The so-called "golden-cross" and "dead-cross" are said to be useful signals to forecast market trends. In this paper, we focus on the Japanese stock market where gold-crosses and dead-crosses are empirically considered as useful investment signals. First, we examined the usefulness of these signals by using historical Japanese stock price data. The results confirmed that these crosses were useful as confirmatory signals for forecasting market trends. The results also showed that the minimum length of period (days) of useful moving averages is shorter in the case of golden-cross than in that of dead-cross. Second, we tried to identify the underlying reasons for the usefulness of the crosses. Because a model, which assumed all investors were rational financial experts, failed to explain the usefulness of the crosses, we were able to assume that the crosses reflected investors irrationality or behavioral bias: conservativeness and representativeness about trends (Tversky and Kahneman 1974). We then developed a model that incorporated this bias. Based on simulations using this model, we identified the mechanism with which those crosses closely relate to investors irrationality or behavioral bias. The analysis also revealed investors tendency that they were convinced by a bull trend more easily and quickly than by a bear trend. This finding is in line with what is generally observed as investors’ bull-bias in the Japanese stock market.