I thought about that, but what if one of your strategies is insanely profitable. What's to keep them from keeping it for themselves and telling you it wasn't profitable?
It doesn't matter what the programmer thinks about or says about the system he programmed for you. You get the coding back so you can backtest it yourself and see what it does. If you are worried him stealing the system, I usually ask him the test the system on different instument/time frame than the ones I will actually use. Even so, he could steal it and try and discover when and where it works. I guess I just don't worry about that...
I'm presently trading these Index futures: eMini S&P, eMini Nasdaq, eMini Russell, EOE (Amsterdam), All Shares (Johannesburg), CAC-40, Dax, Dow-30, FTSE-100, Hang Sing, Topix, Ibex-35 (Madrid), Nikkei, MSCI Singapore, MSCI Taiwan, DJ EuroSTOXX, DJ STOXX-50, Aussie SPI-200, Canadian S&P-60, and a few smaller ones. Standard same-system, same-parameters, many-markets kinds of stuff. Daily data "EOD", fewer than 10 round trips per year per market. Performance has been satisfying, correlation to backtesting in Blox has been quite nice. Hi, MGJ: Are you still using the same-system, same-parameters, to trade many-stock-index-future markets now?
That quotation is from a post made eleven weeks ago: http://www.elitetrader.com/vb/showthread.php?s=&postid=1878403 Have I changed anything in the intervening eleven weeks? No. I still trade these (plural) same systems (plural) on the same very large, globally diversified portfolio of futures markets as before. Stock index futures are just one sector of the portfolio. Often on ET, someone will mention "the stock index futures" when they mean only the eMini Dow, eMini S&P, eMini Nasdaq, and eMini Russell. I posted as a reminder that there are other stock index futures, both domestic and foreign, that people actually do trade.
My contribution: We have spent about the last three years developing our own data set and custom data base with tick data on 15 global meakets to cover high frequency systems and after all the work we have done and the $250K I have invested into this project it comes down to three things that others have mentioned here before. Dynamic systems that can adjust to currect market structure, simplistic methods to help data mine and knowledge of how to adapt to volatility when it moves around and will revert to a mean eventually. I hope this may help some in their quest. Throw out all the indicators and focus on price action and some simple volume filtering and the ability to track volatility.
I backtest with 2 years of tick-level data for YM, which is almost 12M ticks (after removing redundant ones, i.e. ones with no price change). I wrote my own backtesting and strategy optimizing software. It can do a simulated trading run over this data in about 10 seconds, which makes it feasible to automate testing of various parameter values. I used such a long time period so I could find something that is a best fit for all market conditions; 2 years covers bull market, bear market, credit crunch, high volatility, low volatility, etc. My strategy lost money sep-oct '07, but made money for the rest of the period from March '06 until now. It is a fixed mechanical system, just repeating the same trades over and over again, with some features that try to break up losing streaks and minimize drawdowns.