critique my EA | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

Critique my backtesting results of this EA 39 replies

meanVel's Trade Critique Corner 1 reply

Critique my system please 12 replies

Trendline only system- Please critique 14 replies

Beginner requesting plan critique 5 replies

critique my EA

  • #1
  • First Post: Sep 2, 2013 8:34am Sep 2, 2013 8:34am
Hi everyone,
I have recently started to write my own EA and i wanted an opinion from the community.
After months of trials, i think i've finally hit it.
I started running my EA today, (on the 2nd Sept'13) start of the forex trading period.
I am trading on 10 different pairs (CHF, EUR, GBP, JPY & USD).
So far in Day 1, i realized that i'm in the green. Although i know it might just be one of those days that it happens before everything turns downhill.
But my EA executed 42 trades, and i had 41 profitable trades.
Currently i'm in 10 more trades out of which about 8 of them are in the red.
So i hope to hear from you what's on take on the situation.

Btw, based on the risk management every trade i only risk 0.3% of my capital.
And as of current, i have made 6% of my capital.

Would this be an EA that could stand a chance of being profitable on the long run?
Thank You.
  • #2
  • Sep 2, 2013 8:40am Sep 2, 2013 8:40am
Sounds look a good EA, i would study the difference of 41 winning trades, and take those later 10 who went negative
And study them, "what time did it go wrong? what was the volume?" etc..

And figure out how u made 40 profits in a row.. sounds like your EA has alot of great potential
  • #3
  • Sep 2, 2013 8:46am Sep 2, 2013 8:46am
Why not add it here as a Trade Explorer ?
  • #4
  • Sep 2, 2013 8:59am Sep 2, 2013 8:59am
I was shocked that there were 41 profitable trades.
Basically i went with the concept of trailing stop (50 pips), so my trades closed at my modified stop loss.

I will have a run with this EA for this week.
Thereafter, if the profits still keeps coming.... I have something concrete to show.
I might put up my account in the Trade Explorer.

For now i just want to hear it there's any hope for the EA.

Thanks aSTRONAUT and Forex Magic for your comments.
  • #5
  • Sep 2, 2013 9:13am Sep 2, 2013 9:13am
Quoting dranix
I was shocked that there were 41 profitable trades. Basically i went with the concept of trailing stop (50 pips), so my trades closed at my modified stop loss. I will have a run with this EA for this week. Thereafter, if the profits still keeps coming.... I have something concrete to show. I might put up my account in the Trade Explorer. For now i just want to hear it there's any hope for the EA. Thanks aSTRONAUT and Forex Magic for your comments.


Share the logic within the EA and maybe You will get a more accurate answer
  • #6
  • Sep 3, 2013 1:50am Sep 3, 2013 1:50am
Hi dranix,

About your EA:

You say that you did months of trial but what does that mean ? Have you done some backtests on it and if yes, on which time length did you run your backtests ?

What was the maximal drawdown during backtests ?

Have you determined what is the minimum starting capital to run your EA ?

Did you started your EA at different start dates in your time test frame ?



As an example, these are my standards to test any EA I made or I get:

Time test period : 5+ years, from 2008 till now.
Minimum trades to be seen during test period : 2000
I start the EA every 10 days starting 01.01.2008 till now.
Maximum drawdown must not exceed 35%, comfortable zone is 20 to 30%

And I can tell you that I haven't found a lot of EAs that survive these tests !
Simplicity is the Ultimate Sophistication.
  • #7
  • Sep 3, 2013 9:21am Sep 3, 2013 9:21am
Just as amazing Monday was... today was a beating.
A horrible beating. In the red.
But a silver lining is i see a possibility for tweaking my EA.
So Monday's results was 43 Wins and 10 Loses.

Plan to let this EA run it's course this week.
Make changes over the weekend for next week's trial.
  • #8
  • Sep 3, 2013 9:30am Sep 3, 2013 9:30am
Hi broketrader,
Thanks for your sharing.

About your EA:

You say that you did months of trial but what does that mean ? Have you done some backtests on it and if yes, on which time length did you run your backtests ?
I never believe in backtesting. I have always made real-time test like what i'm doing now.
I have a server running my EA and i'm been at it since June this year.

What was the maximal drawdown during backtests ?
No definite drawdown as of yet cos whenever i see the EA doing badly, i'll get back to the drawing board and make my changes.
*Like what i'm about to do now.*
Don't know the drawdown, but in total the loss at a specific moment would be the amount of my stoploss * pairs traded which is about 3% of capital.

Have you determined what is the minimum starting capital to run your EA ?
I always start with $1000.

Did you started your EA at different start dates in your time test frame ?
The first minute of the trading week.
  • #9
  • Sep 3, 2013 1:14pm Sep 3, 2013 1:14pm
Hi Dranix,

When developing/evaluating a EA one needs both backtest and live test. Without a backtest one has nothing. With a backtest one may still have nothing. Only live testing results matching backtest results justify putting an EA trading any decent amount of cash.

A backtest must always be done for two reasons: develop the strategy, and evaluate it's worth. Most recent 2 years of data are used to build/optimize the strategy. Rest of the data (say 3 years) is used to determine profitability. Sample size must always be 500 trades plus.

Live test must be done with real cash (small amount ofc), never demo. You only run a demo to pick up code errors after backtest checked OK. EA must be trading exactly the way it is supposed to before going live.
"To hold, you must first open your hand. Let go." - Lao Tzu
  • #10
  • Sep 4, 2013 1:00am Sep 4, 2013 1:00am
Quoting vox dei
Hi Dranix, When developing/evaluating a EA one needs both backtest and live test. Without a backtest one has nothing. With a backtest one may still have nothing. Only live testing results matching backtest results justify putting an EA trading any decent amount of cash. A backtest must always be done for two reasons: develop the strategy, and evaluate it's worth. Most recent 2 years of data are used to build/optimize the strategy. Rest of the data (say 3 years) is used to determine profitability. Sample size must always be 500 trades plus. Live...
Hi vox dei,
A very insightful explanation. Thank you for explaining the ideal methods of testing EAs.
I have tried backtesting but however the results are almost very right. Maybe it's because of the way i configure the backtest parameters.
I shall heed your advice and take that approach.
Thank you.
  • #11
  • Sep 4, 2013 2:20am Sep 4, 2013 2:20am
Hi again Dranix,

You are welcome. Please note that one should never build strategy/optimize parameters using all data available. Data should always be broken into 2 sets. Strategy is built and optimized using the most recent data set. Then the backtest is done using the oldest data set. This way one guarantees there is no curve-fitting, gets a clean measure of past performance, and the parameters that best reflect recent market conditions.

A simple way to get a good feel of how robust backtest results are is to see the range in which one can vary parameters keeping results constant. If the parameters can be varied in a wide range and results are basically unchanged, the backtest should be OK. If minor variations in parameters produce drastically different results, then the backtest is not reliable. I hope your backtest is OK.
"To hold, you must first open your hand. Let go." - Lao Tzu
  • #12
  • Sep 6, 2013 9:46pm Sep 6, 2013 9:46pm
Quoting vox dei
Hi again Dranix, You are welcome. Please note that one should never build strategy/optimize parameters using all data available. Data should always be broken into 2 sets. Strategy is built and optimized using the most recent data set. Then the backtest is done using the oldest data set. This way one guarantees there is no curve-fitting, gets a clean measure of past performance, and the parameters that best reflect recent market conditions. A simple way to get a good feel of how robust backtest results are is to see the range in which one can vary...
Hi vox dei,
Thanks for your inputs and support.
  • #13
  • Sep 6, 2013 9:48pm Sep 6, 2013 9:48pm
Looking at the EA, it seems like it's performing "OK".
Managing to get 60+% winning trades.
The reason for the overall monetary loss is due to stop loss as i've not coded in the logic to close the trade when the market goes in the other direction.
Working on it over the weekend.
Have abandoned the account and will start a new account on 9th Sept'13.
  • #14
  • Sep 8, 2013 5:57pm Sep 8, 2013 5:57pm
Quoting dranix
Looking at the EA, it seems like it's performing "OK". Managing to get 60+% winning trades. The reason for the overall monetary loss is due to stop loss as i've not coded in the logic to close the trade when the market goes in the other direction. Working on it over the weekend. Have abandoned the account and will start a new account on 9th Sept'13.
Coding an exit logic is a great idea! Often offers better average R:R than getting full stop.

Overall looking promising
"To hold, you must first open your hand. Let go." - Lao Tzu
  • #15
  • Sep 8, 2013 7:19pm Sep 8, 2013 7:19pm
I'm curious to know how you do.
Trading in the zone
  • #16
  • Sep 8, 2013 8:52pm Sep 8, 2013 8:52pm
  •  denozi
  • | Joined Aug 2012 | Status: trade with Patient & Discipline | 82 Posts
Hi,
I am very eager to read more about your EA.
I am a new 2 Forex- Your help would be greatly appreciated
  • #17
  • Sep 8, 2013 8:57pm Sep 8, 2013 8:57pm
Quoting hao
I'm curious to know how you do.
Same here, but without the logic, or trade explorer, we might judge it blindly. As i learn to code MQ4, there are basically 4 sub area in EA should have: (a) Money Management, (b) Logic to determine entry, target (c) Handler for open position - if profit what to do, if loss what to do, (d) Summary display, and emergency exit.
Cannot do any judgement on your EA yet danix.. sorry.
Don't blame the tools. Blame the carpenter.
  • #18
  • Sep 9, 2013 10:28pm Sep 9, 2013 10:28pm
Hi everyone,
Didn't know i'll get such a response/interest in my EA.
Did a little coding over the weekend and think i made a boo boo with my exit strategy.
I'm exiting my trades on the slightest of volatility which is the reason for the losses now.
Will revisit my coding sometime today.
  • #19
  • Last Post: Sep 10, 2013 1:15am Sep 10, 2013 1:15am
Set up several EAs with different characteristics and scenarious and forward test them. You can run more then 1 MT4 demo if you know what i mean. You can do some parallel testing with same MM and different weights on it. You can buy a additional computer or RAMS to play as many MT4 as you wish..

You should add comments for each EA in the code to get a statistical data later in, when you are in a finishing pahse of testing. You could start with a phase one of 1:1 RR to get the best winners and try to build on them.

Definately try backtesting, but im fully recommending forward testing.
0 traders viewing now