Dice Rolls

[ProphetSword]

My friend has been tracking the dice rolls and has noticed that he is more likely to roll 1-3 nearly 12x as often as he rolls 4-6. He tracked these number across several games and came to the conclusion that players have the dice weighted against them.

He noticed in particular at the Crown of Command that the player is less likely to roll 4-6. For three games, he tracked the rolls and got these results:

Game 1:
Rolls Below 4: 12
Rolls Above 3: 4

Game 2:
Rolls Below 4: 7
Rolls Above 3: 4

Game 3:
Rolls Below 4: 12
Rolls Above 3: 6

After that he tracked the rolls in a complete game. In 30 turns, only 8 turns were above 3, while the other 22 turns were all below 4. In another game, the results were similar. I did the same, and got results very close to this as well.

AI players, on the other hand, don't seem to have this issue. The conclusion seems to be that the AI players have an advantaged weighted to help them against human players.

Even with purely random results, these outcomes shouldn't happen. It could be chalked up to the computer not generating random numbers well enough (as a programmer, I understand this completely), so steps need to be taken to make this outcome a little more balanced.

I invite other players to track their rolls across a game or two and compare results. I want to see if this is true across all players or if it is just affecting some of us. Because if it is widespread, then it needs to be addressed.

[Bing987]

Hmm. I haven't tracked my die rolls, but it does seem to take an inordinate number of rolls to kill off the other players when I get to the CoC. One game in particular I recall that I got to the top on the 35th turn and I finally killed off the last player at turn 80.

Granted, there was a Healer in the game, but still. If he hits the Healer every 6th turn, on average, I should be doing three damage to him every time he heals two lives.

I'll keep track next time.

[Bing987]

Okay. I hit the CoC and my opponents had 2, 3, and 4 lives. It took me 17 turns to kill them off, with 10 rolls under four and 7 rolls over three. That's not statistically unusual, but it is suspicious.

Had the Elf done anything other than teleporting between Woods spaces, making Rafts, it would have taken a lot longer (he did cast Healing twice). The Ghoul did manage to land on the Healer twice.

[ProphetSword]

Thanks for the feedback. If you feel like keeping track again, let me know what you end up with. I think this is shaping up to be something that should be addressed.

If anyone else wants to track theirs, feel free to throw it in here. It's odd that the number of rolls below 4 is always higher.

[Drunken Farmer]

I could of told the same thing in a post when it takes me 35 minutes and roll 5 1's in a row in a single game... Something isnt Kosher

[ProphetSword]

My friend tracked his rolls throughout the game again just now, including movement and combat until he died (which happened early on) The results were this:

Rolls below 4: 22
Rolls above 3: 9

Again, we can see the weight is not in the middle or slightly off the middle like it should be. It is completely skewed to one side, which is always under 4.

[Drunken Farmer]

I rerolled with 8 fate on Crown of command just now, only 2 actually passed, and thats from 5 failed first attempts.

[ososober]

is there a way to take from the logs all the dice rolls ?

[greengoo]

it's odd.
i find that i get lower then normal roll. but then i quit a game mid play through and come back a few days later my luck seam ot have drasticly impoved.

then after an hour hour or so, the lower dice results seam to be back again.

so if you leave for a day or so the luck changes and i'm lucky again.

even so the CoP still take way to long to finish. its not fifty percent as it shoudl be.

now as i have never tracked my die rolls i cant says if it's what i perceive or fact. hence i can not truly add to this at the moment.

[ProphetSword]

I just finished a game where I tracked my dice rolls at the Crown of Command. While not as bad as my friend's results, they are still slighted toward the low rolls. The results:

Rolls Below 4: 12
Rolls Above 3: 8

[jshepler]

I've had sessions in the CoC like that, and I've had sessions where I'll roll 4+ 6 times in a row. Is it frustrating to roll under 4 over and over and over - definately. Tracking only your own rolls, though, isn't a very good sampling of the distribution of numbers the RNG generates.

Originally posted by ososober:

is there a way to take from the logs all the dice rolls ?

If the log format hasn't changed, the parser code in this old thread might still work. I haven't tried it recently.

http://steamcommunity.com/app/247000/discussions/0/666825525014399869/

[Mr. Neutron]

Their random number generator is absolute crap. In my experience when I re-roll a die there is at least a 75% change of re-rolling the same number, whereas if it was truly random it would have a 1-in-6 chance of happening. I've been submitting bug reports to them trying to get them to fix the problem but it still hasn't happened. Random dice rolls are essential to a dice-based game and if they can't get anywhere close to true random it's an epic fail.

[sanityismyvanity]

From my last game, all dice rolls:
{"1":69,"2":73,"3":75,"4":64,"5":72,"6":73,"rolls":426}

Dice rolls by just my character in the game:
{"1":11,"2":17,"3":14,"4":19,"5":18,"6":20,"rolls":99}

Looks like I rolled slightly better than the 3 AI players, but well within a reasonable margin nonetheless.

Now, can people just stop posting this nonsense. The dice are not rigged. Period. Just let it go.

[ProphetSword]

Here's my last game:
{"1":227,"2":65,"3":60,"4":53,"5":78,"6":58,"rolls":543}

Of 543 rolls, 227 of them were 1 (that's 41%)


The game before that:
{"1":249,"2":189,"3":102,"4":74,"5":110,"6":81,"rolls":805}

Rolls below 4: 540 out of 805 (67%)
Rolls above 3: 265 out of 805 (33%)


And just for fun, the game before that:
{"1":309,"2":58,"3":65,"4":180,"5":63,"6":66,"rolls":747}

Out of 747 rolls, 309 were 1 (look, it's 41% again)

Rolls below 4: 432 out of 747 (59%)
Rolls above 3: 309 out of 747 (41%)

Oddly, the number of rolls above 3 are equal the number of 1s rolled during that game.

So, still think it's nonsense?

[sanityismyvanity]

{"1":22,"2":28,"3":29,"4":32,"5":28,"6":32,"rolls":171}

{"1":25,"2":14,"3":22,"4":22,"5":25,"6":22,"rolls":130}

{"1":38,"2":29,"3":46,"4":35,"5":44,"6":51,"rolls":243}

{"1":104,"2":81,"3":98,"4":104,"5":73,"6":96,"rolls":556}

However, going back to savegames from build 11819, when individual dice roll seeds are shown in the logs and picked up by the parser, you get results like this.

{"1":311,"2":41,"3":64,"4":166,"5":85,"6":79,"7":null,"8":null,"9":null,"rolls":753}

The huge amount of 1's shown is due to the majority of seeds being 10 digit numbers that start with a 1. If you are relying on these logs as your "proof", then the parser is useless. And the game with 556 rolls was 94 turns, the 753 roll game was only 65 turns. The parser wasn't designed with those extra lines of seed information being displayed. Update your game client to the latest build, and yes, do move on.