|
PureBytes Links
Trading Reference Links
|
John, I think you've gotten to the heart of the matter. You need to
consider both risk and reward, not just reward as Chuck does. Optimizing
your money management for the best Sharpe ratio seems like the way to
go. That will give you good profits and a good comfort level.
An accurate sim should also include the overhead of taxes and periodic
withdrawals for living expenses. The reality is most people, because of
overhead and other withdrawals, trade closer to the constant risk or
constant size models than to the Vince model where we all take marginal
systems and scale up to owning the world in a few short years.
> Can anyone confirm that
> trading results are binomially distributed?
They aren't. The curves have fat tails. Just ask LTCM. :-) I think the
Monte Carlo approach is still a useful research tool but I just ignore
the cases where I flip a coin for 10 winners in a row and assume I'm
going to be the poor schlep with 10 losers in a row. I'll use the MC to
see the worst case scenarios but I mentally fatten the tail and assume
those worst case scenarios are more likely than predicted by the sim.
So, as an example, say your real comfort level is 1% chance of going
bust. Choose your size so the sims say you have a .1% chance of busting
the account.
--
Dennis
|