Hi,
I did many optimizations using just 10 min bars, then I decided to use bar magnifier for 1 min resolution, and the model still performed pretty well... Then I decided to move it to 1 tick and wow, drastically different results...
So now I am pondering what to do regarding this.. Is tick by tick the most realistic ? running optimizations by the tick takes forever, i can increase it by 20 tick resolution and optimize, then when i backtest it look at 1 tick optimization.. is this right?
Also I am trading aapl here, and there are gaps between bids and ask a lot, and i am not sure what is really optimal here, I am asking what is the "proper" approach to optimizing a strategy and how can i see what the realistic results of a model are in the past?
Optimizations min vs ticks very different results
-
- Posts: 186
- Joined: 10 May 2012
- Has thanked: 14 times
- Been thanked: 11 times
- Andrew MultiCharts
- Posts: 1587
- Joined: 11 Oct 2011
- Has thanked: 931 times
- Been thanked: 559 times
Re: Optimizations min vs ticks very different results
In case of comparing backtesting scenarios to find out the most realistic one, it totally depends on what real-time scenario in terms of chart and strategy settings you would have.
In any case even most precise backtesting (and as it derivate - the optiomization) cannot be 100% realistic. Pleae read this article and all linked articles as well to get complete understanding of how it works in MC.
In any case even most precise backtesting (and as it derivate - the optiomization) cannot be 100% realistic. Pleae read this article and all linked articles as well to get complete understanding of how it works in MC.