Simple High Profit low Drawdown Triangular Arbitrage

raffus,

If you talking about the hedging 3 pseudo renko EA please check the following thread: https://www.forexfactory.com/showthread.php?t=335072
Also please read again the #Post 295 on this page

Jacques,

I'm in the UK, and this is still the Standard Demo account, no VPS. Maybe it would be nice to try it out on a small, let say $100 live account with them. Perhaps another one with Dukascopy? Will think about...

Jacques,

I'm setting up my $100 live account with SMFX. Just had some chat with them...
They have a margin call at 200%, but your ea's margin level is 126%. Is it possible to adjust it? Without that can't-do a live test.
My Metatrader's Journal says "Trades: use hosting service to speed up the execution - 1.54 ms via 'MQL5 New York 6 (MQL5 Ltd.)' instead of 111.98 ms". (I'm in Hungary this week.) How can I do that? SMFX says their VPS is a better option so thinking to set up a VPS with them, but they have 3 choices. I don't know which one is better for your ea.

Thanks

@eatrader75
ea 19 is way too unstable to recommend for live trading.
I suggest waiting for the next one.

As for which VPS.
That suggestion must be made by the broker SMFX.
I would think the closest VPS to the SMFX server.

Still I think using EA19 for a live test will be a waste, but that I will leave up to you.
I myself would like to see the response time for trades on a live server and if there is a big difference between it and a demo.

The EA adjusts to the available free margin, but I will test it on that broker when I get time.
Sorry for taking so long.
I'm struggling at work and therefore to get free time.

Absolutely is a problem! Your net position is 0.003 short USDJPY. You're paying through the nose to hold that trade!

I really don't understand the pursuit of this strategy from a retail paradigm. If you want to be successful with arbitrage you need a c++ algo running on a specialized kernel on a colocated server connected directly to LPs via FIX. Otherwise, slippage, swaps, and trading costs will erode any potential profits with this system. This strategy is what billion dollar HFT firms spend millions of dollars to beat out the next poorest billion dollar HFT firm by a micro-second. There is no way to compete with a retail broker and the metatrader platform.

i told them already same. you are first with same opinion and trying to help them too to save time, looks like you trade also in real.

can someone plz tell us that will this EA*(JVB artibrage) close the trades by itself? and what criteria it picks buy or sell for the GBP and JPY?

creator of this EA has brilliant mind but unfortunately has no fund to develop this master piece for all of us . He is trying gradually due to lack of fund so may be this EA will complete in near future or will complete for our children when they grew up.
anyway if someone want a quick up gradation of this EA, should cooperate in funding.
thanks

@Nicholishen
Swap does not play that much of a role.
Problem with EA 19 is the slippage and invalid entry point.
Once a trade is opened incorrectly then EA 19 cannot recover and will never recover, meaning swap will accumulate, but even multiple days worth of swap does not affect this strategy.
Also.
Trades normally lasts a couple of hours at most, so that by itself is showing a problem with EA 19, you can see that by back-testing the EA.
Back testing does not suffer slippage, but will show you the affect, or none affect of swap on the strategy.
The next version of the EA should cater for slippage after open and recalculate the exit point.
Read on for more information.


@T4Trade
Yes it closes trades by itself, but there is a problem with EA 19 in that with slippage it sometimes opens trades out of place and the nature of triangular arbitrage is that it has got fixed highs and lows and therefore a way out of place trade will never hit it's profit target.

I already have a solution planned for it in that I need to recalculate the full history arbitrage levels or as far back as possible after open to check for the correct close level, even if at a loss, to prevent perpetual open positions, which EA 19 is susceptible to.
EA 18 uses a pre-calculated open and close, but suffers from the opposite problem in that it does not take into account slippage during the open and therefore will also close at the wrong position.

The solution is a combination of the 2 methods by recalculating after open and then adapting the close over time.

To see how the balance of the triangular arbitrage works then run it using a MT5 back test for a couple of months and look at the lines it draw on the chart to see the stability and potential entry and exit points.



@badeel
Thanks badeel.
I am already training my 7 year old boy to finish the EA once I am gone in a few hundred years (I plan to live long)

Unfortunately you are right.
I can only develop in my free time.
Laws in my country prevent me from making a profit from my programs and inventions without giving the majority of my income to 3rd party connected companies and therefore I live from hand to mouth.
Therefore free time is few and far in-between with only 5 to 6 hours sleep at night there is not much left.

BUT I will continue until I am done.
I already have about 4 new versions running here.
Even some mutations like an unbalanced triangular arbitrage grid system and a single/double legged triangular arbitrage snowball system.

All of which is unstable and just mutated from the tests so they did not take much time from the main EA development, but looks nice

I will continue and will update this forum with updates as they occur.
Monitor post #1, which will always contain the main updates.





@EVERYONE
Slippage is mainly caused by the slow execution speed of the EA.
The delay seem in part to be caused by metatrader itself, but I hope to be proved wrong by a fast broker somewhere out there, but alas after testing multiple brokers I could not find a really fast one so I went the route of sacrificing some profit to cover the slow speed and slippage.
Still working on that and will continue working on that until done.

I also whent to Dukascopy to test the rumours that they are faster with their Jtrader program.
WOW and are they faster.
According to people I talked to they said that Jtrader is faster on a live account than a demo account, which makes sense for any broker I guess, but I do not have funds to test on live so I went and used their FREE VPS, ANOTHER WOW
Even on their DEMO VPS ..... WOOOOOOOWWWWWW...

300ms to open all 3 trades fully and at the same time.
Another 300ms to close all 3 trades.
I once got 120ms for 3 trades.
Basically 300ms per trade, but you can do multiple trades at the same time and therefore they finish at the same time and not one after the other like in metatrader.
That is an massive amount faster than anything I could get on even MT5.
Or about the same as metatrader for a single fast trade, but JTrader can do multiple trades at the same time with no slow down, unlike metatrader which need to do one after the other and even if you do multi threading then metatrader still have to first reconnect each thread if a trade is needed, which adds even more delay.

I would suggest a live test on them before anything else, but I am getting ahead of myself.
Let my first finish the EA.

I'm now converting the EA to run on Dukascopy Jtrader.

Other than the speed I hear only good things about dukascopy so if anyone know a reason as to not to trust them then please let me know.

This also brings back my theory that to really boost this strategy that I need to incorporate each broker's unique API to really get to the bare bones of the forex to get good execution speed.
The FIX API seem to be the go to for most fast brokers, but to develop it I need a demo account, which I can find no where, so any advice on getting a demo would be appreciated.
A live FIX account I cannot afford so demo is best.

If only more brokers used Jtrader for that is a wonderful platform, so powerful, but FIX is supported by many more, which tells me that once development is done and working on Dukascopy and Metatrader then FIX is next.

Keep you fingers crossed.
Here I go with a mountain of coding to do What fun

I would like to know how to use this EA on MT4. Thank you

Hi everyone

This week the EA did not close the open trades.

On the other hand I used the strategy tester of Metatrader 5, starting from a figure of 10000 EUR and a leverage 1: 100 and trial period of one year, with a speed of 72 ms.
The EA has reached almost EUR 170000 in two months and has not traded anymore.
Questions:
Why stop at two months?
Is it possible that in two months (2017/01/02 - 2017/02/13) you have multiplied the equity / balance by 17?

I attached the test files

The MT5 tester is doesn't work with this strategy because ticks are not accurate and aren't properly synced.

Based on my experience, several other things don't work correctly on MT5. They should just continue allowing brokers to use MT4 until MT5 is sufficiently bug tested in order to get the basics right. Even then, hopefully they continue allowing the use of MT4 since there are already so many tools or solutions created for MT4.

my one ..no trade ...

@ZanderZhang
Please attach you log of the EA so I can have a look.

Basically you just have to attache the EA to a GBPUSD minute chart and have the GBPJPY and USDJPY symbols listed in the market watch.

Also make sure your metatrader is set to be able to run EAs.




@artico01 AND @rizam34 AND @rizam34
I have discovered a bug in EA19 in that it sometimes opens up a trade and then never closes it.
I have already fixed it, but are still working on some other coding enhancements so I have not updated the forum with the updates yet.


@artico01
I am not sure why it stopped, but there is a couple of reasons.
1) There is a 50 lot limit and once reached the EA will stop.
I put it in seeing that the original broker I used had a 50 lot limit and to overcome it I have to open multiple trades, which just adds to the trade delay.

2) There is a problem with EA 19 at the moment that it sometimes opens a trade and leaves it open forever.

As for the accuracy.
You seem to make a lot of profit during the back test, which is telling me you either have a fantastic broker with fantastic spread/commission OR you are using the wrong testing parameters.
When you test on MT5 then do not select "EVERY TICK".
Select "EVERY TICK BASED ON REAL DATA".

Every tick based on real data is more accurate and are time synchronised for multiple currency testing.
Not every broker have got accurate real tick data, but I found myfxchoice to have some good data for the past 3 to 5 months at least.

You will see the profit drop a lot, but still good profit.




@EVERYONE
My phone broke and I could not afford a phone with whatsapp capabilities yet so those that contacted me over whatsapp please note that I am not ignoring you.
I hope to have access to a whatsapp capable phone some time in future.

I also am stuck between multiple odd jobs, which takes away most of my free time and therefore development has slowed down to a crawl.
Will keep you up to date as I get time.

This is what I showed after loading

After loading on the MT4, the EA no orders, I set up according to your settings, I would like to ask what is the next reason? Thank you

@ZanderZhang
According to your screenshots the EA seem to be running fine.
Maybe zoom the chart a bit out so that you can see the lines a bit better, but it is running.
Also you have attached the EA to a 5 minute chart, which will not show the EA lines that good, but will not affect the running of the EA.
I suggest a 1minute chart, but again it does not matter for the running of the EA, just for the display of the lines.

The logs can be accessed by right clicking in the "Experts" tab and selecting OPEN.
Then the log text files will be displayed and that you can attached to this forum so I can check, but the Experts tab screen shot looks good and the lines on the chart is showing a very good spread/commission.

I still do not know it under the law, what kind of circumstances for automatic trading

jacques
if you happen to develop the ea/strategy for jplatform of dukascopy then do share