Subscribe to List View Past Issues RSS translate
facebook facebook Like 0 Comment 0 twitter
Email not displaying correctly? View it in your browser.


Investment technology for the 21st century.
Version 1.04 is here

There is a number of additions and improvements.  Here are the highlights.

Variance partitions

Probably the most important addition is the ability to have linear constraints on the fraction of portfolio variance attributed to each asset.  For example, you can have sector constraints on variance fraction.  This really does what sector weight constraints are often unconsciously thought to be doing.

The partitions of variance that risk.fraction (and linear constraints) can use has been enlarged.  For example the variance fraction adjusted by the marginal contribution to the benchmark is one possibility.

The correlation between each asset and the portfolio can also be constrained via risk.fraction.

Trading costs

There is a new argument ucost that makes it easier to connect the costs to the utility.  You give ucost a single number that multiplies the computed costs before they enter the utility.  This is new to this release -- it was not in either of the beta releases.

It is now possible to give cost.par a matrix so that each asset may have its own set of exponents.

Valuation

The valuation function has had a major renovation.  The most important changes are:

  • it will return simple or log returns
  • it allows prices to be a three-dimensional array
  • it will give valuations, weights and returns for categories of assets (sectors, for instance)
The last of these is new to this release -- it was not in either beta release.

More details

The Change Log http://www.portfolioprobe.com/user-area/change-log/ contains the full list of changes.

There are some minor backward compatibility issues.  These are discussed in the backward compatibility document http://www.portfolioprobe.com/wp-content/uploads/2012/04/PortfolioProbe1.04_Backward_Compat.pdf

Strengthened QA

The quality assurance regimen was enhanced this time around with the addition of a serious bout of using random tests.  That is, values for the various arguments (constraints, utility, costs) are randomly selected.  Several thousand such problems were generated.

These tests were focused on the new additions -- the area that of course is most likely to exhibit bugs.

You have been sent this email because you have opted-in on our online request form.

Unsubscribe pburns@pburns.seanet.com from this list.

Our mailing address is:
Burns Statistics
4-b Jodrell Road
London, England E3 2LA
United Kingdom

Add us to your address book

Copyright (C) 2012 Burns Statistics All rights reserved.

Forward this email to a friend
Update your profile
Email Marketing Powered by MailChimp