Great EA in backtest! - page 119
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
here some backtest i do for GBP/USD, what unexpected profit from this EA, now i do forward test for this EA, on eur/usd and gbp/usd using 1H tf. i use realtrader broker,
you could try to improve that model quality... READ the posts!
model quality < 90% = useless...
you could try to improve that model quality... READ the posts! model quality < 90% = useless...
ooo i see, how to improve model quality tp >90%? sorry for stupid question, i really don't know and i'am still new with forex and mt4..
see post #1181 on improving modelling quality..
Nobody replies
I'm not sure why they call these things a forum if nobody ever has the decency to reply to a post.
FxNorth
I ran this with default settings on backtest on 1 hour timeframe (is that correct?) and it came up with 0 trades. What the?!
I am new to this system, would somebody be kind enough to explain it? Also as far as I can see, the default settings are running with no stop at all? Is this not trader suicide?
Thanks Guys
FxNorthThe reason make no trades depends
initial deposit made..
initial lot size..
take a look at the jounal tab at the strategy tester screen if there are any error it will show up on that
more information about cyberia also the cyberia's manual, could be found at the authors website already posted here..
this EA can use for forward test??? i try it but does open any trade yet, i here that open source CT cannot use in live account/forward test.
Makes sense to me,
Thanks for sharing your progress!
Forward testing is not going great for me, it seems like my kids close the lid on my laptop when I go to work, and it shuts it off, guess I'll disable that. Can't wait till I move to the new house, omg.
I am not answering any more 'what are my settings' questions. I think if people really wanted to know my settings they can look where I have posted and find them. If I were to reply to every reader who doesn't have enough drive to do at least that much independent research to find settings; it would bleed me dry. I'm not hiding settings. And there is nothing that says my settings are perfect. People should do their own testing and start to learn to rely on their own findings. As is often pointed out there are more variables to consider like what broker etc. I am now and always have been since I first came on this forum, using IBFX mini account. That I find myself repeating answers to questions like this only tells me that people aren't actually reading the thread. That's lame. Read the thread, do your own due diligence first. Then if that doesn't answer your question ask me. When I see people ask me the same question two or three times after I answered it already directly to them what does that tell me? I'm wasting my breath. This all stops, I'm not playing that game.
I made some further refinements to the CT filter. The settings and all the changes I made which generated this report are included in this posting of the same version of this EA attached to this post. Unlike other changes I have made to this code I have not made the CT filter outfitted with external user inputs for this filter. The changes are in the code comparision values themselves and not set as variables. I changed them by changing the comparision values in the code so the only place it looks different is in the code not in the pop up user settings window. Any and everyone is welcome to make any changes they wish to this. This is not my original EA, I'm just hacking on it to try to improve it myself for myself. By posting my results I'm allowing everyone else to look over my shoulder while I work on it and to colaborate. The burden of making use of anything here rests on the end user who must do so at their own risk.
From this test it appears that I have done something significant to the long positions. I don't recall ever seeing any previous versions of this ea testing with 90% wins on the long positions. Granted it only took 60 long positions over the course of the year to date. That severely limits the activity but it appears that much of that activity which has been filtered was losing anyway, good riddance. I could lightened the filter to allow more activity and with it reduce the percent of wins but I think that's not how I want to play this. I am allowing this to run in my account just like this with risk = 0.05 to see what it does. This test I just posted is not based on looking for a large net profit from the tester. I know I can raise the risk settings and get a huge net profit of the tester. It's to see just how the percentages of wins in longs and shorts is an what the relative drawdown is in the account with the small risk setting.
I'm happy that people want to improve their backtesting modeling quality. I may eventually develop some tick files myself however this ongoing effort to improve backtesting still is not where it's at, for me. I am allowing this to run live forward with small lots because that shows me 100% actual results of how it performs in my live account. I am happy to see that it made +7 pips last night. It only took one short position and won it. But what I'm more happy about is to see that it DIDN'T take three consecutive long losing positions later on after it's win and as the price dropped. This means that at least for this morning it's kept what it's won. That WAS my goal in making the CT filter.
This means that yes it is not going to take as many trades but if it keeps what it wins better that's what matters to me. I'm playing it really conservative with this right now because I hate losses. The only reason I am allowing this back into my account is because my backtest show that both short and long positions are over 80% now. I want to see if I can get those percentages in live account. If and when that appears to be true then I'll consider raising the risk setting to allow larger positions or else I'll conclude that it's not working and remove the ea from the account.
About the dead wood in the code, stuff that isn't being used..I'm not making release versions of this I'm just hacking at it. I leave all that stuff in the code unless it's getting in my own way of understanding what is going on. I do alot of my coding by copy and paste and this is like leaving my working stuff on a clipboard for me to use when I want to make changes. I leave it there so it's there if and when I return to the project to make future improvements. If all the junk I leave in the code bothers anyone they are welcome to clean it out themselves. I keep it there because I'm never sure when I will want to return to a project and work on it some more.
Here's an angle of interest.
Since I have to watch this thing trade and I don't like seeing trades go against the position I thought I'd do a little study of how far trades moved against the position in the backtester.
So over 40 trades I looked at where it opened and how far it went against the position during the time the position was open. Stop Loss is 17 pips so that's the maximum. as I continue this it will be interesting to see where the average and mean is.
1 sell 1 0.0009 9
2 buy 13 0.0005 14
3 buy 3 0.0005 19
4 buy 8 0.0006 25
5 buy 2 0.0006 31
6 sell 9 0.0010 41
7 buy 2 0.0009 50
8 buy 2 0.0006 56
9 buy 2 0.0008 64
10 buy 4 0.0006 70
11 buy 13 0.0007 77
12 sell 1 0.0008 85
13 sell 4 0.0010 95
14 sell 3 0.0013 108
15 sell 17 -0.0017 91
16 buy 2 0.0006 97
17 sell 2 0.0009 106
18 sell 0 0.0009 115
19 sell 1 0.0007 122
20 sell 3 0.0006 128
21 sell 5 0.0011 139
22 sell 0 0.0007 146
23 sell 4 0.0010 156
24 sell 4 0.0007 163
25 sell 4 0.0010 173
26 sell 4 0.0009 182
27 sell 3 0.0010 192
28 sell 1 0.0009 201
29 sell 17 -0.0017 184
30 sell 15 0.0007 191
31 sell 0 0.0012 203
32 sell 0 0.0009 212
33 sell 1 0.0009 221
34 sell 0 0.0007 228
35 sell 9 0.0012 240
36 sell 9 0.0012 252
37 sell 4 0.0008 260
38 sell 6 0.0011 271
39 sell 12 0.0009 280
40 sell 5 0.0008 288
what this shows me is that I shouldn't be getting too worked up even if it goes as far as 15 pips against the position it can recover and more often than not it does. This also shows that it makes approx 288 pips in 14 days. That's how long it took to make 40 trades. On average that's 20 pips a day and that's descent. Actually there were 285 days from Jan 1,2006 to Nov. 9 2006 and in that range the tester returned +4635 pips on balance. that is 16.26 pips per day.
Today's score. won two trades totaling 15 pips and lost one trade for 17 pips net -2 pips. oy
I'm pulling the risk back to 0.01 again. Doesn't look like steam yet to me.