New release of Portfolio Probe
Version 1.03 of Portfolio Probe is now available.
An addition in the new version that we are excited about is the ability to constrain the fraction of variance attributable to each asset.
Risk fraction constraints
One use of these constraints is to create risk parity portfolios. These are asset allocation problems where all asset classes have the same contribution to the variance.
The implementation in Portfolio Probe allows you to constrain the risk fractions however you like -- the contributions don't have to be equal.
A second use of these constraints is to replace constraints on weights -- constraints like all weights must be less than 5%. These are really attempting to constrain risk, but are used because the technology to actually limit the risk has been lacking.
More on this, including the mathematics of attributing variance to assets, is in the recent blog post: http://www.portfolioprobe.com/2011/04/13/unproxying-weight-constraints/
Benchmark weights
Another addition to the new version is an argument for specifying the weights for benchmarks. This addition was necessary because the benchmark weights are required when risk fraction constraints involve a benchmark.
The addition, however, is advantageous even when there are no risk fraction constraints. When you pass benchmark weights in, you no longer need to add the benchmarks to the variance matrix and the expected returns beforehand. Now those benchmarks are automatically added.
The use of benchmarks is now easier and safer.
The User's Manual, of course, includes details of both of these additions: http://www.portfolioprobe.com/wp-content/uploads/2011/04/portfolioprobe_user_manual_ed4.pdf
From the blog
There have been a couple of blog posts since the last newsletter that involved random portfolios.
One highlighted an interesting paper that discusses how to separate skill from luck: http://www.portfolioprobe.com/2011/02/14/a-tangle-of-luck-and-skill/
The other reports on the publication in Quantitative Finance of a paper that uses (naive) random portfolios for performance measurement: http://www.portfolioprobe.com/2011/02/07/dicing-with-the-market/
Previous Newsletter
You can see the previous newsletter at: http://www.portfolioprobe.com/wp-content/uploads/2010/12/pprobe_newsletter_101216.html
Thank you
Thank you for your interest in Portfolio Probe.
2011 April
|