Summary: Due to the dispute in academic research surrounding the effectiveness of pattern analysis in foreign exchange markets, a method to gauge the information of complicated pattern analysis is proposed. The original models of pattern definition and parametric selection are improved. This opens up a new perspective of the research on the effectiveness of pattern analysis. By comparing the unconditional empirical distribution of daily returns with conditional distribution dependent on specific technical patterns, it is found that most of the patterns provide incremental information and may have some practical value. This conclusion is reached based on twenty pieces of exchange rate data from foreign exchange markets. It is proven that some technical patterns do exist and provide significant information.