|
PureBytes Links
Trading Reference Links
|
Adam Hardy wrote:
>Is anyone out there interested in 'optimal f' or 'secure f'
>fractional money management techniques?
I wrote some software once to examine them in a Monte Carlo fashion.
>I tend to follow Ralph Vince and optimal f up to the point where
>Vince takes the 'biggest loss' from the trade history to which he
>applies the formula.
Two problems with this approach:
1. There is no accounting for comfort level. Applying the biggest
loss in the history to the formula may well mean periods of 80%
drawdown.
2. The sequence of trades is assumed to be unchanged. The same
trades in a different order could result in different drawdowns.
The same trades selected at random over a longer period gives a
different result. Monte Carlo simulation gives you a feel for the
mean and standard deviation for max drawdown and expected equity
growth.
>Alot of persuasive opinions on the net warn about optimal f being
>untradeable and very risky, and I agree with that except that
>mostly optimal f is completely rejected instead of being modified.
The way I modified it is like this: Say I'm comfortable with 20%
drawdowns. I run my sequence of trades through a Monte Carlo
analysis, adjusting my fraction of equity invested until the mean
drawdown is 20%. I then look at the variability of this drawdown
and decide if it's acceptable. Then I look at the expected profit
and ITS variability to see if it's worth the drawdown.
>The secure f solution is the only published improvement on Vince's
>optimal f, but I'm not happy with that either, since it also uses
>the trade history to determine the formula, or worse, advocates
>using just some figure that the trader would be comfortable with
>regarding drawdown.
That's not worse, that's better. Psychology is a powerful thing.
How willing will you REALLY be to continue trading a strategy that
has drawn down your account 80%?
--
,|___ Alex Matulich -- alex@xxxxxxxxxxxxxx
// +__> Director of Research and Development
// \ Unicorn Research Corporation -- http://unicorn.us.com
// __) HTML FORMATTED MAIL SENT HERE WILL BE REJECTED AS SPAM.
|