stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

AGT's EA Portfolio
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=1374		 Page 6 of 60
Author:  AGT [ Fri Dec 06, 2013 3:15 pm ]
Post subject:  Re: AGT's EA Portfolio
harry wrote:..... which one is the best one and where can if find it...
Harry
Harry, dude - you understand of course: we all here (and elsewhere in this world) want to find the HG (=Holi Gral) :D :mrgreen:

As I know no one fond this thingy so far ... :mrgreen: :mrgreen:

If you will find at some day the HG - plz plz contact me - I'll pay you (like anyone here) a big amount of bucks to get it ... :lol:
Author:  AGT [ Sat Dec 14, 2013 8:52 pm ]
Post subject:  Re: AGT's EA Portfolio
So the year is coming to his end – I think the next week will be last in this year with considerable market activity.

I’ll analyze below my 3 accounts with MyFXBook:

The “MANDL2007” on the VPS include the XMeters.
All in all at this moment only one bot is doing good:
https://www.myfxbook.com/portfolio/xmet ... low/747867

All the others are in DD’s. I think I’ll switch off (again) the bot using UrsusBovarius settings – it its really impossible to use the coverage – the DDs are to high or the accs even crash quickly. Also the Greta NSNA isn’t doing a good job for me.
So: the results are so far “mixed”. Will let them run furthermore – lets see. :?


The “TGTFX”:

It’s time to abandon the CF4D hopfi well coded in the summer. Although it seems to be able for surviving so far – the DDs are high enough: cant see any advantage for the bot.
On this acc only the NewCF7SR is running on Cowboy Go Markets: Timeframes M15 and H1. And it’s very good documented that the M15 is really much better then H1 – so as I stated earlier: there is no chance to use the higher TF’s. BUT: Cave – using now different settings it may be only a question of any other setfiles and this bot could perhaps work on higher TF’s ! Anyway atm this EA is doing quite well.

The “TGTFX1”:

Here I have namely the NewCF7SR test running. 2 weeks already I use some different settings and 5 k starting capital + 0.01 lot (GP 2-10, GlobalPrime) !
The bot (see GP 23) is surviving with settings “OOTB” (not the best as I know now) since summer using only 2 k starting capital ! :cheer:
Soon it will be already 6 months from starting trading with this bot. :)
I’m pretty curious about the behavior of all the NewCF7SR-bots running now with 5k and changed settings (see GP 2-10).
The NewCF7SR on TF M5 is to choppy, and TF M30 seems to be not running with optimal settings !

The other two: XM-NewRevoV2 or XM Evo v10 (hedging switched ON) are running for a longer period in deeper DD’s – don’t know when they will recover, but they will surely.

:xm:
Author:  AGT [ Fri Dec 27, 2013 10:11 pm ]
Post subject:  Re: AGT's EA Portfolio
So it seems I had no luck with the latest XMeters. Its a pity ! :(
Using them OOTB gave only bad or disappointing results for me. So its time to abandon them.
May be George (=cuz) is right manual interfering the trading of the XM bots - in any case his results show that he is right !
In fact, for me EA trading means that the bot has to do his job fully automatically - otherwise its better not to use it. Or wait one of the newer creations Jose wil code sooner or later. :)
So now only the "NewCF7SR" is left for me - at the moment. Hope in the new year more and succesful EAs will come out.
Author:  AGT [ Thu Jan 02, 2014 7:13 pm ]
Post subject:  Re: AGT's EA Portfolio
So I had to stop my XMeter attempts using some of the latest versions Jose coded.
Had no luck using them OOTB.
BUT there is one left doing really well for nearly 2 months already:
http://www.myfxbook.com/members/MANDL20 ... low/747867
This baby is a bit older. Have no clue why its doing well (with Alpari ;) ). The only difference to the 2 others (Cowboy IBFX+GP) is: the bot with Alpari demo acc is trading 0.1 lot - not 0.01 lot as I planned. That happened only by chance: the Alpari demo is running only with minilots - so the platform itself started the trading with 0.1 lot ignoring my original setting of 0.01 lot.
So is all this only a random result bcz of the famous: "different brokers - different data sheets - different results" ? I think there is no marging filter or something like that in the code ?
This baby provides since start only a few (and good !) trades and is all the time in a pretty small DD ! Using Cowboy IBFX + GP AND 0.01 lot I have quite contrarian results. :? :uff:

OMG - got it ! Of course there is a margin filter coded in !
"MaxUsedMargin=5" did it - that's the reason for limitation of the trades opened ! :uff:
Author:  AGT [ Sun Jan 05, 2014 8:03 pm ]
Post subject:  Re: AGT's EA Portfolio
:hi:
Ok – starting now into the new trading year beginning with the first true trading week of the year lets make a little resume.
ALL BOTs ARE RUNNING WITHOUT ANY INTERVENTION AT ANY TIME ! ;)
1. So the only one XMeter left is the “Evo v10”: I let it run since without any interventions like I did already since nov 09. No changes planned.
http://www.myfxbook.com/members/MANDL20 ... low/747867
2. The main bot I use now (and already for a longer time !) is the “NewCF7SR” coded by KingHigh. I’m trying this EA using various settings. It seems I’m now on the right way of finding appropriate setfiles for GP and GO. All in all this bot seems to be quite stable in the most various market conditions running it now since summer 2013.
So here is an example:
http://www.myfxbook.com/members/TGTFX1/ ... acc/642659
(it looks not so good at the moment - but firstly I use only 2k capital AND the settings arn't optimal (as I know now) - I use now better setfiles - so I hope this bot will do his job better in the nearest furture and the acc will recover after this hard DD !)
Demos setted up some time ago only with optimised setfies are running much better (so far):
Here are the actual favorites on GP and GO accordingly:
http://www.myfxbook.com/members/TGTFX1/ ... ngs/768916
http://www.myfxbook.com/members/TGT_FX/ ... -11/724801
3. At this Monday I’m going to start a demo with the “Monday Zone Breakout EA” – looks like this one could be a good candidate based on a simple and manageable strategy.
4. Furthermore planning starts with the “BASA” and “Marylin” – but not sure so far how and where to start with. A lot of possibilities are described in the relevant threads. But will turn them on in any case in this year !
So that’s all – nothing more so far is planned for the next weeks, months ... this year :D at my side.
Soon I will delete all the older - turned off - demo accounts in my MyFXBook links.
Besides that – let’s see what new will come up perhaps in this wonderful forum sooner or later ! :yahoo: :clap: :cheer:
Author:  milanese [ Sun Jan 05, 2014 10:39 pm ]
Post subject:  Re: AGT's EA Portfolio
:hi:

Thanks for all your hard work AGT ...much appreciated, I think your testing is for all coders and users here very interesting and useful :yahoo:

Cheers :)

Tommaso
Author:  brinol [ Mon Jan 06, 2014 12:08 am ]
Post subject:  Re: AGT's EA Portfolio
Ditto Tommaso AGT many thanks.. :hi:
Brian
Author:  AGT [ Fri Jan 10, 2014 11:45 pm ]
Post subject:  Re: AGT's EA Portfolio
Ok a short comment.
So what do we have after this trading week and after the first NFP in this year ?
Firstly: Its a pity, but it seems even for the last of my XMeters (evolution v10, acc "MANDL2007") I have to accept the unavoidable fate of going down sooner or later. The DD today went now down up to 30 % ! ... even in possible case of recovering its no more a candidate ... :uff:

BUT: this NewCF7SR thingy is toling hard + nice - it is doing its job unfazed, that means it's affected only a little by all these events like news, choppy christmas movements and so on ... all that gives me some hope ! Even the oldest acc's working with this bot (running smaller money - 2k and unappropriate settings) are still alive - although in deep DD atm.
AND: Perhaps the idea of using higher TF's is back in the game. The point is: other settings are needed, creating a wider grid with a small number of trades only.

:!!:
Author:  AGT [ Sat Feb 01, 2014 2:07 pm ]
Post subject:  Re: AGT's EA Portfolio
Its a pity - but it seems the demos in "TGTFX1" are now finally switched off ! :cry:
But I had to expect that - sooner or later they must have switch them off.
So all in all thats not surprsingly - bcz these demos have been running since summer 2013 on the old GP demo server officially switched off already months ago - but dont know by which reason until yesterday these oldest demos have been up :) And so I could watch the trading of the NewCF7SR bot thanks god for some time longer.
Anyway its over - but I could get some good results about the further use of this bot.
Have to set up now new demos of course on the new GP demo server. Did that already - preparing the new upcoming weeks with this EA ! Hope they will be successful !

Got some very simple insights from these oldest demos:
- Dont use it OOTB !
- I prefer "CalcPipStep=false" and then:
- Use a "PipStep" al least 60 (40 wasnt good) - or more ... 60-80 (100 or even 120 ?)
- Set the BsktTsFactor to 6 or 8 (maybe 10 ?) - low "BsktTsFactor" will considerable lower the gains but is not as much effective as expected lowering the DD - to my view at least. Of course maybe in market storms it would be better to have a "BsktTsFactor" of only 2 or 3: but who knows ? Its all a question of risk you are willing to take - like every time we are playing in fx !
- Higher TF may be better for trading this bot (H1 or even H4) - altough it is originally coded for M5
- M15 will trade quite often - but the DD's may be higher, M5 may survive (I started a new attempt with M5 for a limited selection of pairs - no AUD, no JPY)
- On H4 you'll have trades very rarely only - the idea behind all that is the expectancy of more valid S&P zones on the higher TF's - so we could fewer but safer countertrades - I hope so :?
- I add a typical setfile - I use that on the Cowboy Go Markets 10 (GOM 10) demo acc, BUT the good news is, its possible to set the parameters in a wider range - that differences wont effect much the trading behaviour of the bot, maybe that is perhaps a sign of something like a "stability" of this EA ? So dont takt the attached setfile as the only one you have to use ! :)
- Trading capital of 3k ... 5k is recommended (using the minimal 0,01 lot) - although using only 2k and not optimal settings (OOTB) a demo (GP 27) could even survive all the time since summer 2013 ! :good: That was the main surprise for me and a sign of "stability" - unfortunately cant say what would be now running this older M5 demo with optimized settings further if I could - would it recover completly ?

AND: Never use with Cowboy Go Markets a combination of the parameters "CalcPipStep=true" + "CalcBsktParams=true" - the bot will open very very much trades killing the margin - that cost me 1k !
As I wrote already above - I prefer on principle "CalcPipStep=false" using a fixed PipStep. But its possible to use "CalcPipStep=true" + CalcBskTParams=false" - that may avoid often trade opening too.
Dont know why - but I had this BS only with GO - never with GP. So if you try any broker with this bot watch out its first steps. "CalcPipStep=true" may be dangerous - which brokers arn't appropriate like GO - dont know, its perhaps a technical issue. :roll:
Author:  AnotherBrian [ Sun Feb 02, 2014 6:23 pm ]
Post subject:  Re: AGT's EA Portfolio
Fantastic post AGT!. Glad to see your still testing and tracking.

Your points on PipStep are well taken - I agree. My approach was to use set CalcPipStep=true and add code to set a minimum pipstep, I set that to 100, so it can be higher than 100. I think 60 is probably better.

Timeframe - I like 5 or 1 but as you said, the zones aren't as good. To get around this, do you think the daily zones should be looked at? So trade on the 5 min but don't enter if to close to daily zones, or some form of logic using daily zones?

Trading times and days - do you regulate the time and the day? I don't open new baskets on Friday. I start when London opens Monday. I trade London and NewYork, then it wont open new trades or baskets but can close trades.

I was thing to NOT use ad time filter but only use the spread filter, this should do almost the same thing but is much more dynamic. This is a detail, the zones are much more important.
All times are UTC Page 6 of 60

Powered by phpBB® Forum Software © phpBB Limited