New stuff is available on the Portfolio Probe website.

A new version of Portfolio Probe was released a few days ago.

It includes a new constraint for easily specifying the minimum weight that assets may have in a long-only portfolio.

It improves the speed of the randport.eval function (get information about each random portfolio in a random portfolio object).

There are a few minor bug fixes.

But the key motivator was the speeding up of valuation for random portfolios for the purpose of scenario optimization.

There are more details in the user support email: http://www.portfolioprobe.com/wp-content/uploads/2013/04/pprobe_user_130429.html

Installing pprobeSup

There is a new public domain package available of supplemental functionality to Portfolio Probe called pprobeSup.  It is available from the same repository as pprobeData (and Portfolio Probe):

install.packages("pprobeSup", repos="http://www.portfolioprobe.com/R")

pprobeSup version 1.00

There are two functions that are wrappers for functions in the TTR package.  One that gets closing prices for multiple assets and one that performs MACD on multiple assets.  These two functions do not depend on Portfolio Probe.

This package contains the functions from examples that used to be in the Portfolio Probe User's Manual.

It includes a function to confidently select the best selection of assets of a certain size in an optimization.  This is meant for small problems in which getting the best optimization is important.  Small as in selecting 3, 6,  maybe 10 assets out of a universe of 50 to 100.  Though the size of the universe is probably not especially important.

There are a number of functions for scenario optimization.

Scenario optimization

Scenario optimization is useful when the distribution of asset returns is not symmetric and/or there are specific times in the future when we know things will change.  Bond or options portfolios are examples.

The scheme used for scenario optimization is to generate a set of random portfolios that obey the desired constraints, evaluate the utility of each of those and pick the best portfolio.  If we stopped there, we would be extremely lucky to get a good optimization even with a large number of random portfolios.

The scheme does an iterative approach of generating new sets of random portfolios that obey the portfolio constraints plus they are close to the best portfolio found so far.

pprobeSup includes some simple functions for generating scenarios.

There are some utility functions.  Plus it is very easy to write your own utility function to do pretty much whatever you want.

There are some new pages in the Portfolio Probe Cookbook showing how to do scenario optimization: http://www.portfolioprobe.com/user-area/documentation/portfolio-probe-cookbook/3-optimize-trades/scenario-optimization/