T2W Trading Community

Weekday Patterns

Shifting from months down to days, we can see other patterns appear.  Some of them are most obviously related to the monthly ones note previously. Others are less so.  Again, the focus of the study is trading during the 1999 to 2005 span.

If one were to look strictly at the daily market movements, there would be very little to suggest that USD/JPY or any of the JPY-based crosses is anything other than a 50/50 directional prospect for any given weekday.  Things get much more interesting when the filter of the month is applied as well.  Take a look at the table below to see what kind of patterns show up.

Given that August has a strong downward tendency, it is no surprise to see that USD/JPY has three strong downwardly biased weekdays in that month.  January, however, tends to be an up month for the market, but Monday's that month have a definite negative leaning.  On the upside, there are examples of solid weekday biases scattered all over the place.

These sorts of weekday biases in certain months also can be seen in the crosses, in some cases in even stronger fashion.

Application to Trading

Trading based on seasonal or calendar patterns can be tricky.  One can never quite be sure when a pattern is going to fail, or change all together.  As a result, to blindly trade them with no concern as to risk would be quite foolish.  Yes, some of these patterns are very strong.  One who sold USD/JPY at the start of each August and closed out the trade at the end of the month would have made over 2200 pips in the last seven years.  That's a pretty good take, but there are drawdowns.  Sometimes they can be quite large.

Probably the best use of the type of data presented here is in its application to bias one's trading.  For example, a day trader could make use of the fact that GBP/JPY has risen 75% of the time on Tuesday's in August to favor long positions on those days.  A swing trader could look to sell strength or go with downside breaks in USD/JPY during August with a high degree of comfort that the percentages are in one's favor.  And of course a mechanical trader can incorporate the calendar information in to systems.

The point is that data such as the calendar patterns outlined here should be considered another tool at one's disposal.  It is potentially quite valuable, but certainly not an end in and of itself.