On Mon, Dec 20, 2010 at 01:04:18PM -0600, cory n wrote: > > length(sample(25000, 25000*(1-.55))) > [1] 11249 > > > 25000*(1-.55) > [1] 11250 Besides FAQ 7.31, look also at http://rwiki.sciviews.org/doku.php?id=misc:r_accuracy for further examples and hints concerning rounding errors in simple situations. Petr Savicky.