This article tests the overreaction hypothesis using data from the UK stock market. The study covers a period of 30 years (from 1973 to 2002). The results initially seem to be consistent with the overreaction hypothesis and no obvious seasonal pattern can be identified. Our results do not depend on whether buy-and-hold returns (BHR) or cumulative abnormal returns (CAR) used to compute the returns of the arbitrage portfolio. The overreaction phenomenon is still observable even after controlling for the size effect and the time-varying nature of risk.