Interesting, I never really believed in seasonal trades but there seems to be quite a few opportunities.
Quick dumb question..... why don't you backtest that since 1950 ? Surely you can find basic futures since at least 1980, that'd give you 35 years instead of 15, quite a bit more data to build your idea on....
Fundamentals changed significantly since 1950, and so did seasonals. A simple example:
1950 there were not many exports of soybeans from Brazil , and seasonals were determinded more or less only by the crop cycle in the US. Today Brazil is a large exporter. Seasonals for soybeans now include the South American crop cycle, which obviously is different from the North American one.
Similar changes exist for many commodities. Thus, it makes sense to limit the number of years in backtests for seasonals.
Best regards, Myrrdin
The following 4 users say Thank You to myrrdin for this post:
One of my favorite seasonal trades is the RB-HO at this time of the year. The acual entry data suggested by MRCI for the March contracts is the 17th of November.
Currently this spread shows a very high value. Only once since 2006 the spread showed a higher entry value than this year (it was in 2010). This seems to be caused by the warm weather and corresponding weather forecasts for the foreseeable future.
I intend to wait for a cold snap in the weather forecasts, and then enter this spread. In case there is no such
cold snap, the spread might move upwards without me.
Another option worth considering might be to hedge the RB-HO spreads with some long NGH call options.
Both of those contracts have not been following prior seasonal patterns this year.
Seasonally they usually peak in mid Oct. This year they bottomed in mid Oct. They usually go down the first 2 weeks of Dec. This year they were up.
The spread seasonally rises from mid Oct to early Dec. This year it dropped.
I will say that the spread in Dec has performed seasonally.
I have found that your excellent prior advice that if a contract/spread has not been following normal seasonal patterns that the likelihood of it following the seasonal pattern in the future are small.
Last edited by ron99; December 21st, 2016 at 10:00 AM.
The following user says Thank You to ron99 for this post:
I assume that there is a difference in the seasonal charts we use. MRCI uses the most recent 15 years.
These charts show a high for the seasonal (15 years and 5 years) as well as for the current chart at approx. 10th of October. The spread moves down for seasonal and current chart from this top until end of November. The December high occured 10 days earlier than in the seasonal and was not as significant as usual.
To me the correlation between the seasonal chart and the current chart looks to be reasonable.
Best regards, Myrrdin
The following user says Thank You to myrrdin for this post: