Let's assume that it is the end of February 2005 and we are considering a long position in Cotton. Let's use the All-Months Continuous Contracts program to generate the chart below. From this we see that cotton is low with respect to its historical range, generally a good sign for a long position.

Next, the seasonal tendencies are checked using the Multiple-Month Seasonal program. This chart showns that all the contracts have the same basic behavior, especially in March. Notice that in the marked area the May and July contracts show the greatest seasonal movement. Let's zoom in on this area. This chart shows that we should look for a low entry point for the May or July contract starting around the end of February and hold until around March 14.

Let's now examine the reliablility of the May seasonal. For this, we can use the Single-Month Seasonal program which produced the following chart. Since the standard deviation lines are not relatively close to the seasonal during February and March, this time of year will not be the most reliable for the seasonal.

There is another check of seasonal reliablity using the Seasonal Consistency program. This program takes the last 20 years of data (or less if there is not that much data) and divides it into two equal time periods. Then a seasonal is made from each time period and both are plotted. Let's zoom in on the divergence near the end of the seasonals. This chart shows why the reliability of the seasonal is not great. From March 9 until contract expiration, the "Old" seasonal rises while the "New" seasonal falls. This argues against the trade.