Good slippage value for backtesting?

[skellington]

Hi,

Anyone know what a reasonable/representative slippage value would be for backtesting?

In my comparisons between live and simulated strategies, a slippage value of 2.5 seems to make the simulated match the live results fairly reliably.

What have you guys been using?

Thanks!

[NinjaTrader_JoshP]

Hi skellington,

It is really up to you. Different markets will behave differently so slippage is never something you can pin a hard number to for testing. I suggest you explore a range of slippage that you think is feasible for the instrument you are trading on. That way you can understand how it will behave in a broader sense.

[skellington]

Ok, I'll do that. Thanks.

[skellington]

Actually one other question...

If a system places long and short market orders shouldn't the slippage on the average be symmetrical? As in, the traded price would as often be above the simulated price as below (with perhaps a bias in the direction of the trade).

Is this true or am I overlooking something?

[NinjaTrader_JoshP]

Depends on your school of thought. In theory I can see it being symmetrical, but I can also just as easily see it being asymmetrical. You just can't pin it down. If it were simple everyone would be making tons of money .

[skellington]

Ah, ok...on to live testing then.