Hello everyone,
I would like to ask something to those who write EA.
How do you decide that an EA is good enough to actually be used in real mode?
When you decide that an EA does not work anymore and then remove it?
Can(and how) you automate that decision?
an example of a “good EA” on backtesting, from 2007.01.01 to the end of 2012…but on 2013 :46:
greetings
I usually make the decision based on the observed largest drawdown in the testing period. I will only invest this largest drawdown amount plus the largest MAE per trade observed in the testing period. I am willing to let the system stay alive as long as it has not lost the amount I have invested in it. When this amount is lost, the system has gone bust.
Ok, got it.
I’m trying to put an average on equity but depending on its lenght, sometimes it is worse.
I would like to turn out with a safety-net that will stop the EA when it starts to perform very bad.
The most important thing remain to have a very good system at the beginning.
The problem with the EA you are looking at is that it looks like it was curve fitted to the 2007 - 2012 price series, that should explain why it completely falls apart on 2013 price series
liftoff,
sorry, that I posted was only a stupid example ( short on RSI overbought with 1% tp and sl, H1 tf).
But really do you think that a good system will work forever? I’m convinced that not. Perhaps I’m wrong.
I dont know about self learning EAs, but a tipical EA with fixed rules is like a
suit for a person that sometimes get fat and then slim down.
Btw at the moment my best result is based on two simple strategies that work together and it don’t show that dramatic DD.
In my opinion, EA s with fixed parameters wont last forever. Their advantage will disappear with time. Disappearing with time means they should have been profitable at some present time in the past. Good to know you have two systems working for you. I am still out testing and pruning through systems, I have found one that I like but I want more. #Greed.
I agree with liftoff. The EA settings won’t stay profitable forever.
You need to be able to adjust it with the market. The only time I’ve seen this kind of thing work is when there is an experienced manual trader behind it keeping a close eye on the market.
