kjeldahl.net

In reading the "results" below, please note the format:

result-love-fXX-hYY[-hutZZ].txt

Where fXX denotes the rebalancing frequency (typically f28 for 28
days), hYY the number of stocks to hold (typically h5 or h10 for
holding top 5 or top 10) and hutZZ for "hold-til-drop" (where ZZ
varies from 15 to 30).

The simulation was run using monthly data and holding period from
1997-08-31 until 2004-12-31.

First, the "hold 5" category, with "hold-til-drop" 15,20,25,30:

Updated 2004-12-16.

Below you can find heatmaps and dendrograms that show the correlation between the different SIPRO based long screens over time. The raw data used is also available below each heatmap.

This time some of the major indices from the stock market have been included. TMWX is the Wilshire 5000, RUT is the Russel 2000 and 1866580 is the Oslo Stock Exchange in Norway (for my curiousity only - I'm based in Oslo).

Monthly top 5 data from 1997-08-31 to 2004-12-10

This includes all screens except Zweig-26.

With this little project I wanted get some hands on experience with market timing. Basically, the strategy is as follows:

Inspired by the numerous discussion about using tecnical analysis (TA) together with mechanical investing (MI) and with the data from the SIPRO backtester, I decided to do some tests and see whether MI with TA may show an improvement compared to "straight" MI.

Below is a (old) perl script I wrote to extract and merge data from the various data tables in SIPRO into simpler to process CSV files:

This document tries to explain how the data in the product Stock
Investor Professional (SIPro - distributed by http://www.aaii.com/) is
built up. In the Mechanical Investing (MI) community it seems most
people are more interested in the raw data rather than the software
distributed with the product.