Message246642
Serhiy: there's already a small bias inherent in using `int(random() * n)`, regardless of double rounding, since in general `random()` gives 2**53 equally likely (modulo deficiencies in the source generator) outcomes and `n` need not be a divisor of `2**53`. I don't think the double rounding is going to make that bias noticeably worse.
See issue 23974, which was resolved as "wont fix". |
|
Date |
User |
Action |
Args |
2015-07-12 09:05:49 | mark.dickinson | set | recipients:
+ mark.dickinson, tim.peters, rhettinger, pitrou, vstinner, steven.daprano, r.david.murray, skrah, serhiy.storchaka, Serge Anuchin |
2015-07-12 09:05:49 | mark.dickinson | set | messageid: <1436691949.22.0.168511901917.issue24567@psf.upfronthosting.co.za> |
2015-07-12 09:05:49 | mark.dickinson | link | issue24567 messages |
2015-07-12 09:05:49 | mark.dickinson | create | |
|