7pairs dynamic grid

Avatar
  • updated

Hi guys,

So I was just playing around with various grid step-ups based on volatility. And ended up with an idea to basically go smaller then bigger when calculating the step distance. the only way to do that was to go with custom step coefficients.

Image 4957

As you can se we first start decreasing the step sizes then when we are in a situation where market is still moving against us we start expanding the grid. 

We also use the 5min ATR, but with a look back of a whole day, 288. This gives us the average 5min movement of the day to base your grid spacing.

The last part of the grid was the lot sizes. Usual calculation types often blow the account, so I ended up taking the Fibonacci sequence and stretched it to 15 orders with custom lot sizes just to maintain control on risk. The best custom lot sizes I came up with is the following:

Image 4958

This keeps the lot sizes in check, but makes it bottom heavy to be able to average out the trade a bit easier.

So then I had to work on entry signals, this part I find doesn't matter as much to this idea, but I decided to use Stochastic K and default TDI. (Feel free to choose any signal you want). I ran this for all 28 pairs filtering out any set with greater than 30% relative drawdown. Then looking at open trade correlation, I went with the least correlating pairs to lessen the chance of drawdowns lining up, these are the best picks from the results.

Image 4959


I decided to exclude TP and SL for the time being and go purely with signals. Once I'm confident enough, ill aiming for a global TP and SL, which would have all the different grids interacting together.


Here are the results so far:
AUDCAD:

Image 4960


AUDUSD:

Image 4961


EURGBP:

Image 4962


EURUSD:

Image 4963


GBPUSD:

Image 4964


NZDUSD:

Image 4965


USDCAD:

Image 4966


The above results are from 1-Sep-2022 to 31-Aug-2023, last year, using my broker, blueberry markets raw account.


SO... the idea here is to have these 7 pairs trade together... and since I don't really have a proper way to combine the results, and since the initial lot size is 0.01 per 1000USD, I think that having all 7 will compound each other to get bigger and bigger entries which should result in a higher return overall.

Here are the set files:

AUDCAD.set
AUDUSD.set
EURGBP.set
EURUSD.set
GBPUSD.set
NZDUSD.set
USDCAD.set


Use these for whichever pair you like, the main 3 parameters that you should optimize are:

  • Pending order distance
  • Martin Step Size
  • Whatever entry signal you decide to go with.


The above sets are NOT a plug an play sets, just an idea I've been working on that thought I'd share and brain storm with you guys. And various brokers will sure have different setups and results.


Thank you!

MT version:
Set for MT5 only
EA version:
2.55
Symbol:
AUDCAD, AUDUSD, EURGBP, EURUSD, GBPUSD, NZDUSD, USDCAD
TimeFrame:
M5
Avatar
Chris Tom
Quote from jay hu

I am just curious if it is possible for you may implement this set and use it in the Ultra Hedge project. It might improve the performance.

Very good idea and I already have some accounts runing with pretested SetFiles/pairs.

You can see them here: https://www.mql5.com/en/users/chthb/news

All signals named "Test" are different combinations of your files running on different accounts ;)

Avatar
jay hu
Quote from Chris Tom

Yeah! I really like the SetFiles and the results of them!

Looking forward to also test you improved version. 

But so far...great work man, keep on doing! :D

I am just curious if it is possible for you may implement this set and use it in the Ultra Hedge project. It might improve the performance.

Avatar
Chris Tom
Quote from lytnin

Nice man! I think I need to work on a better entry signal setup, that could increase overall profitability... the grid seems to be ok generally.

Yeah! I really like the SetFiles and the results of them!

Looking forward to also test you improved version. 

But so far...great work man, keep on doing! :D

Avatar
lytnin
Quote from Chris Tom

Okay, so I did a second run with all +70 pairs available on ICM and here are the results: 7 pairs dynamic grid V2.xlsx

Nice man! I think I need to work on a better entry signal setup, that could increase overall profitability... the grid seems to be ok generally.

Avatar
jugulonline

nice set :)

Avatar
Chris Tom
Quote from jay hu

Great. Thank you for your time and dedicated work, Chris.

Okay, so I did a second run with all +70 pairs available on ICM and here are the results: 7 pairs dynamic grid V2.xlsx

Avatar
Yannis Latroche

Interesting, i run on a demo account a multi-pairs grid without your custom stepsize on 5 pairs for now with fixed lot size but the drawdown is kinda meh so thanks for your idea

Avatar
jay hu
Quote from Chris Tom

So...took some time, but now here is the Excel File with the results: 7 pairs dynamic grid.xlsx

Great. Thank you for your time and dedicated work, Chris.

Avatar
Chris Tom
Quote from lytnin

Thanks for sharing results!

are you using the same settings for all pairs?

Yeah, but I do a second run just to be sure. 

ICM an ICM SC should be (almost) similar...

Avatar
lytnin
Quote from Chris Tom

So...took some time, but now here is the Excel File with the results: 7 pairs dynamic grid.xlsx

Thanks for sharing results!

are you using the same settings for all pairs?