Help Needed to Debug Existing EA with proven strategy

@Bullilulli & @detector

It's very good buddies - think, think, analyze, suggest.

I also, related to Martingale, previously recommended to study the system of Voldemar (https://www.forexfactory.com/showthr...8#post10982048)
The results are incredible with my little change (BE version). If someone understands how and why the system works, let me explain...
Unfortunately I have a lot to do and I do not have the time to think how to make the connection between the two rockets ...

Looks like there is still a bug, I have the EA running on GJ&EU, the GJ has opened trades in sequence, however the first trade from the EU was a sell then within seconds opened multiple buy positions so now sitting with 20 open orders, D

The only good thing to come out of the situation was to stop auto trading and close all manually, made a small profit, live testing stopped for now until further notice, anyone else experiencing this ?,D

The best thing to do is to follow the trades with the same settings in the tester and possibly report something wrong has happened.
With general descriptions I can not identify the Bugs.

SW, good idea before continuing, when I get time I'll go through it, in the mean time if anyone is using the EA on demo let the group know, I would not recommend live but if you would like to take the risk it needs monitoring lol, imagine my surprise to see the account maxed out again, phew!!! with that said, the trade volume maxed it out but the account was suffering little loss, I did close out and as stated, cents profit, D

Swingman, we need to stay focus on resolving the core logic/strategy first.

Everyone else - please STOP making any further suggestions on adding in unnecessary features (at least NOT at this moment in time) in particular on any aspects of money management. It is an utter waste of time energy and focus on these utterly meaningless add ons (at this moment in time) when we have NOT yet resolved the deadlock situation to our satisfaction. The EA as it is, is still way from satisfactory although we have made great strives with the implementation of the two filters.

Please, please, please everyone - this is what you can contribute at this moment in time;

1. Forward test the EA and report any bugs (well done to those who had reported the 'multiple orders' bug, I had the same experience); and
2. Suggest new ideas or comment on suggested ideas on how to further reduce the incidence of deadlock situations and or on how to improve the profit level (Not by manipulating money management because it is utterly meaningless if the underlining strategy is not sound).

Please, please, please everyone - refrain from discussing/posting on anything else - once we have resolved the deadlock incidence to our satisfaction, only then we discuss on adding in all these extra features into the EA.

For resolving the deadlock incidence why not reducing the TP dramatically? So every cycle will close faster, hence there will be more cycles daily.

Swingman, let me explain in further details with examples on my idea of a 'one single martingale recovery order' (with slight variation as suggested yesterday - incorporating Detector's suggestion of limiting the number of opened orders) -

Swingman & everyone else, I would like to suggest the following recovery logic to be coded into the EA:

The rationale behind this is to kill off the cancer/root cause of a deadlock situation before it has a chance to develop into a full blown deadlock. The logic & approach is very simple as described below.

Swingman, in order not to make it confusing to you, I will just describe the orders issued by the EA as first order, second order, third order, fourth order etc INSTEAD of my usual manner of describing them as First Order, first subsequent order etc....

To code in the following adjustable parameters:

Recoverymode_enabled - true/false
Recovery_lotsize_multiplier - 3
No_further_orders - true/false

This is the logic to be coded into the EA - when the Recovery Mode is enabled, the EA will implement it by issuing the recovery lot when the pre-set depth had been breached. But the Recovery Mode would kicks in ONLY if and when the order to be issued is an OPPOSITE order of the existing opened order(s), let me give you examples of sequence of opened orders to make it clear:

1. Buy, Buy, Buy - Recovery Mode does NOT kick in even if enabled
2. Sell, Sell, Sell - Recovery Mode does NOT kick in even if enabled
3. Buy, Sell - Recovery Mode immediately kicks in with the second order (ie the second Sell order is THE recovery order) because the second Sell order is OPPOSITE to the first Buy order
4. Sell, Buy - Recovery Mode immediately kicks in with the second order (ie the second Buy order is THE recovery order) because the second Buy order is OPPOSITE to the first Sell order
5. Buy, Buy, Sell - Recovery Mode immediately kicks in with the third order (ie the third Sell order is THE recovery order) because the third Sell order is OPPOSITE to the first & second Buy orders
6. Sell, Sell, Buy - Recovery Mode immediately kicks in with the third order (ie the third Buy order is THE recovery order) because the third Buy order is OPPOSITE to the first & second Sell orders

Before the EA issue the recovery order, it will first calculates the total lotsize of the opened trades and then multiply that total with the pre-set multiplier and immediately issue the Recovery order in that multiplied lot size.

If the parameter of "No_further_orders" had been set to "true" then the EA will stop issuing any further orders after having issued the Recovery Order - if the price moves in the direction of the Recovery Order, the cycle will be closed speedily in profit and hence effectively removes the 'cancer' before it has a chance of developing into a full blown deadlock. However, if it moves the other way, then we will either interfere manually and or let it hit the stop loss (need to set a sensible stop loss).

If the parameter of "No_further_orders" had been set to "false" then in the event of the Recovery Order failing to kill off the 'cancer' (ie price has moved in the opposite direction of the Recovery Order before the TP is hit) then the EA will revert back to the normal preset lotsize and continue to trade in the normal way.

Swingman, I hope the above is clear enough & from my thorough observation of the data collected from backtest on gbpjpy over the period of Jan 2017 to April 2018, the above implementation should resolve at least 70% of the deadlock situation because I had observed that at least 70% of the time that when the price reversed after having earlier triggered an order (in the opposite direction), the reversed price always (at least 70% of the time) moves in the reversed direction for at least another 20 pips - and if we can set the multiplier of the Recovery Order to 4, we only need the reversed price to move in the reversed direction for less than 10 pips in order to kill off the 'cancer/other opened. Swingman, I have no doubts the above implementation will help further reduce the incidence of deadlock situations.

Cheers Swingman.

Hi Manuel - we are already doing that and I have thoroughly tested even up to $5 TP per 1.0 lot traded - it does resolves a lot of the deadlock situation but they are still there - what is more, by setting it so low, the profit level is too low as to be unacceptable to a lot of us (some of course can accept it).

So Manuel, the REAL solution is to resolve the incidence of deadlock situations to a satisfactory level (they can never be totally eliminated) and NOT by manipulating/tweaking the settings & money management (as a lot of you seems to be suggesting).

Hey swingman, the bar range filter still doesn t work, can you fix it?

Yes Swingman, it is vitally important to fix that.

Also time filter please Swingman,very important for back testing.

Agree entirely, the EA needs to just function correctly before moving to any other level, the multiple orders is a must to overcome or there will be no confidence in leaving it to run without supervision, once the account reaches maximum orders open, the deadlock is created, unless in my case the buy orders dominated and the account could only go one way to obtain profit, if it goes the other way there would be no room to open sells to cover and the account would blow, live forward testing on hold for the moment, D

One week ago MetaTrader forum published a Recovery EA.
Who wants can test it.

I started last night the EA on demo (GBPJPY-M5) and the orders ran problem-free.

I just started the ea this morning.

Hi Philip,

No sure if this can be a way out of deadlock situation. Your concern is once a deadlock situation happens, the EA will get stuck in a non-productive period until it really break out of the deadlock (which will eventually happens, theoretically).

Instead now you have 2 or 3 reserve accounts. Once the first account goes into deadlock situation (now you have to set the rule to define a deadlock), you start running the EA on a 2nd reserved account. So while your first account is being held in a deadlock, your second account is starting to generate trades. So this will work if your BELIEF that the first account will eventually unlock in future even if you don't know how long it will be and how much drawdown you will get.

In real world situation for this to work? You have to make sure that you never get a margin close out by the broker, meaning you have to top up your lock account for whatever drawdown you are stuck with until it resolve itself. Don't forget the possibilities of more than 2 deadlock situations. So how many accounts are you prepared to send in?

Hi,
Maybe this is what you are looking for? They have som youtube videos that explain the recovery

https://robotfx.ro/post/135684531817...-auto-recovery

El1000,

Had a quick look at it and - NO, that is NOT what we want - there is no real recovery in that old strategy - its a bloody risky strategy which will wipe out your account in no time during ranging market - here I post my own version of it which was coded for me way back in 2012, in fact, this is precisely the EA I spoke about in my post above (post 432) wherein I had implemented the 'Bar Trigger' filter ie by measuring the length of the current bar in order to determine momentum/velocity. Here are the EAs for you to play with if you want to:

I called it Bar Triggered Break Out Profiteer.





Swingman,

The Recovery EA you posted above looks suspiciously grounded on the same logic/strategy. If it is, then better forget it Swingman, it doesn't work, way too risky. I had also quickly read through the thread you recommended earlier and that I also think is way too complicated. Swingman, believe me, my suggested Recovery Mode above is the simplest and most effective way to reduce the deadlock incidence without adding any risk at all. I sincerely hope you can take it up and code the Recovery Mode in Swingman.

Cheers Swingman!

The results I posted in my post 432 was obtained by back testing the above Bar-triggered BOP EA.