There is still something that I can't seem to understand. If you sell/buy short/cover positions every 20 days why do you need 6 months (long term) to demonstrate the gains ? You should be able to do so in 60-80 days at most. Just a beginner trying to understand.
Alpha, Nice open thread and system. I have to call you on the blackjack comment, though. Most trading systems, including yours, seem to be doing exactly what card counters do at the blackjack table - taking a relatively small statistical advantage and playing that same advantage in many hands to get a positive return. I have to admit that some trading systems seem to have a much higher theoretical advantage than a blackjack player could ever hope for. Of course, knowing the bj table rules you can know the future expectancy with a fairly high certainty. While with trading systems you are assuming future expectancy based on past statistics - the backtesting problem. Keep up the good work.
I will use an example to demonstrate: Lets say I have a strategy that buys any tech stock that starts with the letter 'A' such as Apple or Agilent on the open and sells in on the close. If you started this strategy in the late nighties you would have looked like a star and made tons of money, does that mean it's a good strategy? Of course not. Point being that for any strategy to prove itself it must be traded through various market cycles, bull markets, bear markets and everything in between and that takes time regardless of what the holding period is. Does this clarify things?
You are correct, maybe the blackjack analogy was not a good one. In fact gambling and trading share many of the same aspects. One being that in gambling there is a small amount of people making all the money, while most people loose. The same goes for trading, but less extreme probably. One thing is for sure, going to Vegas is more fun than going to the NYSE
Alpha, I've got a few questions other folks might find interesting too. - which software(s) did you use to develop and backtest your system? - how long did it take to develop including previous systems leading to this one? - how are you backtesting the discretionary/non-fundamental parts of the system?
1. I use a combination of Zacks Research Wizard, Tradestation and Excel to do my backtesting. 2. I have been working on various systems and analysis for around 10 years. It has been a very interesting and sometimes very frustrating road. 3. A backtest is an approximation of how something did in the past. It is not perfect and it does not tell you what will happen in the future. I cannot backtest discretionary decisions or exact entry/exit prices. That said, the backtest is the only possible substitute for actual trading results and it is surely better than pulling numbers out of a hat.