Anonymous user
Talk:Hofstadter-Conway $10,000 sequence: Difference between revisions
Talk:Hofstadter-Conway $10,000 sequence (view source)
Revision as of 21:37, 8 December 2010
, 13 years ago→Mallows' Number=: the main problem is the task descript is wrong
(→Mallows' Number=: the main problem is the task descript is wrong) |
|||
Line 8:
According to [http://mathworld.wolfram.com/Hofstadter-Conway10000-DollarSequence.html Mathworld], the correct number is 1489, yet some of the entries are giving different results. Shouldn't they be marked as incorrect?
:Well, the only ones that are correct according to the task description are the ones giving 1490, but that's because the task description is incorrect. The task description has n >= p whereas the Mathworld page you mention says i > n, not i >= n. So I think we'll need to fix both the task and the programs that were written to that spec. Oddly, at least one of the programs that gets 1489 (C#) appears to get the right answer for the wrong reason; it's returning a result using 0-based indexing, which compensates for adding one to match the bogus task description. The Ada and Algol algorithms are truly incorrect, insofar as they are computing the last *maximum* that was larger than .55, not the last ratio. a(1487) is larger than a(1489), but a(1489) is still above .55, so wins as the last one. --[[User:TimToady|TimToady]] 21:37, 8 December 2010 (UTC)
| |||