There is no problem with this as long as you keep track of your positions and make sure they square at the end of the day, week, or month. It would be a pain in the butt to do day after day. (If you miss closing out a leg you can easily screw yourself with this clever approach.) Hiding your system's trades can also be done by trading several systems in the same account. The number of trades you make could also help/hinder a 3rd party's understanding of your system. For example, a system that trades 5 times a day may be easier to reverse engineer than a system that trades 2 times a month. Simply more data examples. The big picture here is being missed, why would anyone take the time to steal a system? 1. Systems are so tied to an individual trader's personality that if you gave a successful system to someone else they will probably screw it up. 2. In the book on "When Genius Failed", investors thought LTCM had 'unique technology', instead most of their techniques had been published in financial journals and were in the public domain. Pick a popular system, search the web and you can find a good analysis of it. Turtle, aberration, DollarTrader, no problem. Pay $400 for the CD of all past issues of Stock & Commodity and you can find lots of trading system examples. TradeStationWorld and WealthLab also have lots of examples. If your system was something 'really' new it would be very difficult to reverse engineer because it would not fall into the trend, countertrend, breakout models the reverse-engineers were looking for. 3. Without knowing the details of a system why would someone try and reverse engineer it? Some systems are hot for a while and then crash. Watching trades of a hot system, duplicating the results, may only deliver a system that is about to crash.... Why do this? Anyone smart enough to reverse engineer a complicated/successful system is smart enough to do an equal or better system themselves. 4. LTCM was using multiple accounts as much to hide the scale of their specific positions as the techniques being used to buy and sell. Just some thoughts... ramora
I agree. Reverse engineering a trading system would be attempted by running throug a catalog of indicators. Did the trader use MACD? Or moving average crossovers? Or...? If your strategy isn't on the list of strategies someone might compare your trades to, then you don't have to worry. It is just like guessing passwords or solving those series puzzles (like "what comes next after O, T, T, F, F, ...?") These are also solved by trial and error and if you have a password, series puzzle, or trading strategy that isn't going to be tried, then it won't be guessed/solved/reverse engineered.
Almost none of systems can be reverse engineering. If anyone really wants to do it, use ANN to learn the systems - an approach from many others.
I have heard of a commodity broker who, in the phone-age (e.g. before electronic trading) - would piggy back off trades made by a big customer who had a hot hand. Supposedly he traded other customer accounts (who paid high commissions) and not his own account, which he felt made it o.k. I would suspect this went on a lot.
Yes, trade all four (or more) strategies in each account. If you have one system that makes 100% in a year and three that make 25%, you'll still do great and it will much tougher to reverse engineer. It's a lot of work, but what is easy about trading?
Didn't mean to imply I'd develop a system strictly based on it's "hideability". Nobody's going to bother with me while I have a small account anyway. But as you get larger, it's naive to think you won't be on somebody's radar screen...