How can I enhance my system? | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Enhance, Enhance, Enhance! 19 replies

My Simple System (Can we automate this system?) 96 replies

In general, do trailing stops enhance or hinder a strategy? 9 replies

Need a programmer to enhance this Buff indi 0 replies

How can I enhance my system?

  • #1
  • First Post: Sep 19, 2011 8:46am Sep 19, 2011 8:46am
Hi fellow programmers and creative traders,

I have been trading foreign exchange for quite some time using the simple technical tool, basically price action with a little bit of oscillators and moving averages. I don’t have a preferred indicator but I am loyal to support and resistance levels and trend lines. They are in the core of my manual trading.

Beyond that, I recently have learnt programming and developed some EAs and experimented with different tools. The best average results I got for an EA is between 2000 to 4000 pips a year for E/U (most of them are the 2000 zone) which I don’t find very appealing for a system which runs for 12 months, especially that sometimes the draw-down can go up to 1000 pips, not more than once a year normally though.

In the basics of these EAs, I use an oscillator or another indicator for triggering trades with the same or a larger TF moving average filter to identify trends. Another thing I use is stops and fixed trailing stops.

My question to you guys, what can you think I might be missing that can enhance the performance of my EAs?

Your invaluable thoughts are highly appreciated.

Happy trading,

ipnee
  • #2
  • Sep 19, 2011 9:39am Sep 19, 2011 9:39am
if you have the same rules for trading you can see that indicators give you an exact location which repeats itself over time. i don't use anymore supp/res because if you're in a position every level in front of you can cause problems. the solution which resolves this issue is atr. if you have the same entry over and over again then you can start making a system with atr. you can measure how much on average does it go in you're favor and how much do you have to risk in order to achieve that. when creating a system you can have 2 mindsets 1 high winrate or 2 high reward/risk . start with 1 and see if it pays of . if not begin with step 2. don't think in pips because they don't mean nothing. what matters is risk/reward and winrate. make sure that the system pays its spread.
  • #3
  • Sep 20, 2011 8:27am Sep 20, 2011 8:27am
That looks like a philosophy rather than a simple missing piece. Maybe I needed to mention what I am expecting to get. When I used stops at first I notice a significant improvement in backtest. The similar effect, I got from using MA filter. But these 2 enhancers did not seem to go along with each other very well. Here is where I come from.

So I’ve been seeking something which can enhance the performance and go along with either or both of these 2.

Quoting deltatrade
if you have the same rules for trading you can see that indicators give you an exact location which repeats itself over time.
I am not sure what you’re trying to say, but I know and agree with the above statement.

Quoting deltatrade
i don't use anymore supp/res because if you're in a position every level in front of you can cause problems.
I think I see what you mean and I have my own approach to S & R. If you want, give me a chart and I’ll explain a useful way to pick them up.

Quoting deltatrade
the solution which resolves this issue is atr. if you have the same entry over and over again then you can start making a system with atr. you can measure how much on average does it go in you're favor and how much do you have to risk in order to achieve that.
I am not sure about what you mean with the same entry, but I have just added an ATR filter to one of the oscillators EAs and it did not alone make it profitable, although it reduced its lose a little bit. Using it with MA filter reduced the profit a little bit. I still need to try out optimising it, although I don’t think it can be something.

Anyway, Thanks a lot for this idea. And I may post my optimisation report later, if it is worth.

Quoting deltatrade
when creating a system you can have 2 mindsets 1 high winrate or 2 high reward/risk . start with 1 and see if it pays of . if not begin with step 2. don't think in pips because they don't mean nothing. what matters is risk/reward and winrate. make sure that the system pays its spread.
As of this moment, I am fine with risk/win rate, especially as I am trading manually basically.

I would love to hear from everyone, guys, everyone who has a nice idea to share.

Happy trading,

ipnee
  • #4
  • Sep 20, 2011 2:03pm Sep 20, 2011 2:03pm
Quoting ipnee

As of this moment, I am fine with risk/win rate, especially as I am trading manually basically.

ipnee

why don't you use manual backtesting? all the best ideas came to me when i see the transactions in front of me. you can start by looking at a window of 40 trades . see what is best for those 40 trades and after that move to another 40 trades window. see what is best for those 40 trades. now link the rules and come up with ideas which make the whole 80 trades profitable. after that step up to even more trades. when the system rules are all in place then you backtest everything you can. when finishing a system the best feeling i had was when i downloaded sp500 hystorical data from 1970's , aplied the system and worked exactly as expected. i wasn't even born then
  • #5
  • Last Post: Sep 26, 2011 5:00am Sep 26, 2011 5:00am
I did not find something worth sharing about the optimisation that I performed. I also created an ADX filter, but it was not better than ATR.

Quoting deltatrade
why don't you use manual backtesting? all the best ideas came to me when i see the transactions in front of me. you can start by looking at a window of 40 trades . see what is best for those 40 trades and after that move to another 40 trades window. see what is best for those 40 trades. now link the rules and come up with ideas which make the whole 80 trades profitable. after that step up to even more trades. when the system rules are all in place then you backtest everything you can. when finishing a system the best feeling i had was when i downloaded...
I have been doing manual backtesting along. Nothing has come up so far other than my own discretionary level and pattern recognition system, shall I say. I know I'm smarter than a computer

I am wondering, is it that there are no ideas out there or people are tightwad about them?!!
0 traders viewing now