Greed and Fear - Daily trading signals based on mathematics and software, no opinion, no emotion, no ego. Verified trading results!

The Greed and Fear indicator had its first appearance on this website in late September 2010. Previous to that, I have been working on the software that produces the indicator signals. Every new version of the software was tested by 'replaying' history, starting on 1-1-2009, going forward to 'now'.

As the reader can find on other sections of this website, the Greed and Fear indicator is a daily indicator to go long or short on the S&P 500 (also referred to as a long-short model). To compare the performance of the indicator with the index, a chart is made where both (G&F and S&P 500) start out at the same level (low 900's, January 2009). If the indicator made a correct prediction about the next day, the absolute value of the daily change in the index is added to the total score of the Greed and Fear indicator. If it was wrong, this value will be subtracted from the total score. More details are here: How do the daily signals add up?

I was especially interested to see how the indicator would deal with market extremes. When would it 'realize' the top or bottom was in? Could it keep up with a strong bull market? As you can see in the chart below, I have marked some points of interest. The bottom in March 2009 was detected very accurately, so the indicator switched from being mostly short to being mostly long. But later on, there are some unexplainable drops in performance whereas the S&P 500 itself is not behaving extreme in any way.

There are a few of these areas where the performance suddenly drops for some unknown reason. There is no way to exactly know why that happens, because the internal state of the software is a form of artificial intelligence based on a automated learning process (machine learning).

In September 2010 the test results were good enough to actually apply the Greed and Fear indicator in trading with a very small trading size. That's when the daily updates started to appear on the website. As software development is still going on to find better performance, this chart may be updated when that happens. Every time, history will be replayed to see how the new version performs.

Click the chart for a bigger version.