> the TdF total race times would be very weak evidence for it. Actually, the jump in TdF average speed (not total race time) during the early 1990s is fairly o

> the TdF total race times would be very weak evidence for it. Actually, the jump in TdF average speed (not total race time) during the early 1990s is fairly often attributed to EPO (and other drug) usage. The average speed in earlier events was lower because of the inconsistency - cyclists would ride much slower on days before (or after) a big stage, and this would drop the average speed for the entire race. EPO and other drugs improved recovery so cyclists would ride hard everyday. It's hard to know how true this is - it also corresponded to increased TV coverage for full stages, and more money in the sport. But the average speed jump after 1989 is significant.