Dice is not random at all

[Mart]

I have some serious doubt about the dice. I just played a few games in which it was nothing but 5's and 9's. Hardly any 7 or 6 or 8 was rolled. When I look at statistics after the game, a normal distribution cannot be seen in the times a certain number is rolled.
This game is far from perfect. It laggs, it crashes occasionally, but I at least expect that i can play the game in a fair way.

[Ink]

~60 dice rolls isn't going to give a 'normal distribution'. Who knows if the dice rolls are adequately RNG but this argument isn't the one to use.

[Way Smarter]

I respectfully disagree with you. Mathamatically the dice have to do what you experienced every now and then. instead of 60 rolls. try 10,000 rolls to compare how the numbers are distributed.

[fearenough]

lol i had game last night first 5 throws 2 then 2s 4s 9s and 6s but mostly 2s 4s n 9s o with random other numbers in between but every 3-5 rolls 2 the rng been tampered with by dev seemed to have changed on update just the other day. before that other than getting dropped from game rng seemed better was having close games that was fun. i dont like winnning 1 sided or losing

[Smith]

https://steamcommunity.com/app/544730/discussions/0/2247805152849208123/

[jehman]

Originally posted by Way Smarter:

I respectfully disagree with you. Mathamatically the dice have to do what you experienced every now and then. instead of 60 rolls. try 10,000 rolls to compare how the numbers are distributed.

You are correct that given thousands of rolls over all games may be normally distributed. However, I have played probably 100 games and have have found that in any given game two or three numbers other than 7 are generally favored for that whole game. The consistency in which this happens in specific games would indicate something other than "random" for each game.

[Johnny_Retro]

They most probably use the mersenne-twister (or some variation of it) as an algorithm.
One of it's weaknesses is that i can produce badly distributed series of numbers.

[fearenough]

Originally posted by Smith:

https://steamcommunity.com/app/544730/discussions/0/2247805152849208123/

how the hell did i play a game that 2 came up over n over again more times than even 7

[Smith]

Originally posted by fearenough:

Originally posted by Smith:

https://steamcommunity.com/app/544730/discussions/0/2247805152849208123/

how the hell did i play a game that 2 came up over n over again more times than even 7

Read what I wrote.

[Chrisse]

I am pretty sure that the rolls in single player are not random but especially designed that you get a disadvantage in the beginning. I noticed in nearly all of my games, that the numbers which i don´t have in the beginning, show up much more like they should. Pretty common build. I don´t have a 9 - in the first 20 rolls the 9 is at 30%. The next game i don´t have a 5 in the beginning, 5 has 30%. I changed my strategy now that I don´t look on the numbers, but I look on what numbers the PC takes and this is working pretty good now - but i don´t like this - it makes the game - well.....

[noob master]

I am a programmer and have dabbed in random algorithms. I strongly believe the random algo. they use for the dice rolls is either bugged on purpose, or simply poorly written, leading to really bad gaming experiences which wouldn't really happen in real life. Here are a few situations that I found myself in:

1. In one game, I positioned myself well with 2 settlements on a 5. In 42 dice rolls, 5 did not roll even once. The probability of 5 not showing on any given roll is 88.89% (100% - 11.11%). The probability of 5 not showing at all in 10 consecutive rolls is around 30% (88% ^ 10). The probability of 5 not showing at all in 42 rolls is: 0.7% (88% ^42). This means 0.7% chance of this happening. By this time the distribution started to flatten out and after 42 rolls 5 came a few times. Safe to say my game was over around the 30th dice roll that didn't show a 5. Also in a 4 player game, 42 dice rolls happen after all players roll at least 10 times, just so you understand how ridiculous this experience was.

2. Another time I was playing against a player with 2 settlements on a 3. I had him beat positionally and resource-wise. In 25 rolls, his 3 rolled 7 or 8 times, at the beginning of the game. This comes to between 0.0000...% and 0.03% you can work out the math on this using the Bernoulli trial ;) You have better odds at a lottery, yet here this guy was in 4-5 turns winning the game at 10 points, while the rest didn't break past 4 points.

3. Even worse, many times (more than 25%) when I'm ahead, as soon as I get my 8th card, a 7 rolls as if on schedule. Other times if I'm ahead, 3s and 11s or 12s start to show even more to get my opponents even more resources (maybe to balance the game?). This type of dice roll that seems to purposefully slow my game down never seemed to happen to me on Colonist.io. I don't really want to suggest deterministic dice, but the signs are there for at least a very poor random algo. Just saying ;)

All in all, if Catan Universe wants to stay in business they need to prevent this kind of stuff from happening and normalize probabilities within a game, not within 1000s of games with one game getting only 9s and another game getting only 5s.

Otherwise, time to switch to Colonist.io folks, as have I (until Catan Universe makes some actual changes)

[noob master]

And I agree with everything said in previous posts:
1. sometimes in some games the algo gets stuck on a certain number like 9 or 5 (or in my case 3)
2. other times it seems dice come with the deterministic scope of affecting the game. I saw this in single player as well, but also multiplayer
3. the distribution of random dice rolls is actually garbage

[fearenough]

Originally posted by Smith:

Originally posted by fearenough:

how the hell did i play a game that 2 came up over n over again more times than even 7

Read what I wrote.

that not tell me why 2 got rolled more than 11 times

[N o i r]

Originally posted by sparky:

I am a programmer and have dabbed in random algorithms. I strongly believe the random algo. they use for the dice rolls is either bugged on purpose, or simply poorly written, leading to really bad gaming experiences which wouldn't really happen in real life. Here are a few situations that I found myself in:

1. In one game, I positioned myself well with 2 settlements on a 5. In 42 dice rolls, 5 did not roll even once. The probability of 5 not showing on any given roll is 88.89% (100% - 11.11%). The probability of 5 not showing at all in 10 consecutive rolls is around 30% (88% ^ 10). The probability of 5 not showing at all in 42 rolls is: 0.7% (88% ^42). This means 0.7% chance of this happening. By this time the distribution started to flatten out and after 42 rolls 5 came a few times. Safe to say my game was over around the 30th dice roll that didn't show a 5. Also in a 4 player game, 42 dice rolls happen after all players roll at least 10 times, just so you understand how ridiculous this experience was.

2. Another time I was playing against a player with 2 settlements on a 3. I had him beat positionally and resource-wise. In 25 rolls, his 3 rolled 7 or 8 times, at the beginning of the game. This comes to between 0.0000...% and 0.03% you can work out the math on this using the Bernoulli trial ;) You have better odds at a lottery, yet here this guy was in 4-5 turns winning the game at 10 points, while the rest didn't break past 4 points.

3. Even worse, many times (more than 25%) when I'm ahead, as soon as I get my 8th card, a 7 rolls as if on schedule. Other times if I'm ahead, 3s and 11s or 12s start to show even more to get my opponents even more resources (maybe to balance the game?). This type of dice roll that seems to purposefully slow my game down never seemed to happen to me on Colonist.io. I don't really want to suggest deterministic dice, but the signs are there for at least a very poor random algo. Just saying ;)

All in all, if Catan Universe wants to stay in business they need to prevent this kind of stuff from happening and normalize probabilities within a game, not within 1000s of games with one game getting only 9s and another game getting only 5s.

Otherwise, time to switch to Colonist.io folks, as have I (until Catan Universe makes some actual changes)

1) and 2) Considering the number of games played every day, this was bound to happen for someone sooner or later. Also, the probability of observing a random unspecified unusual occurrence without specifying neither the occurrence nor the timeframe of the observation is infinite.

3) Confirmation bias.

You are a developer and don't understand basic statistics? Why would they weigh the RNG, when a perfectly functional one is readily available in pretty much every single programming language?

[N o i r]

Originally posted by Chrisse:

I am pretty sure that the rolls in single player are not random but especially designed that you get a disadvantage in the beginning. I noticed in nearly all of my games, that the numbers which i don´t have in the beginning, show up much more like they should. Pretty common build. I don´t have a 9 - in the first 20 rolls the 9 is at 30%. The next game i don´t have a 5 in the beginning, 5 has 30%. I changed my strategy now that I don´t look on the numbers, but I look on what numbers the PC takes and this is working pretty good now - but i don´t like this - it makes the game - well.....

More empty circumstantial evidence and confirmation bias. No, there is no demonstrable connection between the positions picked by the AI players and the probability of outcome of the rolls - it's all in your head.

[noob master]

(before you read, also check this out: https://imgur.com/gallery/ySnG9RF, there are several ones out there showing very poor distribution over 50,000 rolls, and in all my programs after about 1 to 2 hundred rolls the distribution starts to be pretty fair)

1) when this happened I had played no more than 50 games. I agree, this is mathematically very possible without a flaw in the random algo but makes for a crappy experience for me.

2) the second one is a lot less likely. Just like another game when I got 6 twos in 10 rolls. I did a simulation (see below)

3. Very much possible. Still, this doesn't happen nearly as frequently on other platforms that I've tried, like http://colonist.io. I've held the lead with 8, 10, 14 cards in my hand many times on colonist.io without seven-ing out, unlike in Catan Universe

I made a program for case 1) just to demonstrate. The way it works is:
- in a single trial, the program rolls two dice 42 times
- checks for how often value k was rolled in those n rolls
- the program runs a total of 1000 trials

For 1000 trials (of 42 rolls each) I got 5 not showing in exactly 7 of them, but sometimes 10, sometimes 5. Which matches the math. For 50 games like I had it matches 1-2 times when 5 doesn't show in 42 rolls. Again totally fine, though crappy experience.

Then let's move on to case 2. I created another program with 3 params (N, K, X) which:
- rolls dice N times
- checks if number X was rolled more than K times

Then I ran this program over 1 million times. Not once did the 3 come up like it did in the Catan Universe (as I remember, it was 10 times in the actual game). On the other hand, if I put in 7, or 6, or 8, it was quite possible for these to come up 10 times in 25 rolls (still low chance, but happened, i.e. 0.2-0.5%). I also initially said 7 or 8 times because I didn't save the screenshots and didn't want to exaggerate, but after 25 rolls, 3 rolled 3 more times anyway at which point I quit the game. But even for 3 to come up only 8 times in 25 rolls, my program got literally a 5 in a million chance.

I then wrote another program about the 2 coming up 6 times in 10 rolls, guess what? In 1 million runs of this program, not one time did that happen.

What did I do to deserve to get multiple less than one-in-a-million losing situations, in several games played consecutively? It's a garbage experience, pure and simple. I have solutions on how to fix it of course, but nobody seems to be listening on their end.

It's a nice math exercise to prove their algo is pretty bad, but other than that, no thanks. Even if the math checks out in the grand scheme of things (which I highly doubt given their frequent one-in-a-million scenarios) it is a horrendous experience for the user.