EA Factory | Page 7 | 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

  • #121
  • Oct 2, 2006 10:10am Oct 2, 2006 10:10am
Agree.... but commenting makes it a little easier ... except bad commenting

Quoting GP2X
Hehe, other's code is always hard to read, same to yours, even with comments.
Target: Trading naked.
  • #122
  • Oct 2, 2006 10:13am Oct 2, 2006 10:13am
Thanks for sharing this here! I learned something there.

aha


Quoting RoBiK
please see my posts in this thread for explanation about the modelling quality

cheers
RoBiK
Target: Trading naked.
  • #123
  • Oct 2, 2006 10:15am Oct 2, 2006 10:15am
Thanks, GirlFlyer.

Would you please take a look at the post #99? I think by analyzing the backtesting chart in the way I did in the post, we may solve the mystery why this EA behaves differently in the backtesting of 1M and other time frames.

aha


Quoting GirlFlyer
Hi Aha and Everybody!

I'm posting an update to my Flyer! EA. This updated version will allow the EA to recover in case of power failure, or if you want to shut the program down, or even simply change the chart. This prevents it from "forgetting" previous orders.

Also, you can run the EA live on any chart timeframe you want. But if you backtest it, Strategy Tester must be set to use the M1 setting. Otherwise erroneous results will occur.

GirlFlyer
Target: Trading naked.
  • #124
  • Oct 2, 2006 10:55am Oct 2, 2006 10:55am
Quoting aha
Thanks, GirlFlyer.

Would you please take a look at the post #99? I think by analyzing the backtesting chart in the way I did in the post, we may solve the mystery why this EA behaves differently in the backtesting of 1M and other time frames.

aha
Aha,

I think I can explain why my trades for the first few days of 9/05 are slightly different. It is apparent your dataset is slightly different than mine. If you look at your chart you posted on #99, you will notice over on 09/06 at 08:30 on your chart, the bottom part of the candle bodies for those two downperiods go below the 1.2445 grid line. On my chart that I posted earlier, the bodies of the candles do not go below the 1.2445 grid line.

So apparently, we are seeing a slight discrepancy in our data and that is probably causing the different trade entries.

I have downloaded my data from Alpari. How about you?

Also, again as far as running the EA on different timeframes instead of M1 in strategy tester, you will not get accurate price data out of Strategy tester because it will guess at the prices during its interpolation process. As we mentioned before, during my live trading today, I have run it on all timeframes and it works fine because the prices come in live.

BTW,

6 out of 8 trades live today are up and with 4 of them +45 pip gains so far as of 12:00 eastern time.

GirlFlyer
  • #125
  • Oct 2, 2006 11:19am Oct 2, 2006 11:19am
Hi guys(girls),

Since I started this thread I have received many requests (in public and in private) from system developers asking for help on developing EAs. I felt really bad that I couldn't help all of them partially because my time is limited, and partially because some of them don't really have a well established system.

For those who have seriously developed some EA before, you will understand that developing an EA is hard. You have to really understand the system, and interprete the system thoroughly, analyze the situations even the creator of the system never thought of before, that really takes a lot of time and energy. Like this Mouteki EA, I have to go through 1200 posts before I start to develop it and still worries that I interprete it wrong. And as you may see from this threads, a tester only need to find one situation that doesn't make sense, but a EA developer have to consider all the possible situations and make sure they have been taken care of. Once you started developing an EA, it's like you have attached part of you into the EA and you can't just simply drop it unless you find out the system is a crap. And at those situations, believe me, it's painful.

So this is not a piece of cake and I really have to use my time wisely in terms of chosing the right system to code. I understand that for the systems that is under development, an EA can greatly accelerate development process. But unless the frame work of the system has been establish and have shown some really good potential in the manually test, I couldn't just throw myself in and easily get out of it later. Please understand me.

As I said before I welcome system developers come to me and discuss with me on the possibility to develop an EA together. But before you do, you at least should have a well-written system description and have manually tested your system for at least 6 monthes strickly following your own rules. You don't have to disclose them to me if you don't feel like to but you have to have it so you are ready to talk with me.

When you come to me you are expecting some quality from me, same here.

Again, I feel bad that I couldn't help all the system developers who come to me. I hope you can find someone else to help you, or better yet, learn to develop it yourself if it justifies the time and energy.

aha
Target: Trading naked.
  • #126
  • Oct 2, 2006 11:50am Oct 2, 2006 11:50am
Hi GirlFlyer,

Yeah, my data was from Alpari too. And yes our data feeds are a little different. I think your explanation makes sense. Because your EA open trades immediately after a setup appears, not waiting for bar close or open, it is sensitive to the data within the bar. I guess for this kind of EA we can only test its performance in forward testing.

To others invloved in this thread:

If GirlFlyer's EA is working as his/her 1M backtesting indicates, this EA might be the most profitable EA I have ever saw. Unfortunately I don't have enough resource to test it on all the pairs because I have a couple of other EAs running and if they enter a trade at the same time they will conflict with each other. I would recommend this EA to anyone who are looking for profitable systems to test it. And when you do, I would appreciate if you can post your results here so we can compare it with GirlFlyer's results.

Thank you.

aha




Quoting GirlFlyer
Aha,

I think I can explain why my trades for the first few days of 9/05 are slightly different. It is apparent your dataset is slightly different than mine. If you look at your chart you posted on #99, you will notice over on 09/06 at 08:30 on your chart, the bottom part of the candle bodies for those two downperiods go below the 1.2445 grid line. On my chart that I posted earlier, the bodies of the candles do not go below the 1.2445 grid line.

So apparently, we are seeing a slight discrepancy in our data and that is probably causing the different trade entries.

I have downloaded my data from Alpari. How about you?

Also, again as far as running the EA on different timeframes instead of M1 in strategy tester, you will not get accurate price data out of Strategy tester because it will guess at the prices during its interpolation process. As we mentioned before, during my live trading today, I have run it on all timeframes and it works fine because the prices come in live.

BTW,

6 out of 8 trades live today are up and with 4 of them +45 pip gains so far as of 12:00 eastern time.

GirlFlyer
Target: Trading naked.
  • #127
  • Oct 2, 2006 12:07pm Oct 2, 2006 12:07pm
Quoting rockrat
could you point me to any help where i can get the mouteki EA to work. I have v2.0 in my experts folder, and have drgged it onto the chart, but all i note as different is a smallc ross in the top right hand corner of the chart

what am i doing wrong?
Hey Rockrat,

You have to enable the EA.....

At the top of your Chart screen, click Tools, then click Options, then click Expert Advisors, then click Enable EA

Hope that works.
  • #128
  • Oct 2, 2006 12:29pm Oct 2, 2006 12:29pm
Just so everyone knows. i am forward testing girlflyers EA. I have it going on these pair:


eur/usd
usd/jpy
gbp/usd
eur/jpy

I started the EA around 12:15 pm central time. I only did .2% of my portfolio even though it suggests 8% but I am going to test it. If it shows gains, i will bump it to 2%.
  • #129
  • Oct 2, 2006 12:56pm Oct 2, 2006 12:56pm
Thank you BMW! But you are not running in on your live account, are you? This is for testing only and we would only go live after 6 monthes of successful demo testing of the system and the EA even we have full confidence. I would apologize if you know what you are doing but please don't try this on your live account yet. It's so easy to get a demo account and we don't need to risk anything on this.

aha



Quoting bmwboyee
Just so everyone knows. i am forward testing girlflyers EA. I have it going on these pair:


eur/usd
usd/jpy
gbp/usd
eur/jpy

I started the EA around 12:15 pm central time. I only did .2% of my portfolio even though it suggests 8% but I am going to test it. If it shows gains, i will bump it to 2%.
Target: Trading naked.
  • #130
  • Oct 2, 2006 1:03pm Oct 2, 2006 1:03pm
Quoting aha
Thank you BMW! But you are not running in on your live account, are you? This is for testing only and we would only go live after 6 monthes of successful demo testing of the system and the EA even we have full confidence. I would apologize if you know what you are doing but please don't try this on your live account yet. It's so easy to get a demo account and we don't need to risk anything on this.

aha
aha, just on my demo. sorry i didnt clarify. I am trying to trade it as i would a live account also. i have a live account but it is not for this. I have two computers so i can help with this a little i hope.
  • #131
  • Oct 2, 2006 1:24pm Oct 2, 2006 1:24pm
Thanks again man, . But you did get me worried.



Quoting bmwboyee
aha, just on my demo. sorry i didnt clarify. I am trying to trade it as i would a live account also. i have a live account but it is not for this. I have two computers so i can help with this a little i hope.
Target: Trading naked.
  • #132
  • Oct 2, 2006 1:43pm Oct 2, 2006 1:43pm
Quoting bmwboyee
Just so everyone knows. i am forward testing girlflyers EA. I have it going on these pair:


eur/usd
usd/jpy
gbp/usd
eur/jpy

I started the EA around 12:15 pm central time. I only did .2% of my portfolio even though it suggests 8% but I am going to test it. If it shows gains, i will bump it to 2%.
I started Girlflyers EA at about the same time on the first 3 pairs on your list. I kept mine at 8%. We can compare our returns down the road.

Have you had any trades yet? I have not.
  • #133
  • Oct 2, 2006 1:50pm Oct 2, 2006 1:50pm
POTP: I am not sure, i am at work right now and can't check it. i will tonight though.
  • #134
  • Oct 2, 2006 2:39pm Oct 2, 2006 2:39pm
Hi Everybody,


Update on my EA. I have been running live demo test since the market open Sunday night. The EA has done 8 trades for me so far since I started. Of those 8, 2 have been stopped out (they were on GBP/USD) and the rest remain open still having not hit the exit signal. Below is a synopsis of the current standings.

For reference: My current time is 19:30GMT

EUR/USD - Open Buy at 09:32 GMT Price - 1.2686 - Current Price 1.2739 - Up +53 pips

USD/CHF - Open Sell at 08:30 GMT Price - 1.2500 - Current Price 1.2431 - Up +69 pips

GBP/USD - Open Buy at 08:35 GMT Price - 1.8739 - Current Price 1.8867 - Up +128 pips

USD/JPY - Open Sell at 13:00 GMT Price - 118.06 - Current Price 117.67 - Up +39 pips

AUD/USD - Open Buy at 13:31 GMT Price - 0.7466 - Current Price 0.7481 - Up +15 pips

USD/CAD - Open Sell at 13:47 GMT Price - 1.1160 - Current Price 1.1161 - Down - 1 pip

Two GBP/USD trades earlier

Open buy at 06:49 GMT, stopped out -50 pip loss at 07:40 GMT
Open sell at 07:40 GMT, stopped out - 50 pip loss at 08:31GMT

These two trades came when GBP went up and down and back up over 50 pips early this morning.

Besides those two trades, the EA has performed quite well so far today. Waiting to see how the exits go.

That's my standings so far.

GirlFlyer
  • #135
  • Oct 2, 2006 3:13pm Oct 2, 2006 3:13pm
Hi Aha, Community,

i have done some backtesting with the latest Mouteki EA 0.2 from 2001 till now (90% quality) - set EA to trade all trendline breaks:
Attached File(s)
File Type: zip Mouteki02_4h_from2001.zip   580 KB | 664 downloads
expect the unexpected
  • #136
  • Oct 3, 2006 12:18am Oct 3, 2006 12:18am
Hi GP2X,

I have read through your EA. Nice work! Your EA seems to win on EURUSD and loss on GBPUSD, in contrary to mine. How about other pairs, have you tested them yet?

Could you explain a little on why you think only taking trades with large profit projects would help?

aha


Quoting GP2X
Aha, I am glad to know that we are trying to do the same thing.
Attached is the EA I am working on, please take a look.
The difference is I filter trades by the profit points. Besides, I only look back certain number of bars to create the trend lines. So it generates much less trades than yours.
Target: Trading naked.
  • #137
  • Oct 3, 2006 7:30am Oct 3, 2006 7:30am
Hi Everybody,


I have updated my EA Flyer! to now include a trailing stop feature.

You can find it in my thread in the main discussion section.

1st days tests live turned out good.

Thanks for everybody's help and support.

GirlFlyer
  • #138
  • Oct 3, 2006 8:32am Oct 3, 2006 8:32am
Not yet, and I don't think it will work in current status.
At the beginning, I got the same problem as yours, too many fake signals with only a few big wins. Then I realized the work would be to filter out those fake ones. So I added two, one is to only look back certain number of bars(someone suggested it here too), another is to only risk for big pips. They improved the performance a little. But I got much less trades which makes it hard to test and prove it will work. Currently I just stop here, wondering if there is any other better filter. One way I haven't tried is, like the BGX, use pips filter to enter later. Let's say the target pips is 100, we enter after it goes to the direction for 30. Just hope someone else could provide better ideas. To oly look at the trendline is not enough. Besides, I am not condident with the way to draw the trendlines in my EA, if it's not correct, nothing would work, that's why I asked you to take a look. Just one method there.
Quoting aha
Hi GP2X,

I have read through your EA. Nice work! Your EA seems to win on EURUSD and loss on GBPUSD, in contrary to mine. How about other pairs, have you tested them yet?

Could you explain a little on why you think only taking trades with large profit projects would help?

aha
  • #139
  • Oct 3, 2006 10:55am Oct 3, 2006 10:55am
Hi Aha, Community,

attached you will find the backtesting results for the Bago EA, default settings, with 90% Modeling quality
Attached File(s)
File Type: zip Bago_EA_1h.zip   232 KB | 600 downloads
expect the unexpected
  • #140
  • Oct 3, 2006 3:59pm Oct 3, 2006 3:59pm
sorry for the dumb question, but could someone explain to me in the mouteki EA, in settings, there is a setting called positive SL, it's set to 10, exactly what does this do, I understand the profit to move sl is set at 40, and will move stop to b/e once we are in profit by 40 pips, is that right?

Anyone that has time to explain the above for me would be great. thanks.
1 trader viewing now