collapse

Author Topic: EA Controller  (Read 155086 times)

Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #15 on: March 12, 2017, 04:53:42 AM »
Thanks for alerting me to these possibilities Ian (refer previous few Posts).

At this stage I am going to put all this on the side burner and revert to basics. The Controller is working nicely for me as is, but always room for refinements down the track by following up on some of Ian's suggestions.

And so to my portfolio update - illustrating the power of the EA Controller

I have been running a portfolio of 12 EAs with the EA Controller system; and now have 60 days of comparable history to share.

As shown in image below.

Again, as in a previous Post, look at the % Gain, profit factor etc, not the absolute $values.

My live account (Controlled) is in profit. The demo, trading exactly the same EAs but UNcontrolled, is in loss.

Yes I would like to improve on the monthly gain of 5% on my live account with these 12 EAs. And since my DD on live account is only around 2%, I could at least double the lot sizes on the EAs on my live account and still be trading conservatively.

Now the "glass half empty" people will say that this is just a lousy bunch of EAs since they go into such a DD as shown on the UNcontrolled demo account, so why trade them at all?

The"glass half full" people on the other hand will say, wow, fancy being able to pull off such a good result while trading with such a lousy bunch of EAs!

So what sort of person are you?    :)

« Last Edit: March 12, 2017, 08:47:35 AM by donnaforex »
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Online nwboater

  • DonbonsPrivateGroup
  • Hero Member
  • *
  • Posts: 1620
Re: EA Controller
« Reply #16 on: March 12, 2017, 01:35:19 PM »
The 60 day results look good so far. Since this is the average of 12 EAs I am curious what the results for the worst and the best of them are. Would you mind posting their charts and results?

Thanks,
Rod
EA's (Run on separate accounts): Cyborg, Best Scalper MT4 V. 2.3, PZ Averaging with donbon2's 15 min EU set.

Signal: Hans Trend 2 Beta (On hold until new version)

Darwinex: Portfolio of 12 Darwins

Broker: AxiTrader, Darwinex
VPS's: Amazon EC2
Rebates: HFT Group




Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #17 on: March 13, 2017, 01:43:31 AM »
The 60 day results look good so far. Since this is the average of 12 EAs I am curious what the results for the worst and the best of them are. Would you mind posting their charts and results?

Thanks,
Rod

In Reply #7, I posted the best result so far.

The worst result is shown in image below. Note also how that result could have been better if we had selected SMA 20 from the start instead of SMA 5. Far less On/Off chop. Given my results so far I would be inclined to start with at least SMA 10 when starting a new EA.
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #18 on: March 13, 2017, 01:44:56 AM »

... and here is the image i forgot to load for previous Post!
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #19 on: March 15, 2017, 12:55:14 AM »
A further interesting stat from my testing to date is the frequency of when the EAs are running in profit (closed profit above MA of closed profit) and when they are in DD (closed profit below MA of closed profit) on the demo account (UNcontrolled)

This is indicated by green and red respectively in the image below for 12 EAs for the six week period from 1 Feb 2017. During this period the MA of closed profit setting of the Controller has been moved from 5 to 10 and is now at 20 for all of them.

What we would like to see for the EA Controller system to have maximum benefit, is to have either long horizontal lines of green, or long lines of red. And it is fine if these alternate as I suspect they will over the years.

What would yield bad results is if we frequently chop and change from red to green and red again. That would indicate a very choppy performance curve. And that in turn would likely result in the UNcontrolled account outperforming the Controlled account. EAs that exhibit such behaviour would have to be expelled! Or traded in the portfolio but without Control if the overall choppy performance is still positive. Or perhaps switching to an MA of 50 or higher.
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Offline ianj

  • Full Member
  • ***
  • Posts: 251
Re: EA Controller
« Reply #20 on: March 15, 2017, 01:09:01 AM »
What would yield bad results is if we frequently chop and change from red to green and red again. That would indicate a very choppy performance curve. And that in turn would likely result in the UNcontrolled account outperforming the Controlled account. EAs that exhibit such behaviour would have to be expelled! Or traded in the portfolio but without Control if the overall choppy performance is still positive. Or perhaps switching to an MA of 50 or higher.

I am assuming you have a small "band" around the MA that can be used to protect a little from chopping out (at the cost of catching some more "normal" downside) If the intention is to catch toxic behavior behavior then some choppiness is to be tolerated - I guess you could get some indication of that from a backtest.

Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #21 on: March 15, 2017, 02:03:52 AM »
What would yield bad results is if we frequently chop and change from red to green and red again. That would indicate a very choppy performance curve. And that in turn would likely result in the UNcontrolled account outperforming the Controlled account. EAs that exhibit such behaviour would have to be expelled! Or traded in the portfolio but without Control if the overall choppy performance is still positive. Or perhaps switching to an MA of 50 or higher.

I am assuming you have a small "band" around the MA that can be used to protect a little from chopping out (at the cost of catching some more "normal" downside) If the intention is to catch toxic behavior behavior then some choppiness is to be tolerated - I guess you could get some indication of that from a backtest.

At this stage the Controller makes use of the standard MA features of MT4. So no "band" but we can vary the number of trades to be used to calculate the MA to deal with chop and/or vary the type of MA - SMA, EMA, SMMA, and LWMA.

If we use a line bar of price on the chart to simulate a performance curve (as per image below), we can see the sort of differences we could get using either SMA or SMMA on the Controller. Not the best of examples since hopefully an EA won't perform quite like price! But hopefully it illustrates my point.

Currently, due to a bug in how MT4 calculates EMAs and SMMAs, the Controller can only be set for SMA or LWMA.

However we have found a way to overcome this bug and am now testing the new version which will enable both EMA and SMMA to be selected.

For those interested in the bug and a way around the bug, pls see https://www.mql5.com/en/forum/132647/page2

As you will see at the end of that thread, I would like to be able to add the Hull MA as an option - it is both smoothed and has only small lag.
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #22 on: March 15, 2017, 05:55:05 AM »
What would yield bad results is if we frequently chop and change from red to green and red again. That would indicate a very choppy performance curve. And that in turn would likely result in the UNcontrolled account outperforming the Controlled account. EAs that exhibit such behaviour would have to be expelled! Or traded in the portfolio but without Control if the overall choppy performance is still positive. Or perhaps switching to an MA of 50 or higher.

I am assuming you have a small "band" around the MA that can be used to protect a little from chopping out (at the cost of catching some more "normal" downside) If the intention is to catch toxic behavior behavior then some choppiness is to be tolerated - I guess you could get some indication of that from a backtest.

At this stage the Controller makes use of the standard MA features of MT4. So no "band" but we can vary the number of trades to be used to calculate the MA to deal with chop and/or vary the type of MA - SMA, EMA, SMMA, and LWMA.

If we use a line bar of price on the chart to simulate a performance curve (as per image below), we can see the sort of differences we could get using either SMA or SMMA on the Controller. Not the best of examples since hopefully an EA won't perform quite like price! But hopefully it illustrates my point.

Currently, due to a bug in how MT4 calculates EMAs and SMMAs, the Controller can only be set for SMA or LWMA.

However we have found a way to overcome this bug and am now testing the new version which will enable both EMA and SMMA to be selected.

For those interested in the bug and a way around the bug, pls see https://www.mql5.com/en/forum/132647/page2

As you will see at the end of that thread, I would like to be able to add the Hull MA as an option - it is both smoothed and has only small lag.

Another variation for MA that we can use is the Shift feature. A Shift of 2 for example on a SMA 5 or 10 also allows for a bit of chop but still minimizes the lag associated with MAs because it is a "shorter period", ie 5 or 10 instead of 20. Note that we can do all this testing at any time, visually on the chart. And we can adjust whenever we want. Just becomes part of the weekly checkup.
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Online donnaforex

  • Administrator
  • Hero Member
  • ******
  • Posts: 4267
Re: EA Controller
« Reply #23 on: March 15, 2017, 09:41:06 AM »
Just wanted to note that i had an issue on my account and a couple of the losing USDCHF trades are not the fault of EA controller. All fixed now.

I know the seller has said this is easy to setup, but it is infact tricky imo (when compared to 'normal' EA installations) and needs close watching initially.
Follow DonnaForex on Facebook and Twitter.
 
If you need to contact me, email admin@donnaforex.com or Skype 'donnaforex' or PM via the forum.

Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #24 on: March 18, 2017, 01:52:43 AM »
You can now also follow my own testing of the EA Controller on one of my live accounts.
http://www.myfxbook.com/members/sydap2003/sydap-pro/1675402

As per image below, the EA Controller was introduced on 9 January 2017 when the account was at 16% in profit after a painful DD. It is now at 36% after the introduction of the Controller to control my 12 EAs in this portfolio. Four of these EAs come as a bonus when you buy the EA Controller; these are the ones also being tested by Donna. The others and more will become available over the coming months at $79 for a package of 4, ie only $20 each. All of these can be run in Signal mode which means that a Trade Copier is not required.
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Online reinerh

  • DonbonsPrivateGroup
  • Hero Member
  • *
  • Posts: 2728
Re: EA Controller
« Reply #25 on: March 19, 2017, 03:01:44 PM »

from january 9 on its mighty impressive = your results

this is a very novel approach you came up with. i have yet to fully wrap my head around this, because traditionally one could miss the recovery. but at the same time recovery could fail to arrive and lead to total failure, then of course your approach wins hands down.

also i think it will not work with all eas strategies. so this has to be well thought over.

the ea you run i assume are some sort of trend followers, then your approach makes sense for them.

Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #26 on: March 19, 2017, 11:24:27 PM »

from january 9 on its mighty impressive = your results

this is a very novel approach you came up with. i have yet to fully wrap my head around this, because traditionally one could miss the recovery. but at the same time recovery could fail to arrive and lead to total failure, then of course your approach wins hands down.

also i think it will not work with all eas strategies. so this has to be well thought over.

the ea you run i assume are some sort of trend followers, then your approach makes sense for them.

The 12 EAs are a 50:50 mix of trend followers based on various types and combinations of MA crossovers and trend reversals using CCI and PSAR, all with sound money management. As long as the EA has a SL, ie is non-grid and/or non-martingale, use of the EA Controller makes sense. Because the Controller measures closed profit of trades (ie balance) relative to its MA to determine whether or not trades should be copied to the live account. It does not keep an eye on equity which as we all know can fall off the cliff at any time and wipe out accounts with grid/marty EAs even while balance is nicely in profit and growing.

In Reply ##6734 on thread https://donnaforex.com/index.php?topic=3308.6720 I illustrate how examining an EAs performance curve is probably the easiest way to see whether or not use of the Controller would be beneficial with any particular EA(s).

PS Yes, people are finding it hard to get their head around the concept of the Controller. But when they do their eyes light up as happened at a recent dinner party where I was asked about what I had been doing of late. My lucid spiel that night must have been so good that even my wife was startled for a moment and berated me for not explaining it in that way to her before; and excusing my long absences from the family room stuck on my computer. That has to be a victory!  :)
Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Online reinerh

  • DonbonsPrivateGroup
  • Hero Member
  • *
  • Posts: 2728
Re: EA Controller
« Reply #27 on: March 20, 2017, 12:06:28 AM »
yes i agree, marties grids no good.

since they always win, till the day comes and kaboom.

yes now it does make a lot more sense, thats why i mentioned trend ea before, and its now even clearer.

gee dont have your wive feel abondoned, she might send ya packin :)

glad to see she too is seeing the light.

Offline diyforexskills

  • Verified Vendor
  • Sr. Member
  • ***
  • Posts: 1050
Re: EA Controller
« Reply #28 on: March 21, 2017, 01:11:30 AM »
yes i agree, marties grids no good.

since they always win, till the day comes and kaboom.

yes now it does make a lot more sense, thats why i mentioned trend ea before, and its now even clearer.

gee dont have your wive feel abondoned, she might send ya packin :)

glad to see she too is seeing the light.

Kaboom is a good way to describe martys. I like it!

The "Exceptional EAs' that come as a bonus when you purchase the EA Controller are certainly not martys. They have been coded to work in Signal mode with the EA Controller and a short description and 2016 back test results can be viewed here.
http://exceptionalfx.com/wp-content/uploads/2017/03/Portfolio-1-of-Exceptional-EAs.pdf

Nice performance curves but 2017 will no doubt have different trading dynamics so we should be prepared for some DDs. If and when they do happen, the EA Controller will stop them from trading on your live account, so you can sleep easy. No kabooms with Exceptional EAs!! You will still experience some smallish losses from time to time when trading with the EA Controller but you won't wake up only to find that your account has been wiped out.

Information and product page - http://www.diyforexskills.com/forex-tools/

Live trading accounts
https://www.fxblue.com/users/sydap2003axilive (FXAE Bollinger Band EAs)
https://www.fxblue.com/users/sydapaxilivepro (Magic Wand Trading) Not trading with this at the moment.




Online donbon2

  • Hero Member
  • *****
  • Posts: 1939
Re: EA Controller
« Reply #29 on: March 22, 2017, 09:49:30 PM »
I actually think MG have a place in FX

Personally load up a cent account somewhere and let the thing run -- when doubles your money open another account withdraw half and start another one... so what if you lose an account from time to time .. then use the funds for better long term EAS... once these smaller accounts start producing.

You got to be prepared for blow ups it is actually a good thing rather than keep pumping money into it by trading too large with them.
Broker - IC Markets   VPS - Amazon EC2 + CNS  

EA:
Powerflow NZDCAD
EA Lab MA Crossover

News - Thomson Reuters Eikon
Trade Manager - Visual Trading Console for MT4
Trading Rebates - through CashBackFX
Trade Notification EA - TradeTalk from FX Blue - free
VPS - I recommend Amazon EC2 - if you qualify for a Broker Sponsored one get CNS
PZ Support/Resistance Indicator / Day Trading Indicator


 

browse forum

* Recent Posts

Re: LEGEND by atomico
[Today at 04:18:02 PM]


Re: HotForex by pinoytrader
[Today at 01:01:09 PM]


Re: LEGEND by yshin
[Today at 12:22:37 AM]


Re: LEGEND by atomico
[Today at 12:04:43 AM]


Re: Will cryptocurrencies like Bitcoin ever gain mass adoption? by Humble Trader's Fx
[Yesterday at 10:46:40 PM]


Re: Will cryptocurrencies like Bitcoin ever gain mass adoption? by oportunis
[Yesterday at 10:28:27 PM]


Re: Will cryptocurrencies like Bitcoin ever gain mass adoption? by Humble Trader's Fx
[Yesterday at 09:34:40 PM]


Re: Best Scalper by nwboater
[Yesterday at 03:08:15 PM]


Re: Best Quotes by Eliza Abrams
[Yesterday at 03:03:23 PM]


Re: Best Scalper by Humble Trader's Fx
[Yesterday at 01:12:25 PM]