The fluting of Early Paleo-Indian bifaces and associated failure rates have received considerable archaeological attention---perhaps at a scale exceeding its comparative importance. Nonetheless, if we are to properly evaluate explanations of the origins and purpose of fluting and the function of particular sites in Paleo-Indian settlement systems, estimates of failure rates are required. To date, such estimates have been based largely on experimental replication and rarely on archaeological data, and it is suggested the experimental estimates are inaccurate. In order to overcome these problems, three interrelated, yet independent, methods of estimating fluting failure rates from archaeological data are developed and applied to data from the Parkhill site in Ontario and the Windy City site in Maine. The consistency in the results obtained by these methods inspires faith in their relative accuracy. The results suggest failure rates, at least at these sites, were not on the scale often assumed by previous investigators, being on the order of only 10--15% or less.