Hey Henry! Unfortunately this is not supported right now, there have been a few discussions like this issue about supporting something. You can also read more about what you can and can't do with Gekko in the scope doc. The current reasoning is:
Writing code that creates orders of different sizes is easy, but what becomes very hard is explaining to the user how a strategy is progressing as it's running (how do different "advice" signals impact performance like profit and losses). Right now Gekko has a core concept of a "roundtrip" which is combination of a "buy trade" (of the full amount) and a corresponding "sell trade" (of the same amount). After those two trades Gekko can calculate some statistics (like how much money did you make) as well as aggregate all of them into more general statistics, such as:
- statements like "60% of your roundtrips are profitable".
- your roundtrips adhire to this distribution (stddev, etc).
- what is your "worst" and "best" trade.
- more financial statistics like sharpe ratio (risk VS reward), etc.
This way introducing a way to properly reflect on/compare different strategies over different markets (besides only looking at the "profit" metric).
I'm definitely thinking about and open to more complex systems, but that's definitely pretty far away unfortunately.
-----
That said I am introducing limits on how much Gekko can trade with (limit all trading to 100 bucks a trade for example) very soon.
Writing code that creates orders of different sizes is easy, but what becomes very hard is explaining to the user how a strategy is progressing as it's running (how do different "advice" signals impact performance like profit and losses). Right now Gekko has a core concept of a "roundtrip" which is combination of a "buy trade" (of the full amount) and a corresponding "sell trade" (of the same amount). After those two trades Gekko can calculate some statistics (like how much money did you make) as well as aggregate all of them into more general statistics, such as:
- statements like "60% of your roundtrips are profitable".
- your roundtrips adhire to this distribution (stddev, etc).
- what is your "worst" and "best" trade.
- more financial statistics like sharpe ratio (risk VS reward), etc.
This way introducing a way to properly reflect on/compare different strategies over different markets (besides only looking at the "profit" metric).
I'm definitely thinking about and open to more complex systems, but that's definitely pretty far away unfortunately.
-----
That said I am introducing limits on how much Gekko can trade with (limit all trading to 100 bucks a trade for example) very soon.