I've been modifying one of the systems I developed and used for a rather long time.
Some of the results were interesting, for example in many cases the turnover increased by about 2 - 3 times, in almost all cases the DD was reduced to a third, however the win rate was rather low.
Now I have always intended for the system to be a long term investment system, and so I am not concerned with reliable monthly gains (even though overall profit was quite present throughout the year) , but I would like to know what some of the implications of this are.
I'm just asking because different traders/programmers have had different experiences, the only issue that has come up has been periods of very low growth, and this hasn't been a common issue, I'd say 1 in 10 instruments.
Also a recent modification of the system, surrounding market analysis and the systems mode of reaction, has almost completely removed the periods of low growth issue.
I spread stress tested the system (EURUSD), after 7 years of 10 times the spread capital shrunk by 17%, after 7 years of 5 times the spread capital grew by 10%, after 7 years at double the spread capital grew by 30%, after 7 years of normal spread capital grew by 56%.
After 7 years of current (current was slightly lower, 40%, than what I used as the "normal" spread) capital grew by 76%.
There does seem to be an element of spread sensitivity, but it doesn't look like anything extreme, in fact the system is pickier when the spread becomes larger, which is probably what's saving it from larger draw downs in less conducive environments.
There is no recovery mode in this system.
So any suggestions on what may be causing this?
The time frame is D1 though it has some intraday trading attributes, but only for closing trades.
It is hard to give you any kind of advice not knowing the code. :(
If it is so sensitive to the spread (Ask-Bid) the stops and target might be important .. but what to tell you without code?
Or you are trading in flat markets, or ...
Normally I would restrict the markets for this EA, but because of how low the drawdowns are I through it on almost anything to see what makes it bleed.
I can tell that my code has it analysing the market in different timeframes, higher timeframe analysis just determines whether or not I would like to buy or sell in the next 24 hours and lower timeframes are just for the purpose of analysing the strength of movements and the direction of those movements.
With respect to the spread sensitivity it only shows up when the spread is very large, and it's toned down by parameters that have been put in place with respect to spread.
My issue is with the win rate, there is a win rate of 44%, meaning that 56% of the trades are losses. Do systems like this survive long?
It depends - sorry for this radio Yerevan answer. But in case you mean only 44% of the trades win and 56% are looser it might work if the average win of the winners is e.g. more that twice as big as the lost of the looser.