Below we list several possible issues with your software settings and explain how to set them correctly. So if you run into any of these problems listed below then you are probably setting unrealistic expectations or simply use wrong Generator/Reactor configuration.
Problem #1: Setting too small or too big stops.
If you set the SL/TP size to unrealistic numbers you would never find a single EA even with a profit factor of 1.0. For example, and EA on GBPJPY with SL of 20 pips and/or TP of 30 pips would never create you a profitable EA. It’s just not possible and unrealistic. Or you might find a strategy with one of a few trades if historical backtest period is relatively short.
Think of it like trying to fit the 50mm diameter screw into a hole of 25mm diameter. That’s just not possible.
Problem #2: Using too big lot size for the initial account size.
As an example, if you set the initial account size of $1000 and try to create a strategy using 1.0 lots you’ll hardly find anything. It’s because a single trade can blow such account and would never make it to the upside. So it is important to calculate correctly and set the proper lot size (i.e., 0.01 for a $400 account).
Problem #3: Creating strategies with thousands of trades.
In robot trading, it is a commonly known thing that you’ll hardly ever find a strategy with several thousands of trades, even with a backtest of 20 years or so. It’s just unrealistic that any algorithm that’s not re-optimized periodically would work for some many years and product so may trading positions profitably.
It is common that trading robots usually make 1 or a few trades per week on higher timeframes. As an example, we have profitable robots that are working for more than 2 years and the most amount of trades is only 29 for most of them.
Problem #4: Expecting great Profit Factor (or another parameter) to be very high in the initial strategy variation.
It is quite common for traders to expect to find strategies with a high-profit factor, or unrealistically small drawdown, or very high return/drawdown ratio. But it is quite unrealistic because greater strategy result parameters usually are found after optimization. That’s is why it is advisable to try and find strategies with Profit Factor of 1.0 and then try to optimize every strategy to raise the PF to a higher number. That’s where the REACTOR comes very handy.
Problem #5: Setting to strict Acceptance criteria for the Generator, Optimizer, Monte Carlo, or other FRF tools.
It is quite easy to create EAs that pass Acceptance criteria and robustness tests, but if you set too strict requirements then obviously no strategies will pass them.
Imagine if you were looking for a person who can jump 3m high off the ground. You would not find anyone in the world who can do that.
But if you’ll look for people who can jump 2.4 meters from the ground then you’ll find a few people on earth because world record for the highest jump is 2.45 meters.
Now imagine how many more people you would find if you lower this requirement to 1 meter only.
The same with backtests and robustness tests. If you set too strict rules then no strategies will pass them.
The difficult part here is that each trading instrument (currency pair) have its own characteristics, different volatility, different spreads, different average true range, etc. That’s why we usually need different requirements and acceptance criteria when creating strategies for different instruments.
As an example, 50 pips stop loss would be too small for the GBPJPY trading strategy, but it fits quite well for the EURUSD.
Hope this helps and have a great robot trading year!