Yes, but only for a relatively small sample (compared to the years you mentioned) for the YM in 2005. I was trading a breakout strategy (purely discretionary) and entering with a stop order. No doubt what you found is correct. But there are two details about what I am doing that you must consider: (1) After the waiting time is over, I am placing a limit order entry near the breakout price. I am entering when the breakout retraces. (2) My targets are small (5-10 NQ points) And the big one: This "semi-automated" technique I am trading is based on the idea of trading in today's market conditions. So if I notice most of the breakouts have pullbacks near the breakout price (true, recently in the NQ) I fit parameters on the fly. If next week I start seeing a lot of the explosive type, I can change the parameters once again (including the profit targets). I don't know if all of the will work; we'll see. So far, so good. I am heavily influenced by the first professional trader I met who would first analyze the previous trading week. If the prices showed certain dependencies, he would scratch out a methodology for the current week. He always did very well. In my case, I'm looking at yesterdays prices plus what happened up until "now" which is typically around 8:30 Eastern for NQ trades. I am also reviewing trades from the past few weeks and thinking about how I could filter out the bad ones while retaining the good ones.
Today: buying breakouts of the premarket high (PMH) and buying support around yesterday's close. Back to real $ (no sim was needed); see below for the details. First, I realized that I could fix the repeated bad entry problem (when support becomes resistance and vice versa) by first requiring the price to be more in the center of the zone. THEN I realized I didn't have to write any new code to do this. I already support multiple conditions for an entry, so I used a "breakout" condition to make sure the price was where I wanted before looking for that support/resistance trade. AND THEN I realized that I have actually created a trading programming language, not a just a trading program. It doesn't look like it now, but if I wanted to (I don't), I could create some syntax and rules plus a program to turn such source code into an XML input file. Currently, I specify all the possible trades for the day in an XML input file directly.