Start a new topic
Implemented

New Dice Algorithm / True dice rolls

Game does not use true dice roll probabilities. Many, many times I've had 97-99% chance to win and dont.


HOW TO VOTE FOR THIS FEATURE? Tap the 'Do you like this idea?' below


87 people like this idea

All accounts roll fewer ones than they should. I've created a few new accounts and after a game or two they are beginner and before long master. Even after reaching grandmaster the skewed dice remain. Since most of those accounts games are not played at novice level it demonstrates that everyone has skewed rolls and while not accurate the game is balanced.
Nah, that’s nonsense. Any account with a sufficiently large enough sample size comes out with roughly 15-17% statistic for any given roll probability. If everyone is biased away from rolling 1’s then everybody here above Novice should see a below 14% chance of rolling 1. Everybody reading can check and see that’s not the case. I have several friends that play, everyone has 16% even stats after a few dozen games including myself.
Here is my evidence
IMG0684.PNG
(735 KB)
IMG0683.PNG
(864 KB)
IMG0685.PNG
(736 KB)
IMG0681.PNG
(843 KB)
IMG0680.PNG
(739 KB)
IMG0679.PNG
(866 KB)

1 person likes this
Hmmm, that’s messed up. Thanks for posting. Are all these accounts created recently? I wonder if it has something to do with newer accounts. Anybody else reading should post their stats just like SectaOne so we can see how pervasive this is.
The Wilburn Howitzer account is fairly new. I made it immediately after the update that added spectator mode. The other two are a bit older but stll newer than the last dice algorithm change I know of. But I think all accounts roll with the same algorithm as that makes sense to me as a programmer and SMG says it is the case. the reason old accounts stats are not as skewed is because they have a large sample from before the change to the matrix dice.
My stats as a Novice. Note that I mostly play offline games with the AI, as I have basically given up on online games due to the fact of not being credited for wins which would have increased my rank.
I have a fairly new account. Maybe a couple of weeks old. It is Expert / Master level, depending on my current streak. The dice are clearly biased, because I have played 400+ games now and I have 13% of 1s rolled. No way can this happen after such a huge sample size. I just don't understand how it is so difficult to implement smth. this basic and simple ;/

The dice are clearly biased to roll higher than 1.

 


1 person likes this
I made a new account to enjoy destroying newbies.2 games into it and my dice rolls are 13% for the 1's. If it's true that 1's are rolled less often, then the defender is getting some advantage here (assuming it's the same for the defender). If the defending dice are not affected by this, then the defender is getting screwed.

 

I won a 3 vs 10 attack today but I also had some heavy losses so that by itself doesn't prove much. I do think everyone's rolls are skewed not just the attacker or low ranks or high ranks or new accounts or old accounts but everyone's. Having all accounts use the same dice algorithm is the most logical and best approach to take when writing the game. I can't imagine SMG would change rolls from one account to another. I think the stats we are posting are both defensive and offensive rolls together. I'm not happy about the rolls being different from equal odds, but I doubt I would change much about my decision making based on just a slight adjustment.
The question is. How are these stats composed?
Is it from normal rolls only or from Blitz rolls as well, I suspect it is not from all rools as Blitz is probably using some shortened algorithm.

And what Briand is saying is, that if you 1s are rolled less frequently than the defender is getting a higher advantage than he is supposed to, compared to true dice rolls.

There is another very disturbing statement in the Faq, I quote: "The results are within +-5% of the control comparison, which is acceptable."

Uhm no, it's not acceptable at all, when I attack with 8 vs 1 and the chance of winning is 99,9% and instead it's 95%, because the difference is like night and day.


5% deviation is absolutely unacceptable for a trivial algorithm like dice throwing ;/

Blitz can't possibly be giving data for each and every roll. The point of their algorithm is to bypass that. I believe it does a couple of rolls to confirm the results after rolling for the results. Doesn't seem like it would use dice for that though. I have no clue as that code is not shared as far as I know.

+/- 5% doesn't mean your chances of a roll change by 5% it means that expected averages can vary by that much. And that makes sense. Until you have millions of rolls, it's not going to be dead on to the odds of getting something, speaking by the average. Would love to see a number of rolls stat in the dice counter. One of my accounts has over 700 games on it and the dice rolls are all even. It wasn't always. I think I started it not long before when everyone started complaining in this very thread.

I have a 6th account now that I play with when I can't find good enough ranked players. It is rolling less 1's but we'll see how it pans out. Will have to make more accounts to test the theory. Or SMG can just answer honestly?

 

Yes that's what I suspect as well. Blitz probably does not contribute much or anything to the statistics, but only SMG really knows.


The other thing is, I understand that they have some kind of control comparison and that is basically the limit if you did the battles an infinite amount of times. That is to what you refer as expected average.


And now their algorithm gives results, which can differ up to 5% from this control comparison.


That's completely unacceptable in my opinion, it's certainly possible to achieve a difference less than 0.1% to true dice throw outcomes. Devices these days are so powerful that you could even run a MonteCarlo simulation, which would be pure brute force but still fast enough and accurate to these degrees.


I mean I don't want to sound like I'm only here to complain, the app is great and I really enjoy playing the game, but these numbers just seem off and the statement does not really make me more confident in the accuracy of the algorithm -_-

Agree that I want true Dice rolls.The only thing I'll put in is back before SMG took over, when you did a Blitz with 50+ troops, it would hang your device for a split moment trying to calculate all of the rolls. It was a little aggravating especially if it was on a slow device (I now only use fast ones) so I can see why they wanted to go with a probability matrix.If they'd fix the timer to allow you to finish a whole turn without letting people delay the game then I'd be all over going back to a slower but more real system. If I got 5 seconds for every roll I did I could actually win games that I normally would win without these damn time limits.

 

Dude, I only play 5/8 minute games - solves the time limit problem!


When did SMG take over, and from who? Aren't they the original developers of the game?

I already made a feature request to increse the turn timer by a few seconds when you make a Blitz attack. That would totally solve it and is super easy to implement.

 

Login or Signup to post a comment