At 10:21 AM 11/17/2005 -0500, you wrote:
what you do with the random number to produce a dice roll is more
important. It is easy to make a simple mistake in the mapping of a random
number to an integer between 1 and 6 that results in a skewed distribution.
Here is a link to get you started,
http://www.gamesgrid.com/backgammon/faq-rng.html
a Google search will probably turn up more info as well, although doing the
right thing with poor input still produces substandard results in the end
-- especially if you or your client is accountable for the quantitative
aspects of the game, which includes the statistical qualities of a RNG,
process for producing a shuffle/dice roll/etc., and the means for
introducing entropy into the process. Back in the early days of F7, I ran
a couple industry-standard tests on Math.random() and the results were
sufficiently poor to convince me it was not worth running any more
tests. Math.random() hides the seeding process from the programmer,
meaning that you have no control over entropy. Even if you have a perfect
algorithm for producing a dice roll or shuffle, you can not make an
argument for the fairness of that game that is in line with standards
currently in practice for online casinos. If you don't need to meet such a
standard, that's fine, but consider your level of accountability before
selecting a method of approach.
Gamblers always complain about dice rolls even if you use real dice.
How many times have I done that at the craps tables :) This is, of course,
why the gaming commission insists on rigid tests for 'fairness' of casino
dice. Also why online casinos go to such great lengths to reassure patrons
of the statistical qualities of their shuffles/dice rolls/etc.
good luck!
- jim
----
2112 FX :: Singularity [Business Intelligence][Custom Programming]
Flash Math Blog :: www.2112fx.com/blog
_______________________________________________
Flashcoders mailing list
[email protected]
http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
