Grid/Basket Trading Expert Advisors

Interesting discussion.

Grid and basket systems always look simple at the start, but the real challenge is usually not entry logic, it’s trade management once positions start stacking. Equity-based triggers like what you mentioned are actually more flexible than pip-based grids, but they can also make the system very reactive if market conditions change quickly.

One thing I’ve noticed with basket approaches is that once you tie new entries to drawdown instead of structure, you basically shift the whole logic from “market movement” to “account pressure”, which can work, but it needs very strict limits on exposure.

I set a 'Maximum Loss Percentage', backtest those with the grids, get a strategy that maximizes returns with the sawtooths included. Run a few thousand tests and find that band of optimal defined loss / grid profits.

[quote=poeladifeel;15652261]Interesting discussion. Grid and basket systems always look simple at the start, but the real challenge is usually not entry logic, it’s trade management once positions start stacking. Equity-based triggers like what you mentioned are actually more flexible than pip-based grids, but they can also make the system very reactive if market conditions change quickly. One thing I’ve noticed with basket approaches is that once you tie new entries to drawdown instead of structure, you basically shift the whole logic from “market movement” to “account pressure”, which can work, but it needs very strict limits on exposure.[/quot

I agree. The important parts of grids is the exit. Creating the algorithm to exit is the real challenge.
I read a few post . I do not see why some talk about time frames in grid systems.
Perhaps my work is not pure grid systems, but similar approach. I use renko bars to determine the trend and entry. To me, trend is 1 renko bar. Whatever pips I set the renko variable.
In essence, my systems Triger entries when a renko bar close. For example if bar set at 10 pips, then from current price at start a first bar would close at +10 or -10. The system then would trigger a long or short according to that last brick. Let say that first bar was UP. Next bar would close at +20 or -10 . Then the system would triger a long at +20 or a short at -10. And forever like that until your exit algo triggers a close and reset bar count. If you can imagine, it is creating a grid that follow the price whatever it goes . Here is huge tip. I made code that measure continuation. In 10 pips bars, the continuation rate is 67% . ( 5 years test) Meaning, every time a bar close, say UP, there is a 67% chance that next bar would also be UP.

on what pair ? mean reverting or trending highly depends on the pair used. JPY pairs use to trend more. AUD/CAD/CHF tend to MR more

That’s an interesting way to approach it.

Using renko like that kind of removes a lot of the noise you’d normally get with time-based charts, so I can see why the continuation stats become more consistent. A 67% continuation rate is definitely something you can build logic around.

The part that stands out to me though is how you handle the transition phases, when the market shifts from continuation to reversal. That’s usually where these types of systems give back a lot if the exit logic isn’t tight enough.

Also, since the grid is effectively “following” price, do you find that it behaves more like a trend system than a traditional grid? Because at that point it’s almost like stacking into momentum rather than averaging into drawdown.

Curious how you’re managing exits when the sequence starts losing structure.