[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: WHY BACKTESTING WORKS



PureBytes Links

Trading Reference Links

Don't totally discount intuition.

As a trader I am convinced that back tested mechanical trading systems can
work. However, as a doctor I can tell you that one does develop a sixth
sense of what's wrong and what to do about it even before all the tests are
back. This puts you slightly ahead of the heard - be they bacteria or other
traders - and a few minutes or a few hours can sometimes make the
difference. In the scientific world,  of course, one runs all the tests
anyway because sometimes your just flat wrong. Still, my "doctor intuition"
is right about 80% of the time. Many mechanical trading systems could not
claim such a "percent profitable."

I'm certain there are professional traders out there who have the same sort
of sensitivity about the markets. And they clearly have an advantage over
more mechanical traders such as myself.

Clint


-----Original Message-----
From: The Omega Man <editorial@xxxxxxxxxxxxx>
To: Omega List <omega-list@xxxxxxxxxx>
Date: Sunday, October 11, 1998 9:29 AM
Subject: WHY BACKTESTING WORKS


>
>In the October issue of our favorite magazine (Technical Analysis of Stocks
>and Commodities) there is an interview with James O'Shaughnessy.  This
>interview will be of interest to those who have a view on the validity of
>backtesting as a method for developing trading strategies.
>
>The most interesting point made by Mr. O'Shaughnessy is that backtested
>models work best in all fields of human endeavor, not just in trading.
Such
>models, which he calls "quantitative/actuarial" models, consistently
>outperform non-backtested models, which he calls "clinical/intuitive"
>models.  This is as true for doctors attempting diagnoses as it is for
>handicappers picking horses at the track.  It is as true for college
>administrators judging admissions candidates as it is for parole boards
>judging parolees.  It is as true for loan officers doing underwriting as it
>is for traders judging markets.
>
>The key is to choose a model which performs well in backtesting, and then
>*stick to that model*.
>
>
>Good trading,
>
>The Omega Man
>
>
>A is A
>
>
>
>
>
>