Trapping System: Tunnel Martingale

Firstly, I am excited to see that the radityo.ardi is alive and posting in forex factory again. Hi there!

Secondly, I agree with all your comments, especially regarding mg, however, I still use my own grid and mg ea, that is still, after 2.5 years, based on rdzgrid. This one: "Tunnel MG, in my opinion, just contains a lot of filters, and filters, imo, are every bit as bad as they are good; eg they filter out just as many of the best signals, as they do the bad. That is why I only tested TMG for few weeks before returning to testing of rdz. I have changed the first 2 trades from pending orders, to instant orders, AND also added options for different styles of trail stop/step, AND a sort of "average tp". My new ea contains only about 5% of your coding now, however, I still consider it as rdz evolved .

I did try to contact you via ff and google plus, and linked in, but only got a hello from you. I hope that you do resume coding of rdz or tunnel mg eas. I will eagerly follow your postings!

Whoaaa.. I really didn't expect anyone from the past to reply, but I do really appreciate it!
Excited to see your reply too, buddy!
Anyway, I'm just updating after a long hiatus. Got responsibilities, so held back whatever I left.

I am still testing my cBot with the server's past data. It seems promising though, but honestly too good to be true.


I need more tests with more data as I got more. The rows on red touched 6 times of floor and ceiling of tunnel 6 times. But it is expected as it was Monday, the most "lazy" day in trading.

So, in principle, this new Tunnel Martingale is nothing new, same as the last time. The only difference is just purely money management. I was thinking the same for other people with another strategy too, even manual "chart guy", or automated ones like myself. So for example:

1. Last time we calculate lot size based on the money we have. Now, just tag to 0.01 always.
2. Trade only once a day, avoid the same execution time, avoid the same pair. 2 trades max is also can, make sure the time is > 3 hrs apart.
3. Avoid holidays.
4. Avoid NFP.

So, the key here is don't be greedy. If you want to be greedy, put onto some other pairs. That screenshot is only for 1 pair, with 1,000 USD, generates around 11 USD on that timeframe, which is 1.1% growth. But you got a lot of pairs as well. Maybe we can push for 500 USD (then be 2.2% growth) starting money, or 250 USD. But that means risky margin level when it goes beyond 4x of touching floor + ceiling of the tunnel. To me, 10% growth per month is a decent profit.

For Grid, I've got no idea how to improve it. I'm so exhausted with ideas. I can throw a lot of grid ideas, but none of them can anticipate the switch between sideways and trending. The fact is, grid requires the trade to be on always (because it's the nature of grid trading, you get profit from a lot of orders, sometimes it goes over a week). When you want it to generate profit during sideways, you'll be stuck when it's at a strong trend (and it happened). When you want it to generate profit during trending, the fact is it's not always trending all the time. So, I guess the Grid, whatever the strategy is, will be difficult and I don't see any future in it. I've seen a lot, like someone doing the "so-called" grid but not it. Hedge all the way till even months. But it's difficult to manage and practically stuck.

I am no longer coding with MetaTrader. The programming language is using C++ which is sucks. I am into cTrader now. Will post more if this TM does better.

very cool; yes I read your previous...previous post regarding C++ and cTrader vs mql4 and MT4, and while i agree with your comments completely, Mt4 and mql4 is still the main platform used for the desktop, so while I will be very curious to follow your coding; I will have to try converting it to mql4 for my own use as cTrader, as good as the plusses it has over mt4, mt4 does not have any negatives that i am aware of that may have any effect or any negative effects on my current trading.

i also agree with your grid comments regarding grid having to be kept in ON state. However the first reason i fell in love with rdzgridtraps was the on screen button and setting to stop trading after the current cycle has closed. I am not sure if that feature was in tunnel ea; been so much time and water under the bridge, I think I may get them mixed up haha.

welcome back my friend!

I agree completely, which is why i added to your rdzgrid, i think i used base code 1.13?? The more recent ones used libraries, which constantly gave errors and didnt open trades when it should, so i went back to i am sure was 1.13.

I added in abilities to limit the tradesizes, which eliminated the CreateOrder errors due to mgs getting out of hand , and also added a 1st and 2nd grid setting, so that the 1st grid trade opened at 150 point, and all other pendings open at 90 points Also... code to check that the open prices were accounting for stoplevels and such, whcih i dont need it anymore as i changed the first 2 trades to instant orders, so the stoplevel check is kinda redundant now, but anyways haha, also a setting to add the grid to the spread, which i dont use, but i have it there as an option that i might use on the higher spread pairs. I also check free margen and equity before attempting a new trade, and if freemargin is below XXX my lotsize for the new trade is limited to a teeny tiny size. This means that it will be a guaranteed loss, but a lot smaller loss than if i didnt open a new trade. And then on days when news events are occuring i always click the button on the chart that stops trading on close of the current cycle, eg rdz wont create a new cycle after the current trades all close. This forces me to check the market after end of the cycle and decide to open a new cycle or finish for the day. So i also have an indicator that opens an alert window when i have 0 trades open, so I know when i have to go back to computer to look at it. This is also setup so that the alert is pushed to my mobile mt4 on my phone.

The reason is there is not many automated traders around. The demand for cTrader is way less than MT4. Not even MT5 can compare with MT4 in terms of number of users. That's where their money they can suck in.

Good luck, bro. I can't easily convert C# to mql, especially when someone is using LINQ for arrays. We can easily convert mql to C# but difficult otherwise. My new code uses LINQ as it's simple dealing with arrays. It's just requires too much of efforts. Anyway, the concept's the same. Just that I open orders based on preset date time from a JSON file, nothing special with the new TM.

No negative effects. Just the "terms & conditions" linked to the account only.

No button can be added in cTrader apparently. But it's not really needed, as I can write OnStop event to close 'em all.

Wooow..that's complex dude. I think the bottom line is grid is profitable, just don't put too much weight on it. That's a good concept, do a trade, call it a day and close it, try win some other time, profitable, then repeat.

Are you still using it till today?

Thanks brotha!

I was using it til June. And non trading life has picked up since then, otherwise I would be still using it on gbpusd right now. Grid being grid , I found it did great so long as I was near the computer at almost all times to baby sit, which I was having too much fun with non trading life, so I have not used it since June, all manual stuff since then, albeit I do grid like manual trading anyways haha, you and rdzgrid has made a very big impression on me.

Maybe I can get or create a mt4 bridge. But am not concerned with that right now, and yeah, no real need for a button if you just have a setting in the ea or cBot setttings to stop after the X cycle has closed.

Hi radityo.ardi,

You can add buttons in cTrader. Check the new features added in 3.6.

Best Regards,
cTrader Team

Well, that all happened over about 18 months, so it was a work in progress for all dat time. I am only disappointed that I wasnt able to share it with you before now and get your extra insights and advice on improving it before now. It isnt on my platforms now as the last couple mt4 updates screwwed wih it on the last live platform i had it on. When i am on hols next am going to test and fix the trouble.

I did my research for mt4 bridge, not good in my opinion. Most of them paid (didn't get a chance to test it out), some other which is free, are mostly using network protocol to exchange information between MQL code and C#. The main problem is latency. While the bridge will take time to pick up the instruction from C# code (coz essentially it's using network protocol which will "listen" to instructions), the execution of the order inside MQL itself will also take time. Note: MetaTrader doesn't allow you open order asynchronously. One instruction to open an order must be finished before you can open another order. In cTrader, you can do simultaneously.
I need that millisecond performance...

Didn't know that buttons natively supported. But hey, it's C#, we can pump up a button anywhere.

Just a stupid idea, after multitude tests. I hope pictures can explain better.


Let it flow wherever it goes...


Then close wherever you want.

The catch here is, you must run this only once at any particular time (usually when the high impact news is released)!

Now, here is the deal. I've spent so much time on writing code playing around with all parameters and stuffs like that. It's like breadcrumb pieces, where I got the fact that:

1. Price is always unpredictable (unless if you want to learn how to predict it, which makes your profitability higher).
2. Price is always moving, especially when the market opens, some news released, or any high-impact stuff.
3. Tunnel Martingale is now capable of setting Ceiling and Floor with Stop Orders (not executed, until the price is reached).

But I kept myself enclosed in a box that:

1. I need to limit my profit and call it a day.
2. I missed a huge movement.
3. It's a martingale, you can't keep it running for so long.
4. Your risk doubles as it progresses forward.

One key thing is, yes, your risk doubles, but your profit movement per pips also doubles.

Just a brief info, backtesting is using:

1. $1000 capital
2. StartingLotSize 1.00
3. Multiplier 2 (each lot size will be multiplied by 2)
4. Tunnel height 200
5. no stop loss

When I do that, I get $108 from perfect 4 trades (you see only one of them below).

So much movement missed.

But when I took off the target, let it run...I get $919 from 2 trades.


For sure this is just backtesting. Interesting enough if I can put forwardtesting in the factor (maybe throw in some real $50 to the test).

Now,
here is what I see that you don't see.

A price for EURUSD always moves, from point A to point B, almost daily basis. Why is that?

1. Because of a lot of people doing the bidding for EURUSD almost daily, on a regular basis, and at the same time.
2. One may do trade during European time, one at US time.
3. That's why you see the volume (in the chart above) getting more interesting at certain times.
4. A few occasions it shoots up high, this is something we want to catch with martingale.
5. Some just go up and down then back to the initial position, this is something we willing to take a loss and move on.

On that basis, here is the WRONG conception of Martingale:

1. Doing the trade almost anytime, one to another, back to back.
2. You think martingale is to double the risk of getting losses wider, just for a small piece of the profit.
3. Doing the martingale trade in unlimited duration.
4. Holy Grail is real.

And here is what I thought would be the best strategy for Martingale:

1. Do the trade ONCE a day only.
2. When you think the risk is doubled, don't forget that the profit potential is also doubled. Take the profit as much as you can if the price is in your favor, to cover the future losses.
3. Don't start the trade first, put some trap. Do something like Price Action trader would do, the only difference is martingale is a little bit stupid. You don't calculate, you don't predict, you don't assume. Just put and let it run.
4. Do in very limited duration, take profit/loss for the day then continue afresh for tomorrow. This is the mindset that most people forget, but it's in their minds already. Just like a very professional trader, they always stop trading, be it winning or losing, then continue tomorrow.
5. Don't trade if one side (EUR or USD) is on a holiday (Bank Holiday, or any holidays).
6. Holy Grail is not real, no trading strategy will give you win all the time. It's the risk management you can count on.

After years looking at the best method for martingale, I can confidently say I have found one, but this is still far from perfect. The risk is still there and needs a couple of tests in the long run.

Above backtested from 20th April to 8th May 2020, I am trying to run full test 1 year backward.

Just an update, I took notes during my backtesting, carefully looking at the FF calendar, filter out the currency I like (EURUSD), and tag the news type. Here is what I found interestingly can be attached to Tunnel Martingale.
- Unemployment Claims
- Flash Manufacturing PMI
- FED Announcement (Red one) - this also imposes better movement over other red high-impact news nearby the time
- ADP Non-Farm Employment Change (careful, sometimes sideways)
- Non-Farm Employment Change & Unemployment Rate (15 minutes before, 20:15pm)
- CPI m/m & Core CPI m/m
- Retail Sales m/m & Core Retail Sales m/m
- German ifo Business Climate? maybe?
- Spanish Unemployment Rate
- Advance GDP q/q & Advance GDP Price Index q/q
- ISM Manufacturing PMI
- German Constitutional Court Ruling
- President Trump Speaks about economic after 14:00 to 22:00, no other stuff
- Surprisingly, Building Permits gains some movement
- ECB Press Conference
- if a lot of EUR high-impact news on 16:00, set the timer to 14:00 instead

And have settled through all the types of news backtested until the early week of March 2020.

OK, just to settle in on what's going on with Tunnel Martingale from the last update.

For those who don't follow, let's get a refresher. Tunnel Martingale basically consist of 2 main things: Ceiling, Floor. Both are the price line which guards if the price goes towards past Ceiling, it will put pending order on the Floor 2 times lot size of the previous (Ceiling). If this pending order gets executed, it will then again add pending order on the Ceiling 2 times lot size if the Floor. Until then it breaks out, (in example 0.16 of the lot size), it will take the profit and repeat from start all over again.


Moving to cTrader
The reason for this is obviously the expanding logic of Tunnel Martingale can't be accommodated within Metatrader which the base programming language is C++. I mean, I am a full-time application developer obviously and there's a reason why I choose C# over C++ (it's an old programming language that's just like that, never got so much update). C# is an expanding universe, coding with it is so robust, quick, fast, it lets us focus on creative ideas more, rather than focusing on building unnecessary components to achieve the idea.
Let me get a simple analogy, "you want to go to New York from Seattle". With C#, you've already got your components ready to be used like "cars", "motorcycle", "trains", "buses", "airplanes", just use it. Maybe in the future will get "teleport", in C# will get it. While with C++, no one is going to build for you, you need to build those transportation modes from scratch & handmade!

It's not your typical martingale, more enhanced
My last Tunnel Martingale in Metatrader 4, it's like running continuously non-stop. And this is the ABSOLUTELY WRONG method of doing it! The main reason is, the market is unpredictable. But, as I move forward from time to time to realize, the market is actually not that unpredictable. You can't predict the unpredictable, but you must predict what's predictable. The "WHERE it will move" (upward, downward, or sideways) is unpredictable, but the "WHEN it will move" is predictable.


Example: EURUSD will never make its move during the wee hours, it'll just sideways. But it will move commonly after high-impact news are released, or without news but just common opening time of a specific market (2-3 pm of my Singapore time, for EURUSD, see screenshot above). I don't care whether it will go up or down, as long as it moves toward one direction and not sideways.

The problem I found with the previous version of Tunnel Martingale:

1. it runs continuously, which is bad. Sometimes, the market just going sideways especially in the wee hours.
2. Since it runs continuously, you will frequently face the sideways stuff. The more it gets into sideways, the larger the lot size you need, the bigger the risk, the more difficult you get out, the easier you get stuck and finally withdrawn so much blood!

As I get the rough idea to improve Tunnel Martingale, I had to play with an array of objects, a configuration that reads from a text file (presumably JSON file, since the idea is complex the configuration will be complex too). None of these is instantly available in Metatrader. I had to download/use external libraries, written by another folk that is also (probably) the code is not maintained for years/decades. While in cTrader, it's based on C#, the base .NET framework already fulfil my needs!

The questions that were rolling around my mind (these years)

1. How can I put Risk Management concept in this Tunnel Martingale? Does it just like the typical risk management where we put small lot size for a larger investment?
2. Can I put 2 pending orders X pips above and below the current price? Then whichever pending orders executed (after it gets touched) will be the first tunnel line (either ceiling or floor)?
3. Can I have this sliced as sessions in which I can tag each session to a certain date and time?
4. Can I limit the session into X number of cycles (from 0.01 lot size to X.XX until it turns profit)?
5. Can I close all the positions regardless of profit or not, in order to limit the exposure of risk generated by the martingale?
6. If my first position is active (i.e. SELL) then the price against it until the second position is active (i.e. BUY), then it turns back to the first position's favor again, can I cancel the second position and compound the losses and consider it into the profit?
7. If the session closed prematurely (due to the certain expiry date or similar concept), can I compound the losses into the next session, so that I am not in losses?

Now, here's the idea which currently has been incorporated into the new Tunnel Martingale (in cTrader) of course:

1. Traps
   If you set TM running mode as Traps, it will NOT start the martingale immediately (no 2 below). It will set pending orders above + below the current price with the initial lot size. IF for example it goes up and the upper pending order is activated, then the pending order below will be canceled and replaced with 2 times of the lot size of the first position. Then from this point, the martingale starts (no 3 below).
   
   Attached Image

   
2. Session Dates
   Now, this is a new concept that has been thought for so long since 2 years past! But I can't do it in Metatrader MQL as the programming language which is C++ is too vintage to accommodate my logic. I want to have an array of dates to which I want to trade this Tunnel Martingale at a certain date and time and only once.
   This means, I will have time once a month, defining the date + time I want to trade, based on FF calendar. Once set up for this month, then I will let it run through the month independently and host it in the Virtual Private Server. Why this Session Dates?
   
   Because a lot of occurrences happen when the market actually moves high pips after high-impact news is being released, even on standard open time for EURUSD that is 1400hrs on my local time. Who wouldn't want to catch that with this Tunnel Martingale? I do!
   You can take a look at historical data, any broker, on that specific date and time.
3. Close Orders after a session finished
   Now, this idea just brought recently in the past few weeks. This is the risk management concept in this TM. So, if from each session date there is no profit or it's hanging (not closing) for profit, based on the specified "end time", it will close all the orders, regardless of whether it's profitable or not. Then tomorrow (or the next session date) can start afresh.
   For example, you can start to run a session today at 14:00 which will last for 3 hours, so 17:00 will be the end time. If within these 3 hours, it got reached the profit, it will close all of course, as per normal. If not, then it will wait until 17:00 to close all automatically.

I will tell you about SmartBucket and Compounding SmartBucket next time. I'm exhausted.... sorry.

Wow... very detailed explanation... This is interesting
Thanks for sharing this