Commentary, Methodology

10/8/20 – New Methodology

The trend following strategy is em3.

That 5451 number for SOXL is difficult to grasp, especially when you consider how annoying the return stream is. Seems to be right though.

The consenso strategies con1 and con2 derive the signal by combing the signals of all the other strategies and then following the resulting signal.

I probably need to find a new name for tot, that stands for total but also means dead in German. Tot plays by counting how many of the 26 other strategies have buy signals on a given day. If more than 11 are long, it goes long.

Tot also runs on the trend following side. It also does quite well there. TotTF goes long if more than 5 of the 26 strategies are in trend following mode.

Today, for example, only 2 strategies are in trend following mode, so no trend following. However 22 of the strategies have buy signals so it is long because of that.

I’ll be away for the rest of the day.

2 thoughts on “10/8/20 – New Methodology”

  1. Can this method be applied for midterm instead of daily? executing on daily signal is a bit challenging. A weekly signal will be easier to follow and execute from operational perspective.

  2. When I formulated finite state accounting, i was downloading weekly data as often as daily. I stopped looking at weekly when I decided to look at Specter closely. 3 week combinations probably work okay, maybe 2,3, and 4 week combinations.. For longer timeframes, the individual equities moving averages become more attractive. FF5 works on both Specter and Vanilla moving averages.

    I might test that if I get time.

Leave a Reply