maxcos - or how to beat the market

[Vic Norton]

I have been working on an asset allocation theory based on optimizing  
the Sharpe Ratio of a portfolio of exchange traded funds. The title  
page of the website,
   <http://vic.norton.name/finance-math/sospdr/>  or
   <http://home.dacor.net/norton/finance-math/sospdr/>,
reads

            Sharpe-Optimal SPDR Portfolios
                        or
       How to beat the market and sleep well at night


The computations for this work are accomplished by two Octave  
scripts, maxcos.m and soptf.m. The second script is just an  
application of the first.


The maxcos.m script defines a function
   [ cosu, J, p, Q, R ] = maxcos( u, A, J, p, Q, R )
that uses updated QR-decompositions to maximize the function
   cosu(x) = (u' * x)/norm(x, 2), (with norm(u, 2) = 1)
on the convex hull of the columns of A. The maximizing x is returned  
in the form
   x = A(:, J) * p,
where the columns of A(:, J) are linearly independent and the  
coefficients of p are strictly positive and sum to 1. Moreover
   A(:, J) = Q * R
on output.


These scripts are in a "maxcos.zip" file on the website along with  
various test routines, README files, and what not. If anyone is  
interested in checking this stuff out, I'd very much appreciate your  
input.

To be more specific
   1. Goto either of the above URLs.
   2. Click on the title or the picture of the opening page.
   3. Go to the "Computation" paragraph at the bottom of the
      resulting "About" page.
   4. Click on the "maxcos.zip" link.

Regards,

Vic


*---* mailto:address@hidden
|     Sharpe-Optimal SPDR Portfolios
*---* http://vic.norton.name/finance-math/sospdr