Most of you are probably aware of the popular book "Fooled by Randomness", by Taleb. His other work includes "The Black Swan: The Impact of the Highly Improbable".
Many of you also know that I like to place emphasis not on some mechanical entry signal, derived by some indicator, but rather on keen risk management and trade management.
The point of this thread is to see if risk management and trade management alone can be profitable. Most of you know that my math skills are very impaired, so I need your help. Please do help...
First, I created a NinjaTrader strategy back in Feb 2010:
I never went much further, as I stopped using NinjaTrader basically the same time.
Recently I came full back to this topic, as it does certainly interest me. I created a new random trade framework in EasyLanguage for MultiCharts, and submitted it in the Battle of the Bots simply for fun:
But I'd like to improve on this idea and run with it for a bit, if I can. Here are the basic objectives:
- Try to randomize entries as much as is practical. Do this by using a very large sample size, multiple instruments, multiple time frames, and a random number generator to randomize which bars are used for entry points. I think I have years of tick data for several instruments, which should be enough for billions of possible trade signals.
- Apply risk management in terms of position sizing, stop losses, daily loss limits, weekly loss limits, monthly loss limits, or whatever else makes sense in terms of managing risk. I am open to methods to measure this, such as volatility, risk of ruin, expectancy, true range, standard deviations, whatever...
- Apply trade management in terms of protecting profit on trades, and getting us out of trades. This includes things like moving stops closer, setting ultimate profit targets, etc. It could also include using something to tell us we should be flat (such as volatility).
- Determine if adding more inputs, like input from trade data outside of the instrument being traded, will improve the profitability of the system. Examples might include breadth data like TICK, VIX, etc, as well as correlative info with other instruments, or maybe a spread or hedge framework to diversify risk.