|
Portfolio123
Automatically
Managed Stock Portfolios
I’ve
frequently described how to use Web screening tools to come up with
investment ideas. Here’s good news. I’ve discovered a new website
that takes stock screening to the next level.
The site is
called Portfolio123, and it’s the brainchild of Chicago-based software
engineer Marco Salerno. His goal is to provide investors with tools for
creating automatically managed stock portfolios, and it looks like he
has succeeded. Here’s how it works.
Trading
Systems
You start by setting up a trading
system, which includes a set of rules for adding stocks to a portfolio
and separate rules for selling stocks out of the portfolio.
Once you’ve
defined a trading system, the site uses it to identify an initial list
of stocks to buy. Then, on a rebalancing schedule that you’ve
specified, the site sends you e-mails listing stocks to sell and their
replacements, all based on your trading rules.
Salerno
provides everything you need to define and test new trading systems, but
he also offers predefined trading systems that you can use as is, or
modify to suit your needs.
The site
offers a backtesting feature that you can use to find out how you would
have fared if you had used your trading system to pick stocks over
various periods going back up to three years.
You have to
register to use the Portfolio123, but for now, everything is free. That
won’t last long, however. Salerno plans to start charging for certain
features in a couple of months, although costs and which features will
remain free are still undecided.
Start With
Model Portfolios
The easiest way to get started is by using the pre-defined model
portfolios. The top five portfolios, in terms of returns, are displayed
on the site’s homepage (www.portfolio123.com).
The homepage
shows each portfolio’s total and annualized returns since March
31,2001, which is a far back as Salerno’s database goes. The returns
are impressive. When I looked, a portfolio named “GARP Top Ranked”
was the best performer, up 380% since March 31, 2001, or 68 percent
per-year, on average.
Here’s the
best news. Salerno and his cohorts devised 24 different trading systems,
and the worst performer of the bunch, prosaically dubbed the “Value Lg
Cap LT,” is up 68 percent overall, or 19 percent per year, on average.
To put that into perspective, the S&P 500 Index is roughly flat over
the same period. To make the results as realistic as possible, Salerno
subtracts trading commissions when he computes the portfolio returns.
Here are my
ideas for getting started.
How to Do
It
Select Portfolios (top menu) and then P123
Models to see the complete list of pre-defined model portfolios,
including their performance statistics. Then click on a portfolio name
to see a summary
report listing the portfolio’s detailed performance data. The
summary also shows the portfolio’s top 10 holdings and the 10 most
recent transactions including buys, sells, and stock splits.
Then click on
Holdings (top menu) to see the list of all stocks in the portfolio.
While you’re there, select Asset Allocation on the Select View
dropdown menu to see important information that you’ll need to
evaluate the portfolio. The report vividly shows the distribution of the
portfolio stocks in terms of market capitalization and industry sector.
Ideally, you want the maximum possible diversification in both areas.
Otherwise, the trading system’s historical performance will be skewed.
For instance, small-cap stocks (small companies) outperformed the market
last year, but may not repeat their success in the future. Thus, the
historical results of a portfolio heavily weighted with small-cap stocks
would be misleading.
Reduce
Number of Stocks
A problem that I have with Salerno’s pre-defined portfolios is that
most of them consist of 50 stocks, too many for me to buy. You can
modify a portfolio to include fewer stocks by selecting Portfolio (top
menu) and then clicking on New. That takes you to Step 1 of the
procedure for creating a new portfolio. In Step 1, use the Copy Settings
option at the top to pick the pre-defined trading system of choice, and
then use the Position Sizing option to change the number of stocks in
the portfolio. For instance, you’ll end up with a 20 stock portfolio
if you select 5% for the starting market value percentage of each stock.
Finally, click “Next” and then “Finish.”
Unfortunately,
you lose diversification, and thus increase volatility, when you cut the
number of stocks in a portfolio. So, it’s a good idea to test the
historical performance of any portfolio that you’ve modified.
Backtest
New Rules
That’s easy to do by selecting Simulation
(top menu) and then following the same procedure I outlined for building
the model portfolio. By the way, Salerno suggests a minimum of 15 stocks
in a portfolio, and more is better.
Regular
readers know that I’m not in favor of using mechanical formulas to
pick stocks. But Salerno’s program gives you the option of deleting
any stocks before finalizing a portfolio. He suggests researching each
stock and throwing out any that don’t pass muster before finalizing
the portfolio. Then the program will suggest replacement stocks meeting
the trading system criteria.
Although he
has specialized in financial systems, Salerno is a programmer, not a
professional stock analyst. Nevertheless, he has come up with a
stockpicking tool worth checking out, especially while it’s free.
published 4/18/04 |
