I think we spoke sometime in the past about this. I was having an issue with a code I think and you said generally there are 2 main assumptions one has to accept or work with using blueshift.
1)Slippage
2) You arnt moving the market when you make trades.
Are there more? Was I wrong in trying to remember the 2 main assumptions? Is there a way I can simulate the second one?
Blueshift executes backtesting with a simulation that uses minute level prices and volumes to estimate order fill and slippage. The slippage and fill models are usually based on volumes (for equities). Any instruments or strategies that is not reliably simulated with this set up, will not produce reliable backtest results. This include illiquid instruments (where volume based slippage does not apply), market making strategies (where a passive simulation ignoring reaction to your orders does not apply), or strategies with high order to fill ratio or very short holding periods (where minute bars based simulation does not apply) etc.
I was asking is there any other assumptions of trading algorithms. And am wondering if im wrong about the two assumptions mentioned.
The trading algorithm is something that you write - you are free to have any assumptions as you like. Blueshift engine simulates the algorithm subject to the limitations of the simulation as outlined above. Your two points from the original post is correct but not exhaustive. To test if your strategy can be reliably simulated on Blueshift, check against the points above on how the simulation works.