dtrader - Thanks for the comment and compliment. I appreciate your question very much actually. If you are counting up periods versus down periods in the forex, the ratio is pretty darn close to 50/50 as you suggested. However, the trick is being able to trade that "randomness" with any degree of risk control. If I apply the same form of risk control to random entries that I did in the test, the actual winners to losers ratio drops to 34%. Not to mention that the r/r ratio drops to 1:1.2. If you put those together random entry gives you a very sour trading record. The reason for this is whipsaws. The stop kicks you out too fast. I found that the candlestick identified market momentum well enough to avoid enough of those whipsaws to turn the test profitable. The difference between the real random 50/50 and the 50/50 in the candlesticks is the probability of a favorable r/r ratio goes way up. That is what I am really after. Does that answer your question more completely?
I see. And how many samples (and observations/sample) did you use to determine the sampling distribution of your random sample statistics? And where did the 34% return lie on the confidence interval? The 34% drop looks biased to me.