Alternate Systems
As you will recall, the most basic DV-2 system is to fade the daily signal at the close (go-long negative results; go-short positive results). Here, I looked at six alternatives to the basic system, including: 1) waiting for additional days of signal, 2) waiting for a reversionary 'hook' back towards the mean, and 3) leg-in variations on the two. Also, I compared the raw system results with the ultimate reversion system, merely trading the opposite direction of the prior day's close. Clearly, such on and off-systems will perform best for methods that experience sustained moves.

As shown above, the raw system actually performed the best (117% Simple Return, Rank #1), adding significant alpha to just trading the opposite side (84% Return, Rank #2), or any of the alternatives, for that matter. While the first conclusion is key in determining model value, the second was a bit of a surprise. The reason I believe this to be the case, is that the system is so sensitive and short-term oriented, that it cannot capitalize on delayed entries like less frequent reversion swing methods with more aggressive exit targets. Basically, you are always better going all-in on the first trigger.
Indeed, the next best performer (other than just taking the opposite of yesterday's trade), was to trade 50% immediately, then the other 50% of available capital the first day the indicator began to come back to the zero line and holding until it did so (78% Return, Rank #3). However, the severity of draw-downs were somewhat mitigated by this later method, which brings up the next major section on "knowing your trading environment."
When to Trade the DV-2 (or any) System
In the introductory article reference above, some of the key take-aways were:
- Know what environment you are trading in (reverting/ trending).
- Trading methods themselves revert (addition by David V.).
- Track your methodology's equity curve like it was a stock.
In this regard, the second method overlays a Parabolic Switch And Reverse (PSAR) indicator onto the equity curve (and there again, limitless variations are possible). Using that mechanical overlay to turn the DV-2 system on and off, about 80% of gains were preserved while avoiding extended draw-down periods. Due to the extreme whip-saw reversal early last year, using the PSAR to actually reverse signals was not effective. On a longer-time frame; however, I strongly suspect it would have been. Clearly, systems with extended performance epochs will do best with off-on overlays like this.


3 comments:
I'm surprised there were no comments on this most excellent post.
Have you applied DV2 to intraday movements?
No Grain, but I expect that the bounded variety may show promise on hourly bars, or something to that effect.
Post a Comment