|
PureBytes Links
Trading Reference Links
|
Lionel:
I don't know if Ralph Vince's work is derived from gambling, but he
frequently uses gambling analogies because of their simplicity (ie. known
probabilities, known rewards). Optimal f handles the greater complexities,
though, of variations in returns, changing probabilities, multiple systems
and diversification -- you just have to start with backtest results and
then rerun your calculations, ideally, after each trade.
Ralph Vince points out that you don't have to trade the optimal number of
shares or contracts (in which case you are guaranteed to suffer at least the
maximum drawdown identified during backtesting). You can consciously choose
75% or some other percentage of optimal, and just accept a lower geometric
mean return as the price you pay for sleeping nights. The neat thing is, it
is an informed decision because you can calculate and actually see what the
effect will be on your account equity.
I find it all quite fascinating, but like Walter pointed out, it sure
doesn't help when you're locked limit down for three days straight and you
realize stops don't get you out of trades like during backtesting.
----- Original Message -----
From: Lionel Issen <lissen@xxxxxxxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: July 2, 1999 11:50
Subject: Re: Optimal f
> If I remember correcly, the optimal f is an estimate of what fraction of
> your money you should risk on each trade. Perhaps Vince couold add some
> clarification that would make it easier to use.
>
> Many of these kind of ideas are derived from gambling, where the gambler
> makes repeated tries with a known probability of win or loss. With the
> security market the odds on each trade are not known in advance, it is
only
> after a number of trades that anyone can make an estimate of their win
loss
> ratio.
>
> Try reviewing your last 50 or 100 trades and you may be able to estimate
> your odds, and so estimate what fraction of your money you should risk on
> each trade.
>
> Lionel
> ----- Original Message -----
> From: Walter Lake <wlake@xxxxxxxxx>
> To: Metastock bulletin board <metastock@xxxxxxxxxxxxx>
> Sent: Friday, July 02, 1999 12:34 PM
> Subject: Optimal f
>
>
> > Hi Glen
> >
> > I respect Rick's opinion's ... but you might want to check out what
> William
> > Gallacher has to say about Optimal f in "Winner Take All" to see what a
> > non-enthusiast is saying.
> >
> > Page 194:
> >
> > "... While I do not question Vince's mathematical integrity, I seriously
> > question whether anything he proposes is remotely applicable in the real
> > world. ..."
> >
> > after 5 pages of additional analysis he sums up with:
> >
> > "... Conclusion: Optimal f = phoney optimization."
> >
> > Personally, I don't know what to think. I sounds good on paper but I
> don't
> > know how to make it work with actual data. Maybe you could help me out.
> >
> > Best regards
> >
> > Walter
> >
> >
> >
>
>
|