My current research is focused on providing a valuable context to analysis of candlestick signals, and I have discovered some very exciting aspects of this.
My research is based on applying my hypothesis to virtual trading over a two-year period. During this time, trading options based on candlestick signals and a hypothetical portfolio of $100,000, I have generated overall net credits totaling over $75,000 in a series of 582 trades.
This was all based on the primary hypothesis: The candlestick signal is not reliable by itself as either reversal or continuation forecast. It requires confirmation, must appear in the correct proximity (at or through resistance or support), and its strength depends on the strength and duration of the preceding trend. The "relative correlation" of the signal is a variable relying on these attributes collectively.
I am nearing the end of a current book project, to be published by FT Press, in which I document a series of tests of this theory. I discovered many interesting tendencies among candlestick signals. For example, false signals and failed signals do not simply occur for no reason. When these occur, it is usually due to weak preceding trends, improper proximity, or weak (or non-existent) confirmation.
In other words, the study of candlesticks should include a careful analysis of the current trend, proximity to resistance or support, and strength of both the signal and its confirmation. The majority of published articles and books about candlesticks focus on the appearance of signals and whether they forecast reversal or continuation. My study concludes that this is only a preliminary analysis, and that making candlestick indicators reliable depends entirely on the relative correlation of the signal, confirmation, proximity, and nature of the preceding trend.