Thank you for your explanation, betagroup.
I appreciate your effort of taking slippage into account. Do not take me wrong, I know what I am talking about. I am scalping very profitable longterm already and am aware of the potential effect slippage has. My experience gives me a quite solid overview here.
As far as I understand you are trying to consider the slippage amount by simulating a delayed execution. Pretty nice and smart approach in a numerical way. But I do not think this solves the issue, as slippage is not only a product of a "delayed" execution. No matter if you execute market orders or get pending orders filled, slippage mostly is caused by liquidity shortage, where the LPs are not able to give you the requested price ALTHOUGH it is instantly executed.
In consequence this could lead to really horrible deviations you cannot simulate numerically, that is the sad truth. You simply do not have the information of how bad slippage can/will get to implement it in your backtest calculations as a time delay is not doing the job sufficiently. I have seen cases by myself, where the filled price was far away from any quotations. No backtest would have ever considered this. (This it is not a broker issue, I am using six different ECN's to diversify and you can not avoid such executions, especially when trading higher lot sizes)
If I may advise, just skip the forward test and check the live conditions, everything else is more or less a waste of time.
I appreciate your effort of taking slippage into account. Do not take me wrong, I know what I am talking about. I am scalping very profitable longterm already and am aware of the potential effect slippage has. My experience gives me a quite solid overview here.
As far as I understand you are trying to consider the slippage amount by simulating a delayed execution. Pretty nice and smart approach in a numerical way. But I do not think this solves the issue, as slippage is not only a product of a "delayed" execution. No matter if you execute market orders or get pending orders filled, slippage mostly is caused by liquidity shortage, where the LPs are not able to give you the requested price ALTHOUGH it is instantly executed.
In consequence this could lead to really horrible deviations you cannot simulate numerically, that is the sad truth. You simply do not have the information of how bad slippage can/will get to implement it in your backtest calculations as a time delay is not doing the job sufficiently. I have seen cases by myself, where the filled price was far away from any quotations. No backtest would have ever considered this. (This it is not a broker issue, I am using six different ECN's to diversify and you can not avoid such executions, especially when trading higher lot sizes)
If I may advise, just skip the forward test and check the live conditions, everything else is more or less a waste of time.