But there’s a problem. Take a
dragonfly doji as an example. The candle looks like the letter T with the close at the top of the candle. You would expect price to break out upward sooner than downward just because of where price closes in the candle.
First, the big surprise with such
candlesticks is that
price does not break out upward the next day. In one test I ran on 20,000 dragonfly doji patterns, it took an average of four days to break out upward in a bull market. Downward breakouts in a bear market took exactly the same time, four days, to make the journey. Yes, the position of the closing price does influence whether a candle acts as a reversal or a continuation, but not to the extent many believe.
Second, if a candle is supposed to act as a reversal, then price should reverse direction, regardless of where it closes in the candle. That’s almost by definition. If you placed a stop-loss order a penny below a candle in a downtrend that failed to reverse, you’d be stopped out. So price should reverse quickly or it’s useless as a reversal indicator.
To be fair, I’m not that stringent. I wait for price to close either above the top or below the bottom of the candle before I determine whether it’s a reversal or a continuation. A hanging man, for example, takes an average of three days to prove it’s a reversal.
Finally, I checked the statistical results, visually, with the candles themselves and found that the method agrees with how I would categorize each candlestick pattern. In other words, if the statistics said the candle acted as a reversal, then it did. If the stats said the candle acted as a continuation, then it did. I used my eyes and counted the results, then compared the results to the stats. Both agree.
Why am I making such a big deal about this? Because 31% of the candles in this book fail. That is, they don’t work as advertised, and many work little better than random chance. There are exceptions, of course, and those are the candles you should rely on. That’s why I wrote this book—to find those gems hidden on the floor of the exchange.
