Seasonality in price: the real cycles and the sold ones

Commodities genuinely follow the calendar, and an entire cottage industry exists to sell you that fact. This lesson is about telling the two apart.

The real cycles

Grain supply arrives once a year at harvest and is stored, drawn down and priced until the next one. US petrol demand rises every summer as the country drives. Gas storage fills through the warm months and drains through winter. These are physical facts, as regular as the seasons that cause them, and module 1's supply-and-inventory model predicts each one.

Why the pattern is already sold

Here is the step the pattern-sellers skip. A cycle everyone can see is a cycle the market prices in advance. Winter gas demand is not news in November - the winter months trade at a premium all year, sitting in plain sight on the curve you learned to read three lessons ago. For a seasonal pattern to make money, it is not enough for winter to come; winter must surprise. The calendar is public; only deviation from it pays.

How the over-fitted version is manufactured

Take twenty years of prices. Test every window of the calendar against every commodity. Publish the combinations that scored well - 'this commodity has risen in this fortnight in 14 of 20 years' - and let the buyer assume the pattern means something. Test enough windows and such records appear by chance alone, in random data, reliably. Notice what the sales page never shows: the pattern's performance after publication, on a back-adjusted chart done honestly, with the roll costs included.

The reading skill

• Ask what physical mechanism the pattern claims. 'Petrol before driving season' names one; 'gold in September' names a coincidence count.
• Ask whether the mechanism is already on the curve. If the winter premium is visible in July, the seasonal 'trade' was priced before you arrived.
• Treat seasonality as context, the way this module treats open interest: it tells you what normal looks like for the date, so that deviation from normal - the thing that actually pays attention - stands out.

The honest use of seasonality is a baseline, not a signal. Gas drawing down in January is not information. Gas failing to draw down in January is.