Russell 2k has a fairly high correlation to ES nowadays, I think you would end up simply with higher returns coupled with equally higher drawdowns. Your best bet is to trade futures from a different country or going with something like an emerging markets ETF (much lower correlation as compared to Russell 2k). I won't mention something like energies as your edge in those probably won't work.
Just my opinion but personally that raises a big red flag and strongly suggests your system is over fitted.
Good luck with everything - you were smart enough to create a winning strategy for a while which is more than what most can achieve. Take a break, enjoy life and come back to it with fresh eyes - you might find a new edge
Yes, it's an on-going effort and lifestyle. I already have a strategy for CSI300 which have rendered 250% annually with Sharpe ratio of 3.67 during paper trading, so I have something to look forward to at least...anyhow thanks for the kind words!
Turn A Good System Into An Amazing System with Equity Curve Feedback! Equity curve feedback is an exciting technology that can help you turn a good system that you may have into an amazing system! It can be incorporated either into your own trading strategies or as a filter for other strategies. This technology can help you reduce drawdowns and preserve capital. As any trader can tell you, this is essential to reducing risk and preserving your hard-earned money! Murray Ruggiero's exceptional TradeStation toolkit, the Equity Curve Feedback Toolkit, lets you easily incorporate this powerful technology with as little as one line of code. Designed for TradeStation, this toolkit lets you have a trading simulator that keeps track of the equity curve separate from your real trading strategy. You can then turn your system on when the equity curve is positive and turn it off when the equity curve is negative. Click Here for more information
He optimized random elements coincidentally associated with non-random elements. A fluke that can't work but labeling it over fitting perpetuates misuse of that term. An over fitted strategy is assumed to have targeted non random elements but it found a peak rather than a plateau. If we can't make the assumption a strategy targets non random elements then the difference between peak vs plateau would be moot and the term over fit wouldn't have any meaning at all.
I read this message 3 times but I still don't get it. Perhaps you could explain what 'random' and 'non random' elements are? Maybe with an example? GAT
Non random data features are the recurring logical advantages the machines perpetually embed within the data via their persistent preference for one algo over another.