Help Needed to Debug Existing EA with proven strategy

I do not know that. You would have to test because I'm programming more than testing.
Maybe a combination with the ideas in Voldemar's thread can help, but it takes a bit more time to understand everything.

Hi Guys,

I have been up to my neck with work and haven't really had the opportunity to try out Swingman's EA properly although I did run it for a short while and I could not get it to trade. Maybe I got the settings wrong. Will try it out properly when I have more free time. In the mean time, please try it out and report any bugs or if it is trading in the way it is suppose to trade. I wish I have a bit more time. Just bear with me while I am stucked with my workload for the next few days.

Cheers guys.

If Nihillist replies with his set files, lol, D

As requested,try and tweak and get the draw down down,its running at 34%

Nice one cheers

I cant help but reiterate that we need to add the stoch filter like version 1.6 had. If the first orders loss was not a factor with this new version the drawdown would be so minimal.

I can't get this EA to run as solid without help from some type of filter to support it.

Hi Verocity,
i can't help but to agree with you.

I agree too. Stoch is needed like Version 1.6.

Totally agree..

What would be a good idea would be to limit how many orders are taken at an individual cycle, seems the drawdown is comming from many orders open, the swap rate alone would erode profit while trades are open long term, just my thoughts, D

Hi guys,

Sorry for the few days of silence as I was too busy with work. Ok, had a bit of time yesterday and spent it back-testing Swingman's EA and other related stuff which I will report now.

Swingman's EA seems to be working properly according to the logic/strategy - at least on back-testing - I have not done much forward testing yet which is crucial to identify any bugs.

After thoroughly back-tested the cut loss feature I must say that it is NOT working. I have to admit defeat here - it just eats too much into the profit and makes it not viable - no matter what settings I tried to tweak with.

So, I have to conclude on the Non-Hybrid mode as being very high-risk and the only proper way to employ it is to use it as part of your trading tools doing one cycle at a time after having analyzed the market prior to that. I personally would not like to move any further on this respect as the EAs (both Tictac-nur & Swingman) is already capable of doing that.

As to the suggestions to implement STOCH & other filters to make it trade safer, I must say that it is a dead end and I would not like to waste time on that. Having dug further into my arsenals of EAs, I discovered that I had gone even further to implement 3 MAs in addition to the STOCH. I called the EA Forex RM Trinity which I will post the EA here for you guys to test (sorry, I won't post the source codes). The MA settings are as follows:

0 - Simple
1 - Exponential
2 - Smoothed
3 - Linear Weighted

The rules are the Fast MA must be above the other two MAs and that the pres-set STOCH level (20 set as default) must have been penetrated and EXITED before it executes a BUY cycle - and vice versa for the SELL cycle, namely, the Fast MA must be below the other two MAs and the pre-set STOCH level (80 set as default) must have been penetrated and EXITED.

Here are the EAs split into a Buyer and a Seller:



Well, you guys can test it and post the results here if they are of any values at all.

Let me move on now to the next part of my report.

Again, having dug deeper into my arsenal of EAs, I may have discovered a very effective way to reduce the incidence of deadlock situations trading in HYBRID MODE.

I must state here that this is the only mode which I would like to see further development here as it is very safe way to trade and I believe can be made very profitable.

Digging through my arsenal of EAs, I had discovered another EA which operates on a very high risk logic/strategy which thrives on high volatility and momentum but very susceptible to ranging market (exactly like our EA here trading in Hybrid Mode). I had tried to implement two filters into the EA which I believe had largely and effectively filtered out most ranging situations. One of the filter is the RSI and the second filter I will discuss in the following post. I ran the EA through 4 back test using exactly the same set of data (M15 on gbpjpy from Aug 2017 to April 2018) and settings and the only variable is turning on and off the respective filters as follows:

Test 1 - BOTH filters turned on/implemented
Test 2 - None turned on - EA ran in its pure form without any filters
Test 3 - Only RSI turned on/implemented and the second filter turned off
Test 4 - Only the second filter turned on and the RSI turned off

Here are the results.

Test 1:




As can be seen, when the two filters applied together, they are OVERLY EFFECTIVE and the EA trades very rarely.

Test 2:



As can be seen, when no filters were employed the EA blew up the account in no time.

Test 3:



As can be seen the RSI when applied alone is not effective at all and the account was blown.

Test 4:



Now, this is the interesting part - by employing the second filter alone had produced a very desirable compromise whereby it had effectively prevented the account from blowing up and at the same time had allowed certain ranging situations to pass through producing a higher profit level.

I will discuss the second filter in my next post.

will continue later, got something urgent to attend to.....

yes i m thinking the same, i guess that s why in the beginning of a cycle with few orders it works fine but later with a lot of orders it fails..
but since we won t have access to mq4 of tictac , is there another way to get around this?

Nice to see you again philip! hope it will reboost the thread!

the test looks promising, let s see how it goes!

Hi guys,

The second filter is not an indicator at all but a simplistic common sense approach which honestly, I am very surprised that no one has actually suggested it before nor had I ever read about taking this simplistic common sense approach anywhere here in forexfactory. I am sure I could not be the only person who had thought about it & come up with this simplistic idea - I am sure someone somewhere must have written about this before but I certainly had not read about it here in forexfactory. As can be seen above that it is very effective in filtering out ranging market conditions (though of course not 100% effective) and the simplistic common sense approach is to simply measure the length of the current bar/candle before the EA commence a trade cycle - the length of the current bar/candle must exceeds certain pips before the EA commence a trade cycle. When the EA is applied to the lower time frames, especially 1M - the length of the current bar literally tells us what is happening RIGHT NOW and it can easily confirms with 100% accuracy whether or not there is an on-going momentum/velocity with that particular pair - let's say we set the parameter to 15 pips in the 1M time frame - for the pair to have travelled 15 pips within one minute is a hell of a momentum/velocity - the parameter set for the above Test 4 was 15 pips in the 15M time frame - and it had effectively filtered out numerous ranging conditions and saved the account from blowing up.

I personally believe the STOCH or RSI could also be implemented with the aforesaid simplistic common sense approach.

So, Swingman, can you do us all a BIG favour???

Could you code the following for us:

1. Code in an option to trade in Hybrid Mode
2. Code in STOCH filter - I think only needs to penetrate the pre-set level, no need to wait for EXIT for the EA to start trade - too restrictive if EXIT is also required
3. Code in the minimum length of current bar/candle as an optional filter with adjustable required length - the EA will immediately commence trading once the pre-set minimum length is reached - there is no need to wait for the close of the current bar/candle - so long as the minimum length is reached, the EA will immediately fires away.

Swingman, as can be seen from my extensive tests above - the above filters works very well and hence, Swingman, I am pretty sure that it will make the EA trading in Hybrid Mode a lot more profitable and in view of how safe it is to trade in hybrid mode and if the above really solves/reduces the deadlock incidence - then there is absolutely no reasons why it should not be as profitable as trading in non-hybrid mode and if that is really the case, then wow, we would have given a truly precious gift to everyone here - a very safe way to trade yet so profitable. So Swingman, what say you???

I sincerely hope you can Swingman!!!

Perfect plan, I look forward to that!

Great to have you back Phillip, I feel the coals stoking as I read the posts, what are your thoughts on an input to limit the number of positions open, it is my thought that a system like this with limited trades can quickly accelerate a small account, seeing EA's that need 5k minimum balance all bit too common, dimes to dollars, what you say?, D

Hi Detector,

My thoughts as to your suggestion of limiting the number of opened orders - it is good to have that option as for those more experienced users, they could set a limit to the opened trades and manually trade it out of a deadlock situation and or even to cut the losses and start afresh. If Swingman don't mind coding that feature in, I am all for it - it is not essential but it is a nice option to have.

As to your thoughts of turning dimes into dollars, this EA can literally do that even in Hybrid Mode if the above filters are implemented and proved to be effective as the EA would not waste so much valuable time getting stucked in deadlock situations. From my observation of the back test No. 4 above where I had only applied the second filter alone, I had observed that at least 70% of the closed trades were made out of one single opened order ie the First Order had hit the TP - when you take into consideration the fact that the first order of that EA which I tested was only 0.01 lot compares to our present EA which First Order is a full 1 lot, the amount of profit generated should be, at least theoretically, far higher than the $26K odd profit generated therein (26% in 6 months).

The potential of this EA to turn dimes into dollars is definitely there - if the filters proves to be effective, you can easily start trading with $100 in a micro account trading 0.01 micro lots and since it is only $100, most of us would not mind losing it all and therefore, we could trade it in a much riskier manner by trading multiple pairs.

I hope Swingman will take up my suggestions above and code in the suggested filters as soon as possible.

Cheers detector!

Hi Philip,
I can try to make the adjustments.

I do not quite understand what else @yalgaar asked:
"From EA perspective is there any difference between running Buy + Sell together as compared to running Hybrid Buy/Sell option?"