Seasonality and Inventory Cycles

Averages are not a fill. They are a prior.

Intermediate 6 min read Course 30 of 60

Course 30 of 60 in the futures hub. A seasonal chart smooths many years into one path for the month. How much weight that path deserves depends on how many years are in it and what this year's inventory looks like.

A seasonal average is a prior

The average of old Julys is background. It says what the month did inside whatever sample the chart used. Buying because this July ought to follow a green average means you're trading a drawing. Weather, a policy shift, or a stockpile the average has already digested can delete the pattern for the only year you have a position on.

Ask any seasonal chart how many years went the other way. If it won't say, it doesn't get a ticket. By seasonality I mean harvests, heating demand and driving demand, and even those bend.

1. What seasonality is

Seasonality is a tendency lined up on the calendar. Crops come off in known months. Heating demand clusters when it's cold. Driving demand clusters when people drive more. Stack the years, average the path, and the smooth line on the diagram is what you get. The line is a cartoon of that sample. Your order still has to be a price.

This year has its own inventory, and the stop still has to be a live order. The average won't close the trade when the weather goes the wrong way. If the chart hides the sample size, treat the count as unknown and the prior as wide. A wide prior gets a smaller position, or none, until the footnote shows up.

Seasonality is a prior from history — not a fill sampleN yearsaverage patha cartoonthis yearinventory existsyour stopstill required

2. How small the sample is

A stack of twenty Julys is the diagram's picture of a real sample, and one violent year inside it can still yank the average around and then hide there. Ten years is thinner than the smooth line looks. An anecdote about last July is thinner than ten. A blog line that says the month always rallies has no count under it. Ranked honestly, the slogan sits under the anecdote, and the anecdote sits under a sample you can count.

Seasonal edge 8 ticks. Costs plus the stop come to 12 ticks. 8 is smaller than 12, so the average move doesn't pay for being wrong. Leave the $205 unspent.

An edge smaller than its own stop and costs doesn't pay for the times you're wrong. Before I'd spend any of that $205, I'd want the stop to cost less than the move I'm counting on, in a market I have a reason to trade besides the name of the month. If the month is the whole reason, I keep the money.

Twenty years of Julys is not a law anecdote a real sample a slogan thin sample → wide prior → smaller size

3. Inventory against the calendar

The calendar and the stocks on hand argue, and the stocks should be allowed to win. A heating-season bias sitting on full tanks is last year's weather laid over this year's inventory. The tanks are what's actually there this year.

Grains get a public version of that argument each month. USDA releases WASDE at 12:00 pm Eastern. A harvest pattern with no slot for that print is a calendar that left the bins out. I won't guess the stocks ahead of the release. The report prints them. Until it does, a seasonal lean is a reason to watch the market at a smaller size, and a reason to be flat if you can't sit through the number.

4. How to use a seasonal chart

Use the chart as a list of markets that deserve a look this month. It can point you at grains around harvest, or at fuels when heating demand usually shows up. The look is the start. The trade still needs a price, a stop, and a size that survives the stop being hit.

One use I trust is a haircut on size: less risk in the month the seasonal leans against you, and no extra risk merely because the month leans with you. The diagram calls that a bias haircut. Any entry still comes from the price in front of you this year. Using the green month as your entry is trading the average, and nobody fills you at the average.

5. How seasonal charts get misused

Skip the blog that says a month always rallies. The average contains the year that went limit-up, so the smooth line looks gentler than the path anyone actually lived. It also contains the years that did the opposite, which is why a stop still exists. Inventory can flip the sign of the whole month. When the stocks and the seasonal disagree, I side with the stocks or with cash.

Don't buy a seasonal long and then call it protection because the month looks friendly. Protection is for a risk you already hold, and the hedging lesson covers that. 'I need less risk' and 'I like this month' are different orders, and only the second one is a seasonal trade.

Do not buy a month because a blog says it 'always rallies' Averages are not fills Inventory can invert a seasonal.A limit-up year is in the average.Use seasonality as a bias haircut, not an entry.

6. Where seasonal trades go wrong

Buying the green average. Sizing as if eight ticks of history could cover twelve ticks of stop and cost. Ignoring a WASDE week. Treating ten years, or a pile of Julys, as proof the next one will behave.

Once the edge is thinner than the stop, the remaining question is how small the position should be, and a lot of these ideas should come out at zero contracts. The position-sizing lesson does that arithmetic.

Seasonal patterns fail often enough to matter

Some of these patterns are real in the history. Harvest, heating, driving demand: the reasons are physical, and people trust them too far because of that. They also fail often enough to empty an account that treats the average as a rule. I don't trust a win rate built by counting a decade of Julys and stopping.

Under a stable-looking chart the market still changes. A new supply source, a policy, a shift in where the crop is grown. The old path can describe the sample honestly and describe the contract in front of you badly. The seasonal earns a glance and, sometimes, a smaller size. The click still has to come from somewhere else.

How much weight a seasonal chart deserves

  • Read the sample size before you trust the slope. Ten years is thin, and one limit year hides inside the average.
  • Eight ticks of hoped-for edge against 12 ticks of stop and cost means the $205 stays in the account.
  • This year's inventory outranks the month. For grains, WASDE prints at 12:00 pm Eastern.
  • A seasonal chart can tell you where to look. Price, stop, and size are a separate decision.

If the footnote won't say how many years went the wrong way, leave the seasonal chart on the wall. Educational only. Not a recommendation to buy, sell, or hold any contract or pair.