Not perfect at all, but – to simulate slippage etc for Paper Trading / Demo Forward Test – I add an extra point onto my spread setting.
Yes, that is what I did too, in the past. These days I have a very different approach : causing PaperTrading to behave exactly the same as Backtesting (see attachment) – or maybe it is the other way around 🙂 – and next be able to trust the BackTesting which runs in real time. So a. that includes the Commission (or spread) all right and b. because the program code knows about the settings too, it all works consistently.
Ad b.: You’d sell above the spread or Commission loss.
I don’t want to confuse things in this topic unnecessarily, but there’s another bug regarding this. See 2nd attachment;
That spread can’t have more than 1 decimal, while this is most certainly not enough to mimic reality. So might PRT read up till this post anyway …
If we in-code must simulate the spread in order (this is crucial for PRT-IG, less for PRT-IB) *and* you see me simulate my own PaperTrading in BackTest, then with 1 decimal this won’t work out. Btw, I saw someone (a moderator) tell me the other day that you can’t know the real spread … well, IMO you can. Just look at what you lost after entering the order. Of course this is at the entering of the order only (at the exit you will be too late – haha). That aside, if we can’t have this more precise, we also can’t simulate PaperTrading by means of BackTesting (yes, that is the proper sequence).
So all is related a little …
… and if I can’t simulate properly, my graphing will also be worthless (graphing which can’t be done in PaperTrading nor in real AutoTrading). See 3rd attachment.
Am I exaggerating things ? no man, this is $. 🙂
.
