|
PureBytes Links
Trading Reference Links
|
Dennis,
I'm not sure if I believe that you are better off holding
the bet size during a drawdown (not reducing the absolute
$$$). I also perceived that Chuck quite strongly advocated
this approach over the "fixed fraction of equity" approach.
It didn't seem to be a discussion of the cost of lunch -
rather a choice of which resteraunt to eat at.
Chuck's conclusion was that "Even though our system is not a
good one you would think that it would at least be a
breakeven proposition (we haven't included any costs)
because the winners are always equal to the amount at risk
and we win 50% of the time. That sounds like a breakeven
system, doesn't it? But if we employ the popular money
management strategy of risking a fixed percentage of our
current capital we manage to turn the system into a loser.
However, if we risked a fixed dollar amount on each trade
the system results would improve and we would break even."
That is pretty clear to me.
The entire bulletin is at
http://traderclub.com/discus/messages/107/681.html?SundayApr
il3020001039pm
If Aaron is right and results are binomially distributed
and/or if others are right and results are well represented
by Monte Carlo Simulation then the above statement is wrong
for real trading (even without getting to the possible but
highly unlikely 0 winners or 0 losers (only 1 chance in 1000
in Aaron's example.)
Even ignoring statistics, the larger bet for a given
drawdown sometimes makes up for the effect of reducing the
bet size during your drawdowns.
The experiment to test this is to set up your sequence of
real trades and then apply both money management methods
while optimising for the same maximum drawdown. I found
that I would get different results for different trade
sequences but when I apply monte carlo simulation and
scramble the trades then I was always better off reducing
the bet size.
The one thing that concerns me is that monte carlo analysis
and Aaron's binomial distribution may not be representative
of actual trading results. If it isn't representative then
some types of trading may be better pursued without
reducing the bet during drawdowns. Can anyone confirm that
trading results are binomially distributed?
John
----- Original Message -----
From: "DH" <catapult@xxxxxxxxxxxxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Tuesday, August 27, 2002 9:27 AM
Subject: Re: Geometric Capital Growth / Optimal-f
A lot of people missed the original thread on this subject
several years
ago and Chuck's coin flipping example is being taken
somewhat out of
context. The question at the time was what are the pluses
and minuses of
reducing size during a drawdown. The plus (and it's a big
one) is you
reduce your risk of total ruin. The minus is it takes you
longer to
recover from a drawdown and your total profit over time is
lower. If you
have a really good system, that never experiences the mother
of all
drawdowns, you will come out ahead if you increase size,
according to
the fixed fractional model, but never decrease size. Of
course, we can
never count on our systems remaining good into the future so
most of us
will elect to reduce size during a drawdown. We should do
this knowing
there is a financial penalty.
Several people asked me privately to respond to Aaron's
quite correct
statistical analysis below. It's accurate as far as it goes
but I
believe it misses Chuck's point. We aren't asking about a
hypothetical
future where it's possible to have 100% winners. Rather,
let's look
backward at a series of trades that have already occurred.
The question
now is would our account be bigger if we had reduced size
during
drawdowns or if we had not. If you do the sims for a good
system, the
answer is your account is bigger if you do not reduce size.
The
difference between the two values is the "insurance premium"
you pay to
give you some protection against going broke. Most of us,
including me,
gladly pay that premium because survival is more important
than
maximizing profits. However, as in most things, there is no
free lunch
and there is a financial cost for our increased security.
--
Dennis
|