Tonight we played through a battle where we had 2 Nat 20s and 10 Nat 1s (8 attack rolls and 2 saving throws). It may be observational bias, but last week, we had nearly the same disparity. Is there any way to determine if the random number generator of D&D Beyond is truly random?
Tonight we played through a battle where we had 2 Nat 20s and 10 Nat 1s (8 attack rolls and 2 saving throws). It may be observational bias, but last week, we had nearly the same disparity. Is there any way to determine if the random number generator of D&D Beyond is truly random?
It's probably observational bias. Any random generator is likely to produce erratic distribution in short runs of results, but we only notice it some of the time. (Would you notice an unusual amount of 11s? How about too few 3s?) If it didn't, it wouldn't be random.
You certainly could test the generator to see how good it is; the methodology is well-known. I assume they have tested it, but I'm not about to go digging to look.
Even if it's not as good as it could be, it's almost certainly good enough. We're not doing high-security cryptography here; we're generating small runs of numbers in small ranges.
(Also, you can rest assured that it's not "truly random". Computers can't do "truly random" without some kind of external source of randomness, and that's way more work than needed for the task at hand.)
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
Tonight we played through a battle where we had 2 Nat 20s and 10 Nat 1s (8 attack rolls and 2 saving throws). It may be observational bias, but last week, we had nearly the same disparity. Is there any way to determine if the random number generator of D&D Beyond is truly random?
Thanks in advance, Kiddoc73
It's probably observational bias. Any random generator is likely to produce erratic distribution in short runs of results, but we only notice it some of the time. (Would you notice an unusual amount of 11s? How about too few 3s?) If it didn't, it wouldn't be random.
You certainly could test the generator to see how good it is; the methodology is well-known. I assume they have tested it, but I'm not about to go digging to look.
Even if it's not as good as it could be, it's almost certainly good enough. We're not doing high-security cryptography here; we're generating small runs of numbers in small ranges.
(Also, you can rest assured that it's not "truly random". Computers can't do "truly random" without some kind of external source of randomness, and that's way more work than needed for the task at hand.)