INTRODUCTION:
I have been always interested in price action, the moments of candlestick and volume. The candlestick is itself a mirror of the bias, opinions, thoughts of the participants of the market. I prefer using candlestick and price action and the formation of prices to trade over indicators, not saying indicators, oscillators and other strategies are bad... it lags. That lag is critical as the entry signal by other strategies might be right but if it is not scientifically in distance the market could have totally changed. I have seen there are strategies getting 80-90%WR out of lagging indicator strategies. But the factor of lag, scientific lag is what I am aiming for.
I created Doctor XAUUSD to have an EA that won't be slow. I have merged 5 of various price action and pattern recognition logic, for few example... I have used a logic that will recognize DOJI candlestick pattern, from what I know Doji is a critical price action and tells the story of reversal... if there is a strong directional bias but at the end there is Doji, I understand that the market has lesser confidence, more confusion. And this Doji will create a entry signal, but entry alone is not enough... exit is equally important. For exit, all the strategies have their respective exit logic, with scientific TP, SL and trailing. Core reversal on the base.
COMBINING ALL OF THEM:
The main issue that making a multi-strategy EA is that all the strategies must work with each other in sync, if one strategy is bothering the other one the WR, RR, DD will be largely impacted. Plus the capital pool is used by all the strategies, so we must analyze on every single point on how one strategy is using the capital/deposit. Plus position sizing to position management should be in balance with each other.
Building a strategy can be easy, but combining all of them can be hard. Conflict of interest can surely arise. The Doctor XAUUSD uses multiple timeframe, one strategy is scalping looking at 5min, but other strategy is looking at 4h, 1h and 15min to have a bias and there after generate a signal.
CONCLUSION:
What I suggest one that is building a MULTI strategy EA are:
1) Build individual strategies in different environments
2) Only pick the best out of the best characters of single strategies
3) Put up individual comment tag while backtesting
4) Analyze using the tag in journal and position in visual tester, and look if one strategy is hampering the other.
5) Try multiple position at a single time, and only one position at a single time.
Good luck guys, if you have any question about building this kind of strategies and how I solved problems while building this EA..you can let me know.
