Information on calculator Random Number Generators from PPC Journal articles
|
11-07-2016, 10:35 PM
Post: #8
|
|||
|
|||
RE: Information on calculator Random Number Generators from PPC Journal articles
Hi, Gene: (11-06-2016 08:29 PM)Gene Wrote: These are some of the findings that led programmers away from (Seed + PI) ^2 or ^5 etc. and into the 997 x generators of the best of all 9821 x 0.211327 + models. Frankly, this 9821... is utter overkill for the stated purpose of providing some randomness to calculator games. Back in the days of the 25C, 67, and 41, I did always use the smallest, fastest of them all, namely: RCL 00, R-D, FRAC, STO 00 and it worked great for the stated purpose, calculator games (not statistical analysis, not Monte Carlo algorithms, etc). I sent a mini-article to PPC recommending it and remember that someone subjected it to test after test after test till he finally succeded in finding that it wouldn't pass the Spectral Test precisely, which was a foregone conclusion from the get go. As long as you don use 0 or a multiple of Pi for the initial seed, this R-D generator saves bytes and time and provides decent results for games without quickly repeating, certainly you'll never play any of them thousands of times in a row. Regards. V. All My Articles & other Materials here: Valentin Albillo's HP Collection |
|||
« Next Oldest | Next Newest »
|
User(s) browsing this thread: 3 Guest(s)