Energy moves together, metals mostly do

Three positions - long crude, long heating oil, long gas. Three tickets, three stops, three tidy risk budgets. One OPEC headline later, three losses. This lesson is about counting to one honestly.

Why the clusters exist

Commodity correlation is not statistical coincidence; it has plumbing. Heating oil and petrol are made from crude - their prices are chained to their input by refinery economics. Gas and power substitute for each other where both burn. The grains compete for the same acres, so a corn shock rewrites wheat's planting maths a season out. Where the physical connections run, the prices follow - which makes commodity clusters more durable than most, and more honest about their causes.

The clusters, roughly

• Energy: crude, its products, and gas at one remove - the tightest cluster, sharing OPEC, inventories and the same demand cycle.
• Precious metals: gold and silver mostly move together, silver with the volume turned up - though gold's currency half, from module 1, can decouple it when rates dominate.
• Grains: corn, wheat and the oilseeds, tied by acreage, feed demand and the same hemisphere's weather.
• Across clusters, and to the dollar: looser, regime-shaped, and never zero - every dollar-quoted commodity shares that one driver.

Counting exposure honestly

The discipline is the one the indices track built, applied with better justification: risk budgets belong to clusters, not tickets. Three energy longs inside one budget is one position at triple size wearing three names - and the cluster's shared gap events, lesson 4's reports, arrive for all three at the same minute. Mostly, though: the clusters are tendencies with physical reasons, not laws. Measure the pair you actually hold, over a stated window, with the correlation tool this platform already carries - the FX track's lesson that correlation is a window, not a property, transfers here without edits.