[Shootout-list] k-nucleotide golden answers

[Daniel South]

Daniel South wrote:
| Simon Geard wrote:
|| Daniel South wrote:
||
||| James McIlree wrote:
|||| Somehow, I'm getting a different answer for 30305/100000
|||| than you are. I think the answer should indeed be 30.30, but
|||| I am not a fp whiz.
|||
||| Rounding error. 30.305 rounds to 30.31. 0.005 rounds up, 0.004
||| rounds down.
|||
|||
|| Isn't it round to even:  30.305 -> 30.30 ; 30.315 -> 30.32 ?
|
| Good point. This is statistical data after all so the statitical
| definition of round should be used.

But then again :-)
It's not the data that being used to generate the statistics that is 
being rounded, but the final output. Also the data is empirical and does 
not come under significant digit rounding (statistical rounding, 
banker's rounding) rules. So 0.5 goes up, and 0.4 goes down, when 
rounding to one unit.

Daniel South