EA Factory | Page 25 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Calendar as Fundamental Source at Forex Factory 14 replies

Forex Factory Store 0 replies

Journal Factory posters - how much money are you making? 17 replies

Newbie Factory Journal 12 replies

Attachments: EA Factory
Exit Attachments

EA Factory

  • #481
  • Nov 4, 2006 7:01am Nov 4, 2006 7:01am
Thanks Micardo. I think it should be easy to code these strategies.

I am pretty tightened up this weekend. Are there anyone interested in coding them? Don't worry Micardo, I will find the time to do it if nobody responds.


The way I would do it:
1. Since each lot has their own TP SL, I would place multiple single-lot orders and set the TP/SL in the beginning so I don't have to maintain them.

2. Combine two strategies into 1 by using the following input parameter list:

Timetotrade = 0 (GMT), or 8 (London Open)
NumofOrders = 2 or 3
StraddleDistance = 50 or 20
TP1, TP2, TP3 (take profit)
Reentryifnoprofit = true/false (if false, always reentry when price fall into straddle distance and order was closed)

Please let me know if you are planning to do it. Thanks!





Quoting Micardo
OK guys here they are..... the first 2 strategies.... Is it possible to do the back testing over years before you write the EA please?

Could you guys post the results here please once you have them.

Thanks people I really appreciate your help!!!

How long do you think it will take by the way?
Target: Trading naked.
  • #482
  • Nov 4, 2006 8:20am Nov 4, 2006 8:20am
Hey Aha,

I sort understand what you are saying but the issue is that the stop loss only gets moved to break even once certain TP has been hit. So all stop losses start a the same point for both strategies until TP targets ahve been reached.

Also these are 2 different strategies for different phsycological mind sets so im not sure combining them wll be right.

Rentry this sounds correct to me....

Thanks Aha
  • #483
  • Nov 4, 2006 8:21am Nov 4, 2006 8:21am
Howard this is correct... these strategies are for GBP/USD only.
  • #484
  • Nov 4, 2006 9:34am Nov 4, 2006 9:34am
I wasn't accurate by saying combining two strategies into one. What I actually meant was it's possible to code one EA for two strategies. By using different settings the EA can work for either strategy. That will save a little bit work I guess.

BTW, I will code it in the next few hours.



Quoting Micardo
Hey Aha,

I sort understand what you are saying but the issue is that the stop loss only gets moved to break even once certain TP has been hit. So all stop losses start a the same point for both strategies until TP targets ahve been reached.

Also these are 2 different strategies for different phsycological mind sets so im not sure combining them wll be right.

Rentry this sounds correct to me....

Thanks Aha
Target: Trading naked.
  • #485
  • Nov 4, 2006 1:34pm Nov 4, 2006 1:34pm
Quoting Micardo
Howard this is correct... these strategies are for GBP/USD only.
I do not believe in backtesting using strategy tester at all, therefore I manually backtested strategy 1 for the month of October on GBP the result is as follows:

Number of trades 22
No trade days 3
Days with 2 entries 2

Total pips for 2 lots 263

The result is very subjective as there are a number of days when the price moves right on the entry tic and retreats; in such cases the difference between being filled or not filled is saving a stop of 100 pips which is dependent on which broker one is using, as there is always a slight difference between the data feeds, it really needs forward testing on mini or micro lots on live account.

Strategy 2 is very hard to backtest manually, again it needs forward testing.

howard
Regards
  • #486
  • Nov 5, 2006 5:34am Nov 5, 2006 5:34am
I'll be gone from this thread for about a month--i'm driving across the US on a road trip.

I will, however, look into TD REI when I can..if someone else can pick that up, it would be great.

I've a number of other ideas in terms of backtesting and coming up with a good money management strategy for use with TD combo and TD seq...i'll get to that as soon as I can. (it involves some basic statistics and looking at the price after a perfected 9 or on the 13's for td combo and td seq).

phoenix


p.s. any recommendations for sites to see, or place to go or things to do in the US?
  • #487
  • Nov 5, 2006 7:11am Nov 5, 2006 7:11am
Hi Micardo,

The EA is finished but both strategies are not profitable. I guess your manual backtesting was conducted on daily timeframe. If you go down to hourly, you will see a lot of times limit order was triggered but price pulled back to stop it out before it takes off. Even though the limit order was placed and triggered again, the gain hardly matches the loss.

aha
Attached File(s)
File Type: mq4 Daily Range Strategy by Micardo.mq4   10 KB | 583 downloads
Target: Trading naked.
  • #488
  • Nov 5, 2006 7:15am Nov 5, 2006 7:15am
Nice man, I hope I can get away for a while. Have a great trip!



Quoting phoenix
I'll be gone from this thread for about a month--i'm driving across the US on a road trip.

I will, however, look into TD REI when I can..if someone else can pick that up, it would be great.

I've a number of other ideas in terms of backtesting and coming up with a good money management strategy for use with TD combo and TD seq...i'll get to that as soon as I can. (it involves some basic statistics and looking at the price after a perfected 9 or on the 13's for td combo and td seq).

phoenix


p.s. any recommendations for sites to see, or place to go or things to do in the US?
Target: Trading naked.
  • #489
  • Edited 8:33am Nov 5, 2006 7:23am | Edited 8:33am
Hi Howard, thanks for sharing your work here.

About your comments on backtesting, I understand why you, like many other traders, don't trust EA backtesing: bad experience with bad EAs and, lack of knowledge on how EA backtesting works -- no offence, everybody has their strength and weakness.

If you are from where I am from, it would be clear to you that: if the EA is reasonably well designed and tested, you have no reason not to trust the backtesting result. In fact, a quality EA probably makes way less mistakes than human and works way faster, and when it does make mistakes, it will do it every time so it's easier to correct it.

Micardo's method is an good example. In my EA testing for GBP on Oct. the strategy 1 is not profitable. You will agree with me if you zoom in to 1H time frame and take trades at the opening of the day and see how it plays out. But honestly that would be too much work for human.

Mouteki is another example. In Sept. Mouteki system was highly successful and the forum was crazy about it --- voted #1 as the most profitable system in FF. At the time my EA backtesting result showed that it 's not even profitable if I went back just for one year. So I mentioned gently that it needs filters to complete the system, but was largely neglected because there were manual testing results posted every day showing amazing success. I knew those results were either from too short a period -- you can only go that far if you are manually backtesting it, or were biased by human emotion and/or negligence -- you can't expect a tester doing those boring backtesting rountines makes no mistake. Now one months later the flaws begin to show and people begin to wonder if mouteki is profitable at all.

Bottomline is, a good EA is like a good human tester, but better and faster; a bad EA is like a bad human tester, but more stupid and ridiculous.

Hope this will change your view on backtesting.

aha



Quoting howard
I do not believe in backtesting using strategy tester at all, therefore I manually backtested strategy 1 for the month of October on GBP the result is as follows:

Number of trades 22
No trade days 3
Days with 2 entries 2

Total pips for 2 lots 263

The result is very subjective as there are a number of days when the price moves right on the entry tic and retreats; in such cases the difference between being filled or not filled is saving a stop of 100 pips which is dependent on which broker one is using, as there is always a slight difference between the data feeds, it really needs forward testing on mini or micro lots on live account.

Strategy 2 is very hard to backtest manually, again it needs forward testing.

howard
Target: Trading naked.
  • #490
  • Nov 5, 2006 10:02am Nov 5, 2006 10:02am
AHA.... are you sure, the last 2 months I manually back tested on a 15 minute chart.

There must be something lost in translation here because it was 100% very profitable based on a 15 minute chart and the last 2 weeks of forward testing.
  • #491
  • Nov 5, 2006 10:08am Nov 5, 2006 10:08am
AHA, im trying to read the code and it appears that on strategy one you have 2 lots but 3 take profits... is that right?
  • #492
  • Nov 5, 2006 10:12am Nov 5, 2006 10:12am
Hi Micardo,

Do you use MSN or Skype? We may go through the trades the EA took to see if I was missing anything.

aha
Target: Trading naked.
  • #493
  • Nov 5, 2006 10:13am Nov 5, 2006 10:13am
yes i have skype... what is you skype name?
  • #494
  • Nov 5, 2006 12:49pm Nov 5, 2006 12:49pm
Just chat with Micardo online, exchange a couple of screenshots. We haven't got enough time to find out why there is a difference between his manual test and the EA backtesting result. Among the trades of EA we checked, we did find that adding spread on the top of TP may occasionally cause the EA missing target and grab less pips. Though the effect is minimum, the attached is a new version with a quickfix of the problem.

If you are interested to test the EA, you may find the system in Micardo's post on last page of this thread. The default settings of the EA is for strategy 1 of the system.
Attached File(s)
File Type: mq4 Daily Range Strategy by Micardo.mq4   10 KB | 685 downloads
Target: Trading naked.
  • #495
  • Nov 5, 2006 8:17pm Nov 5, 2006 8:17pm
i backtested strategy 1 manually for 2 months, it shows 545 pips profit since september 4. Will backtest strategy 2 but i don't know what's the london opening and new yok closing times.. anybody can tell me??
  • #496
  • Nov 5, 2006 9:47pm Nov 5, 2006 9:47pm
London open: 7AM GMT
New York close: 22PM GMT

Are you testing GBPUSD? If you do, would you tell me how many pips the strategy 1 wins between Oct. 11 and Oct. 13? What about Oct. 24 and 25? Those days are the biggest loser in the EA testing.

Thanks.


Quoting u_atatunc
i backtested strategy 1 manually for 2 months, it shows 545 pips profit since september 4. Will backtest strategy 2 but i don't know what's the london opening and new yok closing times.. anybody can tell me??
Target: Trading naked.
  • #497
  • Nov 6, 2006 2:17am Nov 6, 2006 2:17am
Quoting aha
Hi Howard, thanks for sharing your work here.

About your comments on backtesting, I understand why you, like many other traders, don't trust EA backtesing: bad experience with bad EAs and, lack of knowledge on how EA backtesting works -- no offence, everybody has their strength and weakness.

If you are from where I am from, it would be clear to you that: if the EA is reasonably well designed and tested, you have no reason not to trust the backtesting result. In fact, a quality EA probably makes way less mistakes than human and works way faster, and when it does make mistakes, it will do it every time so it's easier to correct it.

Micardo's method is an good example. In my EA testing for GBP on Oct. the strategy 1 is not profitable. You will agree with me if you zoom in to 1H time frame and take trades at the opening of the day and see how it plays out. But honestly that would be too much work for human.

Mouteki is another example. In Sept. Mouteki system was highly successful and the forum was crazy about it --- voted #1 as the most profitable system in FF. At the time my EA backtesting result showed that it 's not even profitable if I went back just for one year. So I mentioned gently that it needs filters to complete the system, but was largely neglected because there were manual testing results posted every day showing amazing success. I knew those results were either from too short a period -- you can only go that far if you are manually backtesting it, or were biased by human emotion and/or negligence -- you can't expect a tester doing those boring backtesting rountines makes no mistake. Now one months later the flaws begin to show and people begin to wonder if mouteki is profitable at all.

Bottomline is, a good EA is like a good human tester, but better and faster; a bad EA is like a bad human tester, but more stupid and ridiculous.

Hope this will change your view on backtesting.

aha
Hello aha

I manually backtested strategy 1 on 15m chart, it is impossible to do it on daily chart

howard
Regards
  • #498
  • Nov 6, 2006 3:00am Nov 6, 2006 3:00am
Quoting u_atatunc
i backtested strategy 1 manually for 2 months, it shows 545 pips profit since september 4. Will backtest strategy 2 but i don't know what's the london opening and new yok closing times.. anybody can tell me??
Hi

Your result is in line with and confirms my manual backtesting of strategy 1 for the month of October which produced 263 pips, we have to wait for the consultation of aha and micardo to see why they have different results

Please post your results for strategy 2 as well. Thanks

howard
Regards
  • #499
  • Nov 6, 2006 3:38am Nov 6, 2006 3:38am
Quoting u_atatunc
i backtested strategy 1 manually for 2 months, it shows 545 pips profit since september 4. Will backtest strategy 2 but i don't know what's the london opening and new yok closing times.. anybody can tell me??
I'm worse than you, u_atatunc. What is strategy 1 and 2? On which page should I look for these?

And for Aha: you're the trigger, man. Hope it will be nuclear reaction. Confuse man, like myself, need this thread. Thx, man.

741
  • #500
  • Nov 6, 2006 4:51am Nov 6, 2006 4:51am
Quoting aha
London open: 7AM GMT
New York close: 22PM GMT

Are you testing GBPUSD? If you do, would you tell me how many pips the strategy 1 wins between Oct. 11 and Oct. 13? What about Oct. 24 and 25? Those days are the biggest loser in the EA testing.

Thanks.
that's true. between 11-13, lost 100 profit 4, between 24-25 lost 100...
0 traders viewing now