Using the central limit theorom, as the number of dice rolled grows, the closer they resemble a normalized distribution curve.
We introduced a new function that allows us to roll on a normal distribution and received distributions, though they trend less centrally than the original:
This optimization allows us to roll once on the normalized random generation, rather than N times on the uniform random generation.
This is slower than the uniform randomness at low frequencies, but at higher frequencies we get a near constant runtime instead of the old version which scales with number of dice.
Warning: There is one major drawback to this approach in that it completely collapses the dice pool, so operations such as dropping/keeping/introspection cannot be performed on the result.
Using the central limit theorom, as the number of dice rolled grows, the closer they resemble a normalized distribution curve.
We introduced a new function that allows us to roll on a normal distribution and received distributions, though they trend less centrally than the original:
This optimization allows us to roll once on the normalized random generation, rather than N times on the uniform random generation.
This is slower than the uniform randomness at low frequencies, but at higher frequencies we get a near constant runtime instead of the old version which scales with number of dice.
Warning: There is one major drawback to this approach in that it completely collapses the dice pool, so operations such as dropping/keeping/introspection cannot be performed on the result.