|
PureBytes Links
Trading Reference Links
|
Ross:
> Question: does he have the correct formula for the Kelly Criterion in
> Chapter 7? Kelly % = A - [(1-A)/B]
Yes and no. It is correct, but the Kelly formula was created assuming
only two possible outcomes -- a win and a loss of equal size (a
Bernoulli distribution). Trading results more closely resemble the bell-
shaped Normal distribution, albeit with fatter tails. Because of this, the
Kelly formula was modified to use average wins and average losses.
But, since these are only approximations of the distribution, your bet
size will not be optimal. You could be risking too much or too little.
> I already have a couple of Ralph Vince's books. He is tough to read
> and his books aren't written in a "here is how to apply it" way.
I agree, but you'd be better off reading them again. And again, if you
are still having trouble with it. Even if you choose not to use his
Optimal f method, you will be better prepared for assessing other
money management techniques. Attached is an Excel spreadsheet
created by Futures Magazine for computing Optimal f. This used to
be posted on their download site, but I can't find it there now.
Personally, I tore the spreadsheet apart and rebuilt it to suit my own
style and so I could get a better understanding of the formulas.
----- Original Message -----
From: "Ross Kovacs" <rossrk@xxxxxxxxxxxxxx>
To: "Metastock List" <metastock@xxxxxxxxxxxxx>
Sent: Friday, August 10, 2001 2:38 PM
Subject: Excellent web book on Money Management
> I was doing a metasearch of the web for information on the Kelly Criterion
> and stumbled on a web book by Thomas Pflugl:
> http://keplerweb.oeh.uni-linz.ac.at/trading/index.html
> I seem to recall him making various contributions to this list and others.
> He credits this list and others in helping to write the web book (e-book?).
> Regardless, a good read and you can't beat the price (free).
>
> Question: does he have the correct formula for the Kelly Criterion in
> Chapter 7?
>
> Kelly % = A - [(1-A)/B]
>
> where A = % winners and B = Avg Profit per trade / Avg Loss per trade
>
> Also, any feedback on his writings and other money management writings would
> be appreciated.
> I already have a couple of Ralph Vince's books. He is tough to read and his
> books aren't written in a "here is how to apply it" way.
>
> I put the above formula in an Excel spreadsheet, and it convinced me it was
> a good approach (assuming I programmed the spreadsheet correctly ;>). I
> used random numbers to generate the trades, and the criterion worked well
> (as programmed by me, a dubious undertaking).
>
> Any feedback on spreadsheets that apply the Kelly Criterion would also be
> appreciated.
>
> Note for stock traders: this web book uses Currency futures contracts for
> its examples.
> I don't know if you will be as interested as I am if you trade stocks,
>
> Ross Kovacs
Attachment:
Description: "optimalf.zip"
|