Creating artificial data to run backtests on

AFL Programming
1

Hi,

Ultimately I am trying to create lognormally distributed random walk data to run back tests on, but I'm not sure the best way to approach this. Does anybody have any suggestions or references?

Thanks!

2

Can somebody please direct me to the documentation on how Amibroker runs code against the data vectors, i.e., once I start a backtest, how many times does Amibroker execute the code? I remember seeing this documentation years ago but have not been able to find it recently.

Glad to be back in this community.

3

This is a good place to start: Efficient use of multithreading

You can find others by searching the Help for "thread"

Short answer is that during a backtest your Phase 1 AFL will execute once for each symbol in your testing universe, plus one additional time for the second phase of the backtest when the current symbol is set to ~~~EQUITY.

4

It processes entire vector (entire array) at once.
See
https://www.amibroker.com/guide/h_understandafl.html
and

5

You might also find this helpful re artificial data as per subject: Creating synthetic data for leveraged ETFs

6

Thanks for the reply and for the link. I'm thinking there's a way to force Amibroker to use an indicator with a random component (artificial data series in this case) to run multiple permutations to backtest against without having to actually create new series.

7

Yes you can add mtRandom to your formula to add random component
https://www.amibroker.com/f?mtrandom

1 Like
8

This topic was automatically closed 100 days after the last reply. New replies are no longer allowed.