As we all know, market orders in the testing environment are filled at the exact bid or ask price. But in real market conditions it will be different for sure and it makes a difference over time. Do you have any ideas on how to simulate this, or (re)calculate the results considering slippage?

I suppose a distribution of the individual slippages is needed. It would be hard (costly) to collect this data, so suppose it has a normal (Gaussian) distribution between -2.5 and +2.5 pips (EURUSD). How to proceed from that?