In this paper, we investigate two simple but very popular trading rules, simple moving average (SMA) and trading range break (TRB), by looking at a series of data of Hang Seng Index over thirty-five years, from 1972 to 2006. To avoid the data-snooping biases, we apply the out-of-sample tests in additional to the traditional tests to check for the validity of the profitability of the trading rules. In general, our results show that there is one trading rule, the (1, 50) rule, outperforms the market (HSI) over the 35 years testing period and in the pre-1986 sub-periods in our study in the SMA trading rule. These average returns from SMA rule are higher than the returns found in the US market (Brock et al., 1992) and Hong Kong market (Bessembinder and Chan, 1995). Besides, the VMA (1, 50) rule performs better than the FMA (1, 50) rule because it includes the information of the first nine days into investors' decision. With the flexibility of buy or sell within the first nine days, the VMA (1, 50) rule can generate 2.5% to 7.5% (annual) more profit than the FMA rule before transaction costs. However, the returns of the TRB rules are all small and insignificant. We find that the high trading returns are unlikely to be explained by the difference risk in difference trading strategies and transaction costs. However, they can be partly explained by the positive first order autocorrelation in the Hong Kong stock market and the time-varying premium. Furthermore, the inconsistent subperiod abnormal average returns suggest that the stock market integration leads to a more efficient dissemination of information over time in Hong Kong.