I agree with your comments on Random Number Generators - however -
criticizing my comments - I wonder how long you've played Everquest <g>...
Their RNG is CLEARLY flawed. In particular - I use SSOY in the Lord Room as
an Example. I know MANY people who have LITERALLY camped it for two weeks
in a row maybe a total of let's say 50 hours. They report going days and
days of never seeing one - and then getting five in one session. I've heard
of this happening on several occasions.
Another example is Proc's and Failed Notes. I almost never see just one or
three in a row. I'd say that about 85 percent of the time - I get two procs
or two failed notes in a row. I don't know if I've EVER seen three procs in
a row. I RARELY have seen three missed notes in a row (discounting when low
instrument skills is a factor).
To "Perhaps a Princess" this is not Random according to the "Experts" in the
articles I've read. The issue of extrapolating on a small sample aside
(which again I grant is true in my test on Kaesora but not my example of
procs/missed notes) A RNG should produce a truly random sample. For a
Binary test. Consistently getting 45-50 0's followed by 2 1's 85 percent of
the time is bad code.
On the other hand - perhaps this is all a part of "The Vision" and they
specifically coded the RNG to work that way.
I don't know a lot about a Statistics - but I've been in the IT industry for
16 years and read lots of articles about Random Number Generators for
computing systems. There are tests that can be done on the algorithm as to
how Pseudo-Random they are (obviously - by definition - an algorithm can't
be truly random) I'd be curious if Verant has done this analysis.
This is evidence of poorly coded random number generator. The Prosecutions
rests.
Slyde
-----Original Message-----
From: Jeffrey Sue [mailto:
jsue@...]
Sent: Monday, June 12, 2000 12:49 PM
To:
eqbards@egroups.com
Subject: RE: [eqbards] My day in Kaesora was Re: MM Drum misadventure
and Nostrolo Tamborine
umm, do you know what ''random'' means? and have you taken any statistics
courses?
do you honestly think that a sample size of 7 is statistically significant
for a typical
random number generator?
here's an example. a pair of dice has 36 possible combinations. roll the
dice 7 times.
even assuming you don't duplicate any of those combinations during those 7
rolls, that
means you haven't rolled 29 of those combinations. by the stats you've
accumulated
they have a 0% chance of occurring. obviously false. and that's only 36
combinations.