Matthew Toseland wrote:
> I do think that we should take into account the median ... fewer
> arbitrary parameters is generally better.

Sounds good, let's go with the median.

Cheers,
Michael

Reply via email to