From: Randy Kramer Date: 2005-07-17T09:17:13+09:00 Subject: Re: [QUIZ] Sampling (#39) On Saturday 16 July 2005 04:01 pm, Bill Kelly wrote: > From: "Olaf Klischat" > > > See? One number per 200-numbers interval. Every time[1]. This hints at > > a wrong implementation. > > I used the same approach. I wasn't sure if it was legal > or not. But it was sure easy... :) This is off-point for this particular post, and I apologize for not reading/remembering the original challenge, but when someone commented about it seeming odd that there were so many numbers prefixed, iirc, with something like 999999.., I got to thinking--the odd thing about this challenge is sorting the numbers. I don't think many of us are used to looking at a sorted list of random numbers and trying to judge whether they are random (if that's possible). Two examples: If you took 5 random numbers from in the interval 0..10, you might see something like this: 9, 3, 2, 6,4--"looks" fairly random. Now sort that list: 2, 3, 4, 6, 9--"looks" a lot less random. For even more fun, do it for a series of coin flips (with H for heads, T for tails, (and E for edge ;-))--you might get a sequence (for 6 flips) like H, T, T, H, T, H. Now sort that list: H, H, H, T, T, T--doesn't look very random at all. regards, Randy Kramer PS: Just for the record, if someone implemented a simplified approach which took a random number on each interval of 200 numbers in the original range, I would have to say that was the wrong approach (or, maybe more accurately, I would hope that I had specified the problem well enough so that such was not an acceptable approach).