Scaling up in both sides Grid EA? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Me thinks those guys play both sides 0 replies

Bolgar Strategy - for taking on both sides of the trend 41 replies

Grid, Burst Grid, Grid Risk and Grid Profit 1 reply

both sides of pair for confirmation of trend 10 replies

scaling in/scaling out - how do you do it? 2 replies

Attachments: Scaling up in both sides Grid EA?
Exit Attachments

Scaling up in both sides Grid EA?

  • #1
  • First Post: Edited 9:47am Dec 31, 2019 7:53am | Edited 9:47am
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result.

The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds new pending orders in both sides.

Also, if you know any free or commercial EA package for this purpose, kindly share the information.

Thanks!
Simplicity is the ultimate sophistication - Leonardo da Vinci
  • #2
  • Jan 1, 2020 9:22am Jan 1, 2020 9:22am
Quoting heispark
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result. The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds...
I have programmed some EAs with similar ideas in the past. It wouldn't work (to be profitable with low DD) in the long-run and under any market condition.
Trading is the hardest way to make easy money...
  • #3
  • Jan 1, 2020 10:27am Jan 1, 2020 10:27am
Quoting heispark
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result. The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds...
heispark.. u can check up nononsensefx thread here.. i think it was the one that u looking for.

https://www.forexfactory.com/showthr...9#post12679159
  • #4
  • Jan 1, 2020 10:58am Jan 1, 2020 10:58am
Quoting heispark
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result. The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds...
Here is such an EA: Snowballs and the anti-grid. If run in bi-directional mode it does exactly what you want. I implemented my own version (SnowRoller) and currently also trade it. As @MathTrader7 said the concept is only profitable under very specific market conditions. For this reason I specifically removed the feature you are asking for: the bi-directional mode.
It takes blood, sweat and enough tears to wash away the blood.
  • #5
  • Edited 2:12pm Jan 1, 2020 1:36pm | Edited 2:12pm
And here comes your strategy: As you probably know by now the breakeven characteristics of such a bi-directional pyramiding strategy look like a parabola turned side-ways with the vertex being your trade starting point. The blue line is the breakeven line, every position inside the parabola is a "loss" and the moment you cross one of the side lines you enter "profitable" area.

Attached Image


Traders who have been there tell you that such a strategy is only profitable in specific situations. If the market trends strongly for some time you will be able to leave the parabola and come into profit. As the parabola widens over time the longer you stay in loss the harder it becomes to get into the profitable zone. This fact can be turned around by reversing the strategy. A reversed bi-directional pyramid will stay most of the time in profit, and rarely get into negative territory.

Wouldn't reversing be a typical Martingale? No, it isn't because a bi-directional pyramid is not anti-martingale. It looks similar but it is not.

Can it be profitable? Yes. Is it the holy grail? No, it isn't. Try to implement it and you will realize that plenty of things can go wrong, correct implementation and execution is a challenging task. You need automation, don't even think about managing pyramids manually. This means automation and PL/DD management of multiple such sequences.

The critical point is the vertex, i.e. the start. There the distance to the BE line is the shortest and crossed the fastest. You have to define a SL and will hit it a few times before the winner comes. At the time the winner comes you need to know before where to stop. Can't be done manually. It's definitely fun as you learn a lot about things that can go wrong - I would recommend playing with it only for that experience, not for the potential.

The real potential is a one-directional pyramid. Distance to the BE line shrinks which is good but you have to be right about the general direction of the market. This means a pyramid needs to be combined with standard trend determination/forecasting with all it's potential for failures and bad signals. This is what I do.
It takes blood, sweat and enough tears to wash away the blood.
  • #6
  • Jan 1, 2020 7:19pm Jan 1, 2020 7:19pm
Quoting emmzett
{quote} Here is such an EA: Snowballs and the anti-grid. If run in bi-directional mode it does exactly what you want. I implemented my own version (SnowRoller) and currently also trade it. As @MathTrader7 said the concept is only profitable under very specific market conditions. For this reason I specifically removed the feature you are asking for: the bi-directional mode.
Hi emmzett, would u mind to attach your mq4 or ex4 file here.

Thanks in advance.

I like to backtest this strategy
  • #7
  • Jan 1, 2020 7:26pm Jan 1, 2020 7:26pm
Thank you all for your kind replies and detailed information, especially MathTrader, Emmzett and TokanCorner. I really appreciate it!
Simplicity is the ultimate sophistication - Leonardo da Vinci
  • #8
  • Jan 2, 2020 3:49am Jan 2, 2020 3:49am
Quoting heispark
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result. The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds...
High Spreads with scale up in comparison of small targets is worst enemy of this strategy.
In the court of lord, only love and devotion is counted...
  • #9
  • Edited 5:40am Jan 2, 2020 4:10am | Edited 5:40am
Quoting nadiyakinare
{quote} High Spreads with scale up in comparison of small targets is worst enemy of this strategy.
Spreads are not an issue at all with this kind of strategies as long as you stick to a sound base ratio spread/ADR/gridsize. There is an issue with slippage in fast moving markets as such strategies use stop orders and not limit orders. With careful automation that's not a blocker. When the market starts moving fast (NFP) you will even experience skipped gridlevels, not only one but 1-2-3 in a row. Also this does not break the strategy.

Real issue (worst enemy) here is that such a system needs to reliably run 24/7 with all the possibilities of server shutdown, network loss, trade server timeouts etc. It needs to be able to handle all of those and dynamically recover from them. Once a sequence is out of sync it's very difficult/sometimes impossible to bring it back. You have to start over with a new sequence. An out-of-sync strategy is not able to trade according to the formerly achieved winners/losses. So here is the real challenge.

Once you have selected an instrument and a gridsize/lotsize spread is almost a non-issue.
It takes blood, sweat and enough tears to wash away the blood.
  • #10
  • Jan 2, 2020 4:21am Jan 2, 2020 4:21am
Quoting emmzett
{quote} Spreads are not an issue at all with this kind of strategies as long as you stick to a sound base ratio spread/ADR/gridsize. There is in fact an issue with slippage in fast moving markets as such strategies use stop and not limit orders. With careful automation even that's not a blocker. When the market starts moving fast (NFP) you will even experience skipped gridlevels, not only one but 1-2-3 in a row. But even that does not break the strategy. Real issue (worst enemy) here is that such a system needs to reliably run 24/7, with all the...
Hi Emmzett,
Could you kindly let us know how to download the mql.4 from your Github link. I would certainly like to test the system you programmed.

Thanks
Your salary is the bribe they give you to forget your dreams
  • #11
  • Edited 5:40am Jan 2, 2020 5:18am | Edited 5:40am
Quoting horvat
{quote} Hi Emmzett, Could you kindly let us know how to download the mql.4 from your Github link. I would certainly like to test the system you programmed. Thanks
It's a standard Git repository. You can clone the full repo, you can download the full repo as a ZIP archive, or you can download each single file (doesn't make much sense but it's possible via the "Raw" button). There are green buttons on every project main page. To compile it you will need the full project and need to follow the compiler instructions in the bin folder (do not try to compile with a current version of MT5-MetaEditor, it will not work). I will also create a release package with precompiled .ex4 files.
It takes blood, sweat and enough tears to wash away the blood.
  • #12
  • Jan 2, 2020 5:58am Jan 2, 2020 5:58am
Quoting emmzett
{quote} Spreads are not an issue at all with this kind of strategies as long as you stick to a sound base ratio spread/ADR/gridsize. There is in fact an issue with slippage in fast moving markets as such strategies use stop and not limit orders. With careful automation even that's not a blocker. When the market starts moving fast (NFP) you will even experience skipped gridlevels, not only one but 1-2-3 in a row. But even that does not break the strategy. Real issue (worst enemy) here is that such a system needs to reliably run 24/7, with all the...
spreads are issue in this, they force you to scale up more rapidly which also turn out to become a higher DD and unwanted martingale in long run. and again High Spreads with scale up in comparison of small targets is worst enemy of this strategy... trends noise keeping in mind with spread to TP % comparison with scale up boosts and trying to achieve as less open position as possible to keep dd low is difficult and does works for short time but not in long run. for NFP you can hedge prior to event. most people use leverage as their edge as you might or might not be doing, but after 2-3 years its very common to see just few trades make them to close the account in loss. martingale is not bad until one knows how to work around them.
In the court of lord, only love and devotion is counted...
  • #13
  • Edited 7:33am Jan 2, 2020 6:10am | Edited 7:33am
Quoting nadiyakinare
{quote} spreads are issue in this, they force you to scale up more rapidly which also turn out to become a higher DD and unwanted martingale in long run. and again High Spreads with scale up in comparison of small targets is worst enemy of this strategy... trends noise keeping in mind with spread to TP % comparison with scale up boosts and trying to achieve as less open position as possible to keep dd low is difficult and does works for short time but not in long run. for NFP you can...
That's all correct what you are saying. All I'm saying is compared to reliable execution spread is a small problem.
It takes blood, sweat and enough tears to wash away the blood.
  • #14
  • Edited 8:30pm Jan 2, 2020 7:19am | Edited 8:30pm
Quoting horvat
{quote} Hi Emmzett, Could you kindly let us know how to download the mql.4 from your Github link. I would certainly like to test the system you programmed. Thanks
I added a pre-release package with precompiled binaries to the Git repo. All files in "libraries" are required, the scripts "scripts\SnowRoller.*" and the used indicators. If I find some free time I may add a backtest.

ps: I kindly ask everyone not to ask questions like "How can I download from Github..." in this thread. There is countless material on the net and you have to do your homework. Thank you.
It takes blood, sweat and enough tears to wash away the blood.
  • #15
  • Edited 10:30am Jan 2, 2020 10:15am | Edited 10:30am
Quoting heispark
Just curious if it can work. I am pretty sure somebody already tried it as it's very basic concept. Please share its test result. The EA builds scaling-up (adding to winning position) grids in both sides (broker should supports hedging), for example every 50 pips. The pending orders should be stop orders (buy stops and sell stops). The EA takes profit at pre-determined profit setup (for example 50 pips or $100). No stop loss. Once this profit target hits, the EA closes all the open positions and removes all the existing pending orders, then it rebuilds...
Yes tried it. Except it's using limit orders to scale. Stop orders are not a good idea!

Based on random entries. Assuming the market is random, how to keep the random walk on the positive ?
Here is 20 years, 1H TF on the pound , yen and euro, same code and parameters and algo
I wanted to take into account some dramatic moves in the market for a stress test ...

Attached Image (click to enlarge)
Click to Enlarge Name: eurusd.png Size: 12 KB


Attached Image (click to enlarge)
Click to Enlarge Name: Gbpusd.png Size: 14 KB


Attached Image (click to enlarge)
Click to Enlarge Name: usdjpy.png Size: 14 KB
  • #16
  • Jan 2, 2020 7:18pm Jan 2, 2020 7:18pm
Quoting Mingary
{quote} Yes tried it. Except it's using limit orders to scale. Stop orders are not a good idea! Based on random entries. Assuming the market is random, how to keep the random walk on the positive ? Here is 20 years, 1H TF on the pound , yen and euro, same code and parameters and algo I wanted to take into account some dramatic moves in the market for a stress test ... {image} {image} {image}
It looks not bad.... Very appreciated for this backtest result. It's interesting indeed.
Simplicity is the ultimate sophistication - Leonardo da Vinci
  • #17
  • Jan 29, 2020 3:41am Jan 29, 2020 3:41am
Quoting emmzett
{quote} That's all correct what you are saying. All I'm saying is compared to reliable execution spread is a small problem.
To extend a little what I mean with reliable execution I want to post a screenshot from this morning. I think the picture is self-explaining. These are the real issues one faces when running any strategy that is constantly active in the market (opposite to a simple "open position with TP/SL").

Attached Image (click to enlarge)
Click to Enlarge Name: networkerror-handling.png Size: 33 KB


The challenge here is not a simple "wait and recheck" until "everythink works again". In between such events the market moves and grid levels will have been passed. A strategy needs to be able to handle it - however that is defined by the trader - without losing track of the current plan.
It takes blood, sweat and enough tears to wash away the blood.
  • #18
  • Jan 29, 2020 3:51am Jan 29, 2020 3:51am
you cannot make a grid without an edge, believe me i've tried numerous strategies, and none worked
the spread is also a problem which makes the edge to be stronger
  • #19
  • Jan 29, 2020 3:53am Jan 29, 2020 3:53am
Quoting ionone
you cannot make a grid without an edge, believe me i've tried numerous strategies, and none worked the spread is also a problem which makes the edge to be stronger
Obviously you didn't read the thread or you don't know what you are talking about. This thread is not the place for you to tell us that "grids don't work". Please go elsewhere for giving such advice. Thank you.
It takes blood, sweat and enough tears to wash away the blood.
  • #20
  • Apr 15, 2020 1:40am Apr 15, 2020 1:40am
Quoting Mingary
{quote} Yes tried it. Except it's using limit orders to scale. Stop orders are not a good idea! Based on random entries. Assuming the market is random, how to keep the random walk on the positive ? Here is 20 years, 1H TF on the pound , yen and euro, same code and parameters and algo I wanted to take into account some dramatic moves in the market for a stress test ... {image} {image} {image}
Mingray, thanks for sharing your results. (I'm a fan of your commentary on market randomness from the other thread)

Could you share the code you wrote for testing this? What general rules did you apply? Specifically:

  1. Were your limit orders equally spaced?
  2. Did you close a previous buy order when price triggered a sell order at the same level? Or did you have buy and sell orders at same/similar levels?

0 traders viewing now