I guess you really need to think about your data before you decide what method to use.
That could be a very large pitfall though because people can choose their rounding method to make their data fit better. Although in most cases it's not going to make a big difference.
George, I agree. The one thing people say is that you aren't rounding .0, and thus, it doesn't count. I can't understand why though.
]]>0.00 0.01 ... 0.49 50 numbers
0.50 0.51 ... 0.99 50 numbers
fair
0.000 0.001 ... 0.499 500 numbers
0.500 0.501 ... 0.999 500 numbers
fair
so the rounding can be inferred (imagined) as fair.
when talking about infinity, you can only infer and define.
It really is about "horses for courses", but the "5 up" method is simply the most common and agreed-upon method.
I guess you really need to think about your data before you decide what method to use.
]]>Finally, 5 is exactly halfway, so there is no fairness.
Teams with an unfair zero on a team is not fair.
Sorry we already talked about this before, but I couldn't resist mentioning it again.
]]>