[MUD-Dev] Expected value and standard deviation.

[Robert Zubek]

From: Ben Chambers [mailto:gtg983q at mail.gatech.edu]

>> The thing is, xDy dice numbers do the same thing, simply
>> increasing the floor of the curve.  Lots of dice means a high
>> center and floor,

>> while large dice increase the spread.

...

> I agree.  But as a developer, it is in my opinion more intuitive
> to be able to manipulate the spread and center independently.  ...
> The other difference is that a normal distribution produces
> numbers that are continuous -- a decimal number with theoretical
> range from -infinite to infinite.

Yes! And here's another benefit: non-uniform random number
generation lets you pick not just the parameters, but the frequency
distribution, too.

Dice fall into normal distribution, but the developer may want to
model processes that don't fit a Gaussian. For example, sometimes we
may want to model event success using an asymmetric Poisson, or an
exponential distribution. That's a bit awkward to do with dice. :)

Rob

--
Robert Zubek
http://www.cs.northwestern.edu/~rob
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev