Wednesday, June 03, 2009

Detrend Redux

While continuing my review of the Qs Detrend system posted yesterday, I noticed a fundamental mistake in my testing of the original system back in March. That mistake was in not allowing a larger window for the fixed bar exits when I optimized the inputs. Had I expanded the potential range to 20 instead of 10, as I had done, the results would be as shown above. . . reflecting twice as many trades and twice as much equity gain.
And, perhaps surprisingly, given the recent market runup for the past 3 months, the original optimized inputs are identical to the current ones: (2, .75, -1.25, 14, 14).
As per yesterday's post pyramiding is NOT turned on, as reflected in the maximum position sizing in the performance summary.
I've received several emails from readers reporting a respectable performance daytrading the Detrend on 3 and/or 5 minute bars with somewhat different settings for the ES and ER, so I'm encouraged to learn that the system seems relatively robust and applicable to a variety of trading vehicles.

No comments: