To introduce some science to this debate. Mørkeberg and others (Changes in blood profiles during Tour de France 2007. Ugeskr Laeger 170:1916-9, 2008) found significant decreases in hematocrit and hemoglobin concentrations in riders in the 2007 TdF on days 12 and 19 compared to the day before the prologue, with hematocrit 12.1% lower on day 19 compared to the baseline.
In another paper by Mørkeberg and others (Changes in blood values in elite cyclist. Int J Sports Med 30:130-8, 2009), which looked at hematocrit, hemoglobin concentration [Hb] and % reticulocytes in two professional cycling teams in 2007, they found that hematocrit and [Hb] decreased during in-season compared to out-of-season and, more importantly, during the TdF the [Hb] decreased by 11.5 %, with individual decreases ranging from 7.0 to 20.6%.
Lastly, Schumacher and others (Haemoglobin, haematocrit and red blood cell indices in elite cyclists. Are the control values for blood testing valid? Int J Sports Med 21:380-5, 2000) found that hemoglobin, hematocrit and red blood cell count decreased significantly with increasing training workload.
In summary, such a dramatic spike in hematocrit and OFF score during the TdF defies all logical explanations from a physiological point-of-view unless there is some form of supplemental enhancement (ie, autologous transfusion).