Trends don't exist until after the fact in non-stationary(random) series, which prices are according to statistical tests.
Now, if you have a stationary/mean reverting series so that a long-term average does exist and the series has finite variance, then you can trade it all day long.
Question is, once you have the stationary form, how to detecting turning points without false positives? Rolling regression looks promising..
Now, if you have a stationary/mean reverting series so that a long-term average does exist and the series has finite variance, then you can trade it all day long.
Question is, once you have the stationary form, how to detecting turning points without false positives? Rolling regression looks promising..
