Do you remember the Top Gear experiments using a driving simulator shown on the BBC? One group of drivers were plied with alcohol. The results showed a small initial improvement in driving skills that was put down to the relaxing effect of the alcohol. This was followed by a gradual diminishment in skill levels over time, with the number of errors increasing as the booze took effect. This was compared with a second set of drivers who did not drink, but were deprived of sleep. This group maintained their performance level until they reached a point where the the graph nose-dived off a cliff.