Screen Results - How to Improve the Performances

I have tried to sort out the results from the lists generated the last 2 weeks by the Combination Screen and here are some thoughts. At first I tried combining all the fundamentals numbers into one rating and applying it to rank the stocks in the list. The first week, it generated results that beat that overall list but not by much. And it did not carry over to this week. I tried using a technical rating based on multiple technical indicators, but so far, I cannot find a correlation. Then I went back to the books that inspired the screen to begin with and remembered the fundamental criteria that was cited more often - Price-to-Sales. As an experiment, I ranked the stocks by the Price-to-Sales ratio in ascending order and analyzed the numbers. Not surprisingly (in hindsight), this makes a big difference! Here are the results:

After the first week, the top 10 stocks in the list (highlighted in yellow in the list) are up an average of 5.93% with only one loser. This beats the market by 3.83%. After 2 weeks, the top 10 are up 6.99%, beating the market by 3.36% so losing some ground. But it is expected after such results in the first week! Still some very big winners and only 2 losers (20%). Not bad... 


Seeing the first week results, I went back to the list generated by the screen last week and ranked the stocks by their Price-to-Sales ratio again and again, they beat the market average.

The overall market was up by 1.49% last week, but the top 10 on the list were up 1.89%. Not as good as the first set of stocks, but still noticeable. But 3 losers out of the 10. 


This is something worth monitoring over the long run! This is the reason why I listed the screen results for this week in ascending order of the Price-to-Sales ratio. I will keep on monitoring the results based on that criteria as well.


Also, the set of data generated by the screen is not representative enough to do a complete analysis so my plan now is to try to find a correlation between fundamental values and performance over 6 months to a year. I will generate a much larger list, track the performance and analyse the results on a monthly basis! More on this in a later post.