Research tends to be a process of starting wide and then narrowing in.  We see something universal and then we start considering more specific applications via filters, qualifiers, etc. 

The past few postings have shown how the day of the week can be a clue in determining likely open-to-close direction.  Since we've demonstrated many other elemental biases over the life of this column as well as success in combining them, the following question might jump out at us: could we improve the day-of-the-week indicator by ignoring days where other of our signals are contrary?

Figure 1 shows the results of following the-day-of-the-week indicator unless the majority of our original combined 3 indicators are in opposition.  As you may recall, the three indicators are (for buys):
1. The 2-day close is lower than the 5-day close.
2. The close is greater than the previous 40-day close.
3. The highest close of the last 50 days comes before the lowest close of the last 50 days.
(And all three denote short sales where the opposite is true).

Clearly an improvement, as you can see in the average trade and percent profitability columns.  We'll check out other filters tomorrow.

Art Collins is the author of Market Beaters, a collection of interviews with renowned mechanical traders. He is currently working on a second volume. E-mail Art at artcollins@ameritech.net.