Thanks Mike
I tried kelly and it misleads people.
later I began going with van thorp methods.
rvlv
--- On Thu, 11/19/09, Mike <sfclimb...@yahoo.com> wrote:
From: Mike <sfclimb...@yahoo.com>
Subject: [amibroker] Re: Very useful for capital efficiency -- of great help
To: amibroker@yahoogroups.com
Date: Thursday, November 19, 2009, 12:58 PM
Ralph Vince says in "The Handbook of Portfolio Mathematics" that the Kelly
Criterion is not suitable for trading because it assumes that all wins will be
of the same size and that all losses will be of the same size, which simply is
not the case in trading.
Mike
--- In amibro...@yahoogrou ps.com, Deepak Patade <iamdeepakpatade@ ...> wrote:
>
> *I read this in a book and just pasting it here as it is*
> *can be of use for efficient use of capital*
> *very useful*
> **
> *""" I was using something called the Kelly formula that Ralph Vince*
>
> *had popularized. I found the Kelly formula in a little blackjack*
>
> *book, and Ralph and I started fooling around with it and it led to*
>
> *the whole thing of money management. Actually J.L. Kelly Jr. originally*
>
> *devised the formula for the flow of electrons through circuitry.*
>
> *But he said you could apply it to blackjack betting to measure risk/*
>
> *rewards, too, and then we applied it to trading. It's very wild and it*
>
> *will exponentially increase the number of contracts you're trading*
>
> *when you're winning, so you see that we stepped up from 1 lot to 30*
>
> *lots rapidly. But when you lose money you've got to step back down.*
>
> *I use a variation of that in my own trading today.*
>
> * *
>
> Can you explain the specific formula?
>
> *Let's put F as amount of capital to trade. The amount of capital*
>
> *equals the payoff ratio — that's the risk/reward ratio — plus 1,*
>
> *divided by probability — the system accuracy — minus 1.*
>
> *So F =* *(payoff ratio *+ *1) */ *(probability *- *1)*
>
> *I did some tables on it. Let's say your system is 40% correct and has*
>
> *a 2.5 risk/reward ratio, you'll use 16% of your money. If your system*
>
> *is 63% correct with a 2.5 risk reward/ratio, you'll use 55% of your*
>
> *money on every trade. So it varies by accuracy. It's a matrix of the*
> *accuracy versus a risk/reward ratio.*
> **
> **
> *can we have a small afl for this*
>
