Re: Limited life span of mechanical systems?

["Kent Rollins" <kentr@xxxxxxxxxxxxxx>]

The equity curve filter does work, you Lucky Bastard.  All you do is
maintain a hypothetical curve in which the system takes all trades.  This
curve is the one you create a moving average against.  It only generates
signals when the equity curve is above the moving average.  The real equity
would then be generated by TS or whatever vehicle you're testing with.  It
requires more programming and that's why some people don't understand it,
but it does work and is a valid technique.

Monte Carlo sims can be extrememly misleading with respect to drawdowns
estimations.  With an MC sim, you are taking a series of trades in the order
they were actually traded and scrambling them randomly according to
parameters input by a human.  It's like throwing a performance report in a
blender.  The random order of trades has nothing to do with what occurred in
reality.  In reality, the system may go thru a long phase where it is out of
step with the market thus creating a drawdown.  An MC sim would completely
destroy this string of trades and randomly assemble a string that has no
basis in reality.  MC sims are totally at the mercy of a) the overall
profitability of the system, and b) the parameters input by the person doing
the sim.  Once you understand that, you will understand the limited utility
of them.

Kent


----- Original Message -----
From: "'Lucky Bastard'" <hadrada@xxxxxxxxxxx>
To: "OmegaList" <omega-list@xxxxxxxxxx>
Sent: Sunday, July 14, 2002 11:35 AM
Subject: Re: Limited life span of mechanical systems?


: One thing you can do from the get-go when developing a system is build in
a
: rule to only take trades when the system's equity curve is above a moving
: average of it's equity curve.  This way, when the system goes into a
: drawdown, it will stop itself.  Of course, this becomes another parameter
: that you must consider when deciding whether you've curve fit something.

This is not going to work. It's the old question of : which came first, the
chicken or the egg? Gary covered this in detail before but essentially,
using this filter on an equity curve changes the equity curve,etc.

: Best answers: 1) test on as much data as possible, 2) don't scale too
: aggressively, 3) trade multiple non-correlated systems, 4) be sure you
have
: enough capital to survive MaxDD * 2 and still eat.

Or do a Monte Carlo Simulation to get a realistic probability of drawdown
and then build trhe account size around that.