Key: Exchange Implemented Stops means stops that the exchange supports, Dynamic means we have to run code to continually update the stops on the exchange and/or generate our own Exit signals.
Exchanges support vastly different stops technology - especially when we consider trading on spot exchanges. For long term edges not paying the swap fee might make sense. Or using spot might have different tax calculations - no idea.
If we use stops that are re-calculated daily (say based on ATR on the daily candles + an inital fixed percentage stop) then it's easy to use heartbeats to make sure the Dynamic stop re-calculation is happening.
Stops based on each tick are hard to backtest/test as we need tick data to test them.
It's much easier to backtest/test something that just has one fixed stop price at each given daily candle. No profit target, open ended profit, and one stop price.
We assume the edge code wants to detect all it's own exit conditions so it can be used as a backtester.
This also allows the Edge code to generate performance data when live trading without relying on the Trading Engine code to actually enter positions. i.e. if portfolio becomes fully allocated.
Having an API/ABI/Interface the Edge code plugs into that allows it to be tested and used as either a live Edge or backtester seems ideal.
We might want to use a harness that allows us to use Python for performance analysis.
So: Candles Data -> Blackbox -> Data File of Trades/Signals -> Python Performance Analysis
Moved to Backtesting Architecture.