Nickola,
Long time members of the old VIT group know that this is exactly what happened to me last year (2010). I had a compounded YTD gain of over 80% in my discretionary accounts until near the end of May. Then, a series of losing discreionary trades compounded 6-7% trades losses without interruption to bring my accounts in the red just prior to the explosive September bull run. I had lost all confidence in my trading, so when I went long at that time I only started trading 1/3 of my previous position sizes and I could never make it back in spite of the best opportunity to do so.
Besides the psychological healing process, I scrutinized every one of my losing trades to find out what could have avoided the disaster. It was not risk management. My discipline was strict about cutting losses fast according to my trading plan. It was the compounding effect of the "small" losses that was the culprit.
My risk management was and still is based on my multi-pivot methodology. It gave me "optimal" buy and short entries for each day with an "optimal" stop loss. But the decision to take the buy or short entry on any day was entirely discretionary; like you I was following my feeling for market rhymes. And i've been 100% wrong for three long painful summer months. Then I looked back at all the 20 DMF signals over the period and I immediately saw that simply trading my methodology in the direction of the 20DMF instead of my best market direction guesses would have avoided most of my disaster. And, at a minimum, I would not have experienced the vicious negative compounding effect. And I would have started the bull run in September on full margin from accounts up 65% YTD.
On his side, Pascal was doing very well compared to me but analyzing his trades he noticed he could improve a lot the risk management of his trades using my methodology. His Einsteinian reflex was to backtest and backtest and backtest, filtering out every noise that didn't provide any risk-adjusted edges.
And the robot was born, optimizing the mixing of proven outperforming market signals with a proven outperforming risk management system.
Billy