Are the Dice Rigged?

[LaSpeakeasi]

I keep failing 95%+ checks and it's infuriating. 90% is more like 50%. The predictions are not accurate at all.

[Hannibal_PJV]

I sometimes win in 5% checks... it is how propability works. It is only propability. In some scenarios the propability is not calculated rigth though. Like in scenario where 1-3 Are counted as 0 the propability Number is not show'n rightly!

Edited September 12, 2017 by Hannibal_PJV

[Ethics Gradient]

Are the dice themselves rigged?  No.  We've been over that topic several times since the app first came out a year go.  It has a pretty solid and industry-standard RNG running the show.

 

Is the probability stat rigged?  Well, that's a different question.  That feature is relatively new, and I'm not sure that the community has taken the opportunity to chase it down and verify that it is behaving properly.   You could always go to somewhere like anydice.com and see what it thinks about a roll's chances.  If you input something like:

output 2d6+2d8+3 >= 11

It will show that: 2d6 + 2d8 + a bonus of 3 has about a 97% chance of beating an 11 check.

 

Still, it is possible that the dice stat may be overestimating chances a little.  Without investigating further, it does sometimes seem like the results don't quite follow a standard bell curve.

[Jenceslav]

The dice that you roll for checks seem to be approximately random, only a very large set of observed rolls can prove or disprove it. However, some automatic rolls are very often extremely suspect. Once upon a time, in several of my Legendary plays (in Thistletop), the goblin monster difficulty increase was always rolled 5-8 on 2d4. Similarly, there was a later scenario where every automatic throw of 2d4 (don't know what for, either damage or difficulty increase) was almost always 7 or 8 during my play. Both these things are in Legendary difficulty, so a slightly harder experience is expected.

On the card-drawing barriers (Large chest and such), 1 comes surprisingly often, but that is also OK if it is a way to limit farming.

And lastly, my Seoni very often gets 3 on 1d4 when rolling for Wand of Enervation's 1d4+4 difficulty reduction.

Everything may be just my perception of course, not based in reality, but in my skewed opinion

[King_Momo]

Who knows? There were/are so many bugs in this game, I won't be surprised if the dice were one of them too.

 

And to justify what I just said, the vast majority of my friends whom I have recommended this game to and did not like it complained that they felt like that the dice/rolls felt rigged.

Edited September 12, 2017 by King_Momo

[Bullwinkle]

And to justify what I just said, the vast majority of my friends whom I have recommended this game to and did not like it complained that they felt like that the dice/rolls felt rigged.

This is a very common psychological effect, but is almost never true. Someone gave a great explanation sometime back as to why this game feels that way especially, but it comes down to:

 

1. You make a ton of rolls in this game. This means you will roll many extreme rolls, much more than in the average game. This also means you will get long chains of bad rolls more often than in other games.

 

2. You remember bad rolls more than good rolls. This is amplified in this game because a single bad roll (against a villain, or closing a location) can basically cost you the game.

 

3. Because you only need a minimum roll to beat something in this game, you *really* ignore the super high rolls; they just wash past like you beat the check by just a few points. But you remember every low roll. So the already psychologically amplified effect of point #2 is hugely amplified again.

Edited September 12, 2017 by Bullwinkle

[Ethics Gradient]

Ok, I'm relatively convinced that the stat calculator is working fine.

 

The only catch is that it just models dice, bonuses, and the target roll.  If there are any post-roll powers that modify dice or allow a reroll, it won't take those into account when it calculates a percentage.

 

It probably could handle things like custom dice for the Keen Rapier 2d4{1,2,3,5}, but at least it shouldn't be over-reporting your odds.

[Eumaios]

The formula for calculating chances based on dice sides and numbers of dice is straightforward. I can't imagine that a computer programmer would mix it up. However, I think it does fail to take into account things like "all 1s are turned into 3s" and the scenario where 1s, 2s, (and I think 3s) are turned into 0s. It's not even the RNG. Before they implemented the odds calculation thing, I would just see how many dice could come up as 1 before I lost a roll. This stuff is simple statistics, so you could calculate it out pretty quickly for a small number of dice with a regular ol' tiny calculator.

[Ethics Gradient]

The formula for calculating chances based on dice sides and numbers of dice is straightforward. I can't imagine that a computer programmer would mix it up.

No, it isn't hard to do.  It is just sometimes be tricky to implement in an elegant fashion.

 

For instance, when you you're throwing all the blessings you've got at a villain and wind up rolling something crazy like: 2d12+2d8+3d6+2d4, it's checking just under 32 million combinations for how many beat the check, then updating the results in a few milliseconds.  Sure, you can brute-force an answer using some raw GHz, but the percentage feature would get old really fast if the UI were lagging every time you swapped dice around.

 

MrBishop, the dev responsible for dice percentages, pretty much said exactly that during a Reddit AMA.

 

Anyway, despite using basic math and standard RNG implementations, there's always been a lot of suspicion about how this game handles random numbers.  I'm strongly of the opinion that things are fine enough (they're at least more random than most physical dice throws/shuffles).  Others seem to have different experiences.

[Eumaios]

Hadn't really thought of it that way, Ethics, but it makes sense. ten minutes sounds like a long time, but I guess that's basically because of the quick dice changes. Moving dice in and out of the equation could cause the calculation to take a lot of time.

 

I haven't had a problem with the number generator myself, and I'm a little distrustful of anecdotal evidence that it's got serious problems. Not that it cannot, just that if I force myself to keep track of what's going on, the numbers more or less bear out over the long haul. In my experience, that is. The other thing is, 99% means 1 out of 100. With all the dice rolls, you're bound to hit that 1 out of 100 fairly frequently. However, since we don't have fractions, 99% also means 1 out 1000 or more. So I try to keep track of the difference, although the truth is it's a game and you have to rely on luck sometimes.

 

Now something that happened to me so often I thought there might be a problem a while back was evading and having the same card pop up every time. I even started keeping track of that because it was starting to feel a bit weird. ...But I stopped playing for a while because I was busy and when I started back again it wasn't happening nearly as often, so I figured either they fixed a problem or I was just suffering from confirmation bias or something on my end.

[Ethics Gradient]

Now something that happened to me so often I thought there might be a problem a while back was evading and having the same card pop up every time. I even started keeping track of that because it was starting to feel a bit weird. ...But I stopped playing for a while because I was busy and when I started back again it wasn't happening nearly as often, so I figured either they fixed a problem or I was just suffering from confirmation bias or something on my end.

 

Shuffles were (are?) one of the randomness issues that probably weren't a result of RNG flukes.  The shuffle algorithm itself should be pretty solid when it is needed.

 

The likelier issue was that the game calls for different types of shuffles depending on the situation.  Sometimes it shuffles everything, sometimes it shuffles  everything below the revealed card, sometimes it shuffles everything except some certain number of X cards, sometimes it doesn't have to shuffle at all.

 

I think it is better now too, but I also swear there used to be an improbable likelihood that evaded monsters and fleeing villains were often found on the top of location decks.  I'm also pretty weary of observational biases where RNGs are concerned, but as the patterns bore out, it didn't seem so much a randomness issue as it was the app potentially firing off the wrong sort under certain circumstances.

[LaSpeakeasi]

The amount of 90%+ rolls that fail is waaay beyond any reasonable doubt that the percentage indicator isn't functioning properly. Anything below 75% is almost sure to fail.

Edited September 13, 2017 by LaSpeakeasi

[BBprojet555]

All I know is I roll a laughably high number of 1s. Like, ridiculously high number. I have to use practically every blessing in the party as insurance. If there's a 95% chance I'll succeed, somehow my dice will find the 1s.

[Eumaios]

I don't think the dice are rigged.  Seriously, I wish I could see how many hours I've put into the app.  Hmm, maybe I can.  No.  in Steam alone, which I only started using well after the tablet app released, I've got 94 hours.  I wouldn't be surprised if I had over a 1000 hours on the app.  Sure, there's a spread, but that's what probability is.  It goes from 0 to 1.  Anything between is fair game.  There are bugs.  There are irritating glitches.  There's even a few design decisions with which I disagree.  I don't see the dice being rigged.  I could be wrong I suppose, but I can't help but believe that the design team is using industry standards to test the RNG.

Edited September 14, 2017 by Eumaios

[Irgy]

It's well known that a lot of people experience randomness in games as being rigged against them even when it isn't. There's lots of factors as to why, someone else already went over some.

 

Now, maybe this is the one game where they have somehow managed to rig the dice for no clear reason, and you're the one person to have actually realised it. The trouble is, even if you are, there's no way to distinguish you from all of the other people on every other forum for every other game involving randomness on the internet saying the same thing and being wrong. If you want to be different, gather some stats.

 

If you gather some stats, one of three things will have to happen:

1. The game will notice somehow that you're gathering stats and stop misbehaving. Problem solved, for you at least.

2. You'll discover that the dice rolls are distributed fairly and evenly.

3. You'll be the first person on the internet to actually uncover a dice conspiracy and provide proof. Eternal fame will be yours.

[Hannibal_PJV]

Yep... I have had extremely good Luck and extremely bad Luck month in the app and real tabletop game. IMHO it seems to be that I have better Luck with dice in the app than real world, but I have not collected hard facts, so maybe it just feel like I am less lucky in real world. But have not seen extra bad dice rolls in the app by my self. Except the normal that you sometimes roll 3 times one when rolling 4 dice, but that is just normal to me in the real world so I Expect to see them Also in the app.

Maybe I am more conserned with dice rolls in real world because other players succes Also depends on my dice, so I remember those bad rolls better...

[Ethics Gradient]

3. You'll be the first person on the internet to actually uncover a dice conspiracy and provide proof. Eternal fame will be yours.

This.  Sure the game has had it's share of bugs, but it isn't running some sketchy, home-brew random number generator.  It leverages the same basic RNG that just about every other Unity game relies on.  Unity's standard functions may not be ideal for repeatedly shuffling a thousand-card deck or casino gambling, but it should still provide a lifetime of "better than tabletop" dice rolls before any pattern sets in.

 

To believe that the RNG is skewed in a human-noticeable fashion, you would have to argue that almost every other Unity game is severely broken too, or that Pathfinder does something to derp the numbers (intentionally or not).  You don't have to believe me or anything, but I've been looking at this for a while and have found little to suggest that rolls are tampered with somewhere between the RNG outputs and what appears on-screen.

 

Anyway...  Just as a couple others have pointed out, low rolls sting and the memory sits with you.  It isn't nearly as noticeable when you steamroll over Pillbug Podiker for the hundredth time and nearly hit him for triple-digit damage.

Edited September 14, 2017 by Ethics Gradient

[King_Momo]

I usually find discussions like these tend to pop up when there's nothing new to play in the game.