Why does a least squares fit appear to have a bias when applied to simple data? 00/04 • stats.stackexchange.com