D&D 5E Roll20's RNG - or Evil Electronic Dice Gods?

[DaveDash]

So in my session I rolled 4 crits against a player during one combat, and two in one turn. Basically 2 crits in a row. It turned what was meant to be an 'easy' encounter into something else. Basically if my player didn't have a belt of Dwarven Kind, I probably would have killed a Level 15 Paladin with a CR6 Wyvern.




Fast forward to tonight's session in which I am a player. In the very first combat of the night, the DM (the same player as above) rolled 2 crits in a row against my character. I am not kidding.



What the hell roll20? Is this the trick of some cruel dice god?

Although seriously, I think the same rolls come up way too often in roll20. Anyone else notice this?

 

[Blackbrrd]

I did something similar in my 4e game, but with my usual dice, rolling openly in front of the players. One of the players then gave me a new dice, and I kept rolling high. Random is random and random might be rolling 20 several times in a row.

 

[KarinsDad]

I did something similar in my 4e game, but with my usual dice, rolling openly in front of the players. One of the players then gave me a new dice, and I kept rolling high. Random is random and random might be rolling 20 several times in a row.

While this is true, computer programs are often notorious for having poorly implemented random number generators, either the underlying code, or how the programmer using them implements it (the latter often being the most likely culprit). My experience with web based random number generators is that they often fall way short of being "random enough" with certain numbers coming up higher than statistically expected.

A simple test is to use Roll20 to roll 10,000 D20 numbers in 100 groups of 100 to determine patterns or statistical anomalies (if that is possible with that tool).

 

[iserith]

Information on Roll20's random number generator:

Quantum Roll Wiki
Quantum Roll Blog
Quantum Roll Status

 

[Shiroiken]

They just had to do major maintenance their system last week, and it might still be a bit buggy. I rolled 3 20's in a row for skill checks yesterday. It could have been random chance (with pure RNG, you can never be sure), but I suspect it might have a bug or two to work out.

 

[wedgeski]

While this is true, computer programs are often notorious for having poorly implemented random number generators, either the underlying code, or how the programmer using them implements it (the latter often being the most likely culprit). My experience with web based random number generators is that they often fall way short of being "random enough" with certain numbers coming up higher than statistically expected.

Every programming language has its own random number generator, and it is very, very hard to use these things incorrectly. Yes, you could seed with the same number every time. That would be a mistake. And yes, the algorithms used have some limitations at the extreme edges of statistical consideration. But all in all, silicon-based pseudo-randomness is entirely adequate to roll a few fake d20's, and any pattern the OP is seeing is just, well, random.

 

[GlobeOfDankness]

i've had a player get three crits down the line in one of my games. i've also gotten four crits against a PC in the same round (though advantage was involved.)

 

[Psikerlord#]

I've never seen anything like this with Roll20, but then they did just upgrade their dice system, didnt they?

 

[DaveDash]

Every programming language has its own random number generator, and it is very, very hard to use these things incorrectly. Yes, you could seed with the same number every time. That would be a mistake. And yes, the algorithms used have some limitations at the extreme edges of statistical consideration. But all in all, silicon-based pseudo-randomness is entirely adequate to roll a few fake d20's, and any pattern the OP is seeing is just, well, random.

I see the same number coming up twice in a row far too often though. Granted I could just be "lucky", and the sample size isn't large enough to be conclusive. On the other hand though various RNGs used by programs have been proven to be flawed in the past, so it's something I am suspicious of.

 

[lkj]

I can't say whether roll20 has a problem in their dice roller or not. But I can say that humans are notoriously poor at understanding and identifying random patterns. We tend to think random results should be evenly distributed. But truly random results will have all kinds of streaks in them. One can usually identify a truly random pattern from one attempted by a person by choosing the one with more streaks (like a bunch of 20s in a row for example).

AD