%)
19
Dice roll: percentage of 6 (~14%)
15
Dice roll: percentage of 7 (~17%)
14
Dice roll: percentage of 8 (~14%)
13
Dice roll: percentage of 9 (~11%)
10
Dice roll: percentage of 10 (~8%)
9
Dice roll: percentage of 11 (~6%)
8
Dice roll: percentage of 12 (~3%)
1

2024 23 Oct 23:02 • gecko8989 • הצעה זו לא נותחה על ידי היזמים עדיין:

Hey devs
Can you please look into it?
There is no way the dice randomization is correct.
I just got a game with 8+ times number 12......
There is a cluster on a few numbers.

I play a this game daily and I got to see 5 times in a row 11 and other strange things.

The feeling is shared with lots of players.
Some have stopped this game due to that....

2024 4 Nov 12:07 • SvenOver • הצעה זו לא נותחה על ידי היזמים עדיין:

Catan #584453091
6 consecutive 7's, starting from move 46: boardgamearena.com/gamereview?table=584453091
Overall, 7 was thrown 28% of the times, instead of 16,67%
Doesn't look like the rolls are independent of each other.

2024 11 Nov 05:46 • 98225Bromi • הצעה זו לא נותחה על ידי היזמים עדיין:

Just had a game with 8 rolled 36 times. 6 rolled 0 times.

2024 17 Nov 06:57 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

OMG @98225Bromi, the game you described with 36 8s and no 6s is super interesting to analyze. Could you go into your game history and post the game number here?

2024 18 Nov 07:53 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

@98225Bromi, I found the game and watched it and it was truly interesting. Again confirmed one pattern I've noticed that when a player doubles up on a settlement (in your case 8), the randomizer is especially imbalanced. I also noticed the extreme amount of 3s rolled in your game, also benefitting the same player.

boardgamearena.com/13/catan?table=590449529 Here's another game I played where one player (who started the game with the lowest probability of getting a card on a single roll) doubled up on two numbers (8 and 10) and ended up winning the game in a landslide.

2024 18 Nov 17:36 • Gamerfleet • הצעה זו לא נותחה על ידי היזמים עדיין:

The dice roll is definitely clustering. Is there another website where we can play catan that has a balanced dice roll?

2024 3 Dec 23:00 • Coisinho • הצעה זו לא נותחה על ידי היזמים עדיין:

How is it possible that after #205 plays, game at 55% completion, not a single 8 has come out. No point in trying to play for ELO when you get screwed up like this. Check your implementation of the game.

2024 23 Dec 21:05 • superjulio64 • הצעה זו לא נותחה על ידי היזמים עדיין:

Table# 605524260 / boardgamearena.com/table?table=605524260

Played through an entire game without a single 7 rolled.

Game duration
27 min
Players' average level
80
Average score
5.5
Remaining development cards
19
Dice roll: percentage of 2 (~3%)
3
Dice roll: percentage of 3 (~6%)
11
Dice roll: percentage of 4 (~8%)
8
Dice roll: percentage of 5 (~11%)
16
Dice roll: percentage of 6 (~14%)
27
Dice roll: percentage of 7 (~17%)
0
Dice roll: percentage of 8 (~14%)
11
Dice roll: percentage of 9 (~11%)
11
Dice roll: percentage of 10 (~8%)
11
Dice roll: percentage of 11 (~6%)
3
Dice roll: percentage of 12 (~3%)
0

2024 30 Dec 19:23 • Jimbobbery • הצעה זו לא נותחה על ידי היזמים עדיין:

@Nekojin - I need to disagree with you here. I understand the law of averages. I also understand software development and functions used to generate sudo random numbers.

I would not expect people playing a game for enjoyment to, as you suggest, "play ten thousand games, diligently record EVERY roll across those games, and submit that" in order to contribute to a bug report. Nor would I describe other posters contributions as "worthless nonsense."

Please remember, we're talking about the minutiae of engineering code here. Not the probability of dice rolls.

There is no contribution here which won't help a software engineer look at and understand the issue.

My best guess, based on the behaviour reported is that to aid with performance the dice rolls are being determined at load time rather than run time from a random number generator which is seeded from the current time in milliseconds. The issue with this is that, depending on the OS the code is running on, the system clock won't be updating every millisecond. This will lead produce the clustering that's being reported.

2025 4 Jan 03:50 • MountainHeart • הצעה זו לא נותחה על ידי היזמים עדיין:

Please look at this. In game boardgamearena.com/13/catan?table=610336178, the dice repeatedly rolled 9 and 10, for example. I have seen other games where the numbers that are supposed to show up more frequently, such as 8, just don't. It makes it challenging to strategically place settlements and cities in the opening round.

2025 4 Feb 11:28 • Julyvee • הצעה זו לא נותחה על ידי היזמים עדיין:

I did an Anderson-Darling test on multiple Catan games because I was experiencing the same clustered rolls, and over multiple games the test rejects the hypothesis that the dice sums follow normal distribution (which they should). This suggests that the dice rolls are indeed not random and evenly distributed. So surely there is something going on with the coding that makes the results not truly random.

2025 9 Feb 15:08 • mokiran • הצעה זו לא נותחה על ידי היזמים עדיין:

サイコロの目が明らかにランダムではなく、複数の数字を示している時がある。

2025 11 Feb 00:04 • mokiran • הצעה זו לא נותחה על ידי היזמים עדיין:

出目がかたよりすぎています。人が抜けた途端に都合のいい目ばかり出るようになるのはなぜですか？

2025 11 Feb 10:30 • mokiran • הצעה זו לא נותחה על ידי היזמים עדיין:

Obviously there are too many 3's.

2025 15 Feb 15:26 • ts_hey • הצעה זו לא נותחה על ידי היזמים עדיין:

When I accumulate more victory points than others, the die rolls are too much in favor of others. I had accumulated 8 points and others had about 4 points, but from then on, only the die rolls were against me. And as a result, I lost. Of course my play is not perfect, but these dice discourage me from playing.

2025 15 Feb 17:04 • Roskill1 • הצעה זו לא נותחה על ידי היזמים עדיין:

The dice generator is very annoying. the team that chose this generator was made up of mathematicians not gamers. after playing more than 50 games you can't resist when you notice that the crap generator gives higher chances to 3 and 11 dice than 6 7 or 8. which makes a player who placed well often lose because of your dice generator idiocy.
what the hell, you see that you have 1 million complaints about this and you don't do anything. I PROPOSE THAT NO ONE BUYS PREMIUM AGAIN UNTIL THE DICE GENERATOR IS REPLACED. IF YOU PLAY US - Then we have to cut your invested money. or to move to another online gaming platform

2025 16 Feb 23:37 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

Please take a look at this game
boardgamearena.com/table?table=631239836

2025 18 Feb 11:51 • Franco Frenulo • הצעה זו לא נותחה על ידי היזמים עדיין:

Question: could be possible that dice is scripted in favor of premium players?

2025 21 Feb 19:15 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/12/catan?table=633586482

20% of rolls were 2, 11 or 12 (should be 11% of rolls). And this was even after a bunch of 5s, 4s and 10s finally rolled at the end of the game. Interestingly, one player settled on every number except 2,3,11,12.

2025 2 Mar 16:14 • mirandacv • הצעה זו לא נותחה על ידי היזמים עדיין:

I don't know if it's a coincidence or not, but it seems that the dice favor premium players, no matter what number they choose, even 2 rolls more often when they choose it.

2025 10 Mar 17:20 • StinkyWizzleTeats • הצעה זו לא נותחה על ידי היזמים עדיין:

I have a spreadsheet of about 500 games.... they are random

2025 14 Mar 22:39 • Jimbobbery • הצעה זו לא נותחה על ידי היזמים עדיין:

@StinkyWizzleTeats

would you mind expanding on your statement please?

When you say 'random,' do you mean:
- The set of dice rolls for a game is not predictable prior to any visibility of the set;
- OR The set of of dice rolls for a game is not predictable even after a small subset of the set has been observed?

2025 19 Mar 20:47 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

I have a theory that when one player settles two spots with the same three numbers at the beginning of the game, that the roll distribution is especially skewed. Here I settled 8, 4, 10 for both my initial settlements, one of which should roll about 1 in 3 rolls. Instead it was less than half of that with only 2% eights by the game end. Not fun for anyone, even the winner.

boardgamearena.com/11/catan?table=646391502

2025 24 Mar 16:31 • -Peachy- • הצעה זו לא נותחה על ידי היזמים עדיין:

Hi, here also a weird game. 10 and 11 got rolled 3 to 4 times in a row. Which seemed really fishy to me. Defintly the weirdest game I had for now.
boardgamearena.com/5/catan?table=648719540

2025 26 Mar 03:24 • fearofthequark • הצעה זו לא נותחה על ידי היזמים עדיין:

At our latest game it was six times 8 right after another. Then three times 10 and three times eleven... I as the host of the game was on field 5, 6, 3, 9 which didn't come for 12 dice throws in a row.... Like that you don't have any chance on an fair play... of corse the person throwing 5- times 8 won... Sometimes I think the host is sabotaged by the game...

2025 31 Mar 01:35 • cjcheck • הצעה זו לא נותחה על ידי היזמים עדיין:

Dice are buggy
in game
651846181
the number 9 was rolled 31 times....

2025 23 Apr 23:46 • cboers18 • הצעה זו לא נותחה על ידי היזמים עדיין:

boardgamearena.com/13/catan?table=663231825

Less than 1% chance if no 6 in the first 31 rolls, favoring the lowest elo player (half of the first 16 rolls were 8). Also, many more 7s rolled by the player with the lowest elo.

2025 29 Apr 22:01 • beArandallboardgamer • הצעה זו לא נותחה על ידי היזמים עדיין:

The dice algorithm is completely unfair. This game was 50% 6's Ridiculous. STOP trying to make games happen...just use a straight dice roll randomizer !! game is not fun

2025 1 May 03:06 • brunajmeller • הצעה זו לא נותחה על ידי היזמים עדיין:

Look at this game: boardgamearena.com/8/catan?table=666623439

38 TIMES THE NUMBER 8!!!!! only 3 times the number 6. This is insane!

2025 20 Jul 05:14 • NathanMTZ777 • הצעה זו לא נותחה על ידי היזמים עדיין:

Overall dice is giving numbers that are not random and this should be fixed.

2025 23 Jul 08:14 • Saith of Dusk • הצעה זו לא נותחה על ידי היזמים עדיין:

imgur.com/a/hFK0ggT

No 10's the whole game...

2025 27 Jul 20:51 • _Bighouse_ • הצעה זו לא נותחה על ידי היזמים עדיין:

I have a lot of games where 8 is thrown all the time, while the amount of 6 is almost zero.

This game seems really bugged dice-wise.

2025 12 Aug 16:55 • bensenfe • הצעה זו לא נותחה על ידי היזמים עדיין:

Tiene pinta que prevalecen los números que tienen los jugadores permium. Sólo fijaros quien es premium en la partida y la mayoría de veces es obsceno la cantidad de veces que les salen números de sus terrenos. da igual que sea un 6 o un 12, he visto partidas que sale más un 2 que un 6.

2025 21 Aug 16:26 • Hungry_Black_Hole • הצעה זו לא נותחה על ידי היזמים עדיין:

I'm not paying much attention to clustering, but I ran a statistical analysis on the dice rolls from my last 15 games. Most of the time (about two-thirds of the games), the dice behaved as expected and seemed fair. But in a few games, the results were really strange, and some outcomes were extremely unlikely.

I made a table showing the results for each game (table ID, counts of each sum from 2 to 12, χ² statistic, and p-value), and I included my Python calculations for anyone curious:
colab.research.google.com/drive/1Cmyp6RRjgLr0XbXgf9lrj_37zQXoc1MZ?usp=sharing

In short: I compared what actually happened in each game to what should happen with fair dice. The p-value tells you the probability of seeing results as far from the expected ones, assuming the dice are fair. If the p-value is above 0.05, the dice are likely fair; below that, it’s statistically unusual.

In my 15 games, 6 had p-values low enough to suggest the dice weren’t fair. The most extreme case was table #715991625 — I’ve never seen dice so biased before!

Honestly, it’s frustrating when the dice are this unfair — something should really be done about it.

2025 26 Aug 00:59 • ribidilus • הצעה זו לא נותחה על ידי היזמים עדיין:

premium players can see the stats easily... have to come back to this thread, when i'm premium to put the tables values into that python script. The problem is, that people only come here, when something is wrong, so we get biased data... well. good position for bga to lean back and pretend all is fine and nothing is rigged

2025 26 Aug 03:47 • ribidilus • הצעה זו לא נותחה על ידי היזמים עדיין:

I've put all numbers in the script from @Hungry_Black_Hole starting with a 6 (table_id). The most extra ordinary is Table 666623439 from 1. Mai 2025 3:06 • brunajmeller

2025 18 Sep 11:16 • obviousowl • הצעה זו לא נותחה על ידי היזמים עדיין:

Look at the logs on the 9 rolls. This is unnatural. We are bound to have some 7 rolls mixed in.

2025 23 Sep 22:47 • obviousowl • הצעה זו לא נותחה על ידי היזמים עדיין:

unnatural dice rolls

2025 26 Sep 11:21 • obviousowl • הצעה זו לא נותחה על ידי היזמים עדיין:

unnatural dice rolls

2025 3 Oct 02:18 • obviousowl • הצעה זו לא נותחה על ידי היזמים עדיין:

unnatural dice rolls (9 rolls)

2025 3 Oct 21:42 • PashaPupsik • הצעה זו לא נותחה על ידי היזמים עדיין:

It’s not about only CATAN. In many games, the dice, that has small chances to be seen, comes twice in a row or never comes. Probably the randomiser is working wrong

2025 10 Oct 07:44 • PashaPupsik • הצעה זו לא נותחה על ידי היזמים עדיין:

In my game, I just had something like 10 7s in a row! It’s table number 741669154

2025 30 Oct 22:55 • cvramen • הצעה זו לא נותחה על ידי היזמים עדיין:

What method is used to generate random numbers? The native random functions of PHP or JS? Some external data with high levels of variation like atmospheric noise? I know some people are irrational, and will claim unfairness no matter what you say, but I feel like that could quell some folks' concerns if you tell them how the numbers came to be.

2025 5 Nov 23:33 • 19Anubis83 • הצעה זו לא נותחה על ידי היזמים עדיין:

pieldauer
25 min
Durchschnittliches Spielerniveau
142
Durchschnittliche Punktezahl
6.66667
Verbliebene Entwicklungskarten
13
Würfelwurf: % der 2 (~ 3 %)
3
Würfelwurf: % der 3 (~ 6 %)
10
Würfelwurf: % der 4 (~ 8 %)
5
Würfelwurf: % der 5 (~ 11 %)
14
Würfelwurf: % der 6 (~ 14 %)
12
Würfelwurf: % der 7 (~ 17 %)
27
Würfelwurf: % der 8 (~14%)
5
Würfelwurf: % der 9 (~ 11 %)
10
Würfelwurf: % der 10 (~ 8%)
7
Würfelwurf: % der 11 (~ 6 %)
3
Würfelwurf: % der 12 (~ 3 %)
3
l

das kann nicht sein und macht keinen spass, weiss nicht was geaendert wurde aber sowas passiert am laufenden band !!! Das die 11 oder 3 oefter kommen als 6 oder 8 und das extrem oder es kommt 5 mal hintereinader dieselbe zahl ...das passiert in wirklichkeit auch nicht vielleicht einmal bei 10 partien... bin von einem elo 290 zurueckgefallen weil da ws geandert wurde ...bitte tut was

2025 9 Nov 01:34 • Taraxacum • הצעה זו לא נותחה על ידי היזמים עדיין:

That's so true.

The dice in this game are so unnatural, in about 70% of the games (I played) there are specific numbers (different in every game ) that roll extremely often, others nearly never.
Change the dice programming - this game is no fun that way.

Just ask the Catan- people on "catanuniverse.de" how thes do it, the dice there are natural.

These dice are ridiculous, really!

2025 9 Nov 01:38 • Taraxacum • הצעה זו לא נותחה על ידי היזמים עדיין:

Just take the dice programming from Space Base

The Space Base dice are really random and natural, they roll like RL.

Catan dice do not, no way !

2025 11 Nov 10:06 • tangerinka • הצעה זו לא נותחה על ידי היזמים עדיין:

I just played a game and 6 was NOT ROLLED for 50% of the game... table #759014466. Guess around which number I had my first towns built? :)

2025 13 Dec 14:21 • daseinwm • הצעה זו לא נותחה על ידי היזמים עדיין:

I see 3 years of reports about completely improbable (read: mathematically impossible) dice rolls (errant randomization) in this game/on this site, and no apparent interest/action from the developers. ('This suggestion has not been analyzed by the developers yet:').
Is it possible to get this topic addressed? I see that I am not alone in noting the issue.

#66082: "Are dice results really random ?"

על מה דוח זה?

מה קרה? אנא בחר מהרשימה מטה

מה קרה? אנא בחר מהרשימה מטה

בבקשה בדוק אם יש כבר דוח על אותו נושא

תיאור מפורט

היסטוריית דיווחים

הוסף לדוח הזה

שנה את מצב הדיווח ל:

עזרה

תרום

נווט

עקוב אחרינו