Jun

12

In an idle moment I checked the classic Donchian moving average cross (5/20) over the last 3,000 days and gave it the advantage of only trading the long side. Even so the whipsaws were so many that it produced a loss, and that before commissions and slippage.

But it was interesting to see that this system produces a profit when applied to a stock/bond ratio over the same period, on both the long and short side and in both stocks and bonds. It even beats buy and hold, though I didn't include commissions and slippage.

I mention this because the first just gave a short signal after a long run on the long side. But the second has not.

I don't actually think this is a good system or I wouldn't be posting it. But the result seems kind of interesting and may be food for thought.

J T Holley adds: 

Having read and heard numerous presentations from well known and lesser known Donch's over the last five years it is always amazing to hear them down-speak equities. and if you look, their "diversified approach" index futures tend to be a smaller allocation. Their answers to questioning tends to be:

"Equities tend not to trend."

"The whipsaws don't justify a bigger allocation."

"This is the last frontier in trend following."

Nigel's study shows why they make those replies, but they still try. I often asked them "why not just take them out of your allocations?" with no good answer. I guess it is some Edisonian effort to persist or maybe Vic's often quoted "lose more than they should."

Once the Donch's deviate from the fixed system and try to screen, filter, and curve-fit a new fixed I have often wondered why in their attempt at scientific discovery don't they realize the "Law of Ever Changing," and abort the 5/20 or any variation or breakout in same fashion for another method?

One is only left with the assumption that they too have been Bodysnatched! 


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search