Path B · Futures GC/MGC · Module B4
Spot gold was a position you could hold indefinitely — pay the swap and it sits there. A futures contract is the opposite: it has an end date stamped on it from the moment it lists. This module follows a position to that end date and shows you what to do before you get there. It's the most operational module in Path B, because the calendar is not optional and the consequences of ignoring it are not abstract.
What this module does
The dated instrument
Go back to what a futures contract actually is, from B1: a binding agreement to exchange a fixed quantity of gold at a future date. That date is not decoration. It's the contract's reason for existing. A GC contract is a promise about a specific delivery month, and once that month arrives the promise comes due — the metal can change hands, and the contract has done its job and ceases to exist. A promise about a date cannot be open-ended; the date is the whole point.
This is the deepest difference between the two instruments, and it's easy to miss because the price on your screen looks just as continuous as spot's. In spot, you held an undated exposure to gold and paid swap for the privilege of holding it indefinitely. In futures, you never hold "gold" in the open-ended sense — you hold a particular contract that is counting down to its own expiry. To keep exposure beyond that expiry, you don't extend the contract. You can't. You close it and open the next one. That act has a name — the roll — and it's the centre of this module.
So the mental model has to change. A futures position is not a thing you set and forget. It's a thing with a shelf life, and staying in the gold market through futures means deliberately stepping from one dated contract to the next, on a schedule the exchange sets and you don't. Miss the step and the contract expires under you — into the delivery obligation B1 warned you to stay clear of. The calendar, therefore, is not background. It's part of the position.
The dates that govern you
Gold futures don't list every month with equal life. Liquidity concentrates in a recurring cycle of even-numbered delivery months — February, April, June, August, October, December — with the nearest of these, the front month, carrying the most volume and the tightest market. That's the contract you'll trade, for the same reason you trade the front month of anything: it's where the participants are.
Three dates govern the end of a contract's life. They're durable relationships, not numbers to memorise — the exchange publishes the exact days, but the structure never changes:
Throughout the contract's life
Front month — peak liquidity
The nearest even-month contract carries the volume. You open and hold here. As expiry nears, participants begin migrating to the next month.
First Notice Day · your real deadline
The last business day of the month before the delivery month
The first day you could be assigned actual delivery of metal. Retail traders must be out — rolled or closed — before this. It arrives earlier than most people expect: for an even-month contract, it lands at the end of the odd month before it. This is the date B1 told you to know.
Last Trading Day
The third-last business day of the delivery month
The last moment the contract trades at all. By now any retail trader is long gone; this date matters to the delivery process, not to you, precisely because First Notice Day already moved you out.
Expiry / delivery
The contract becomes metal — or is gone
The promise comes due. For the few holding to the end, gold is delivered against a depository warrant. For you, the contract simply no longer exists — you left two deadlines ago.
The one date to burn in is First Notice Day, and the trap inside it is the timing: because the even-month contract's First Notice Day falls at the end of the preceding odd month, the deadline to be out of, say, the June contract arrives at the end of May — not in June. People who anchor to the delivery month get caught a month late. Most FCMs will auto-close a lingering retail position before delivery rather than let you stumble into 100 ounces of physical gold, but relying on your broker's safety net is not a plan. Knowing the date is the plan.
Stepping to the next contract
The roll is mechanically simple and worth stating plainly, because the simplicity is the point. To keep a long position in gold past your contract's expiry, you do two trades: you close the front-month contract you're holding, and you open the same-size position in the next delivery month. One contract ends, the next begins, your market exposure continues unbroken. That's the entire operation. There is no special "roll" button required — it's a close and an open, though many platforms offer a combined roll order that does both at once as a calendar spread.
The judgement isn't how to roll; it's when. The signal traders watch is the migration of activity from the front month to the next: as expiry approaches, volume and open interest drain out of the expiring contract and fill up the next one. The practical rule of thumb is to roll once the next month's volume overtakes the current month's — that's the moment liquidity has moved, and rolling into it means tight spreads in the contract you're entering. Roll too early and you're trading a back month before the crowd arrives, in a thinner market. Roll too late and you're stuck in a draining front month with widening spreads, racing First Notice Day. The window is comfortable if you watch for it and unforgiving if you don't.
Notice this is a relationship, not a date: "roll when volume crosses over," not "roll on the 12th." The crossover happens on a different calendar day every cycle, but the signal — back month's volume exceeds front month's, comfortably before First Notice Day — is the same every time. Learn the signal; read the dates each cycle.
What the roll costs
B3 promised this module would take the roll cost apart, so here it is in full. The two contracts you roll between are almost never priced the same. Gold is structurally in contango: each further-out delivery month trades at a slight premium to the nearer one, because its price carries the cost of storing and financing the metal for the extra time until that later delivery. This is not a market quirk — it's the cost of carry, the same economic fact that produced swap in spot, expressed as a shape in the futures curve.
Now watch what that does to a long roll. You close the front month at its (lower) price and open the next month at its (higher) price. You are selling cheap and buying dear, by exactly the carry premium between them. That difference is a real cost, and it has a name: negative roll yield. Hold a long gold position across many rolls and these premiums compound into a meaningful drag — the futures market collecting, roll by roll, the same carrying cost spot collected night by night as swap. This is the precise cash-out of B3's principle that no swap is not no carry. The carry was never gone. It was waiting here, in the curve, to be paid at the roll.
The honest catch — that "gap" on your chart
When a contract rolls, a raw futures chart appears to jump — gold looks like it gapped up or down overnight, even though nothing happened to the actual price of gold. It's an illusion of the data: the chart switched from quoting the expiring contract to quoting the next one, and those two contracts trade at different prices because of contango. The metal didn't move; the contract underneath the chart changed. Platforms paper over this with continuous or back-adjusted charts that smooth the discontinuity for analysis — useful, but know what they're hiding: a back-adjusted chart's older prices are not the prices that actually traded. If you ever reconcile a historical level against a back-adjusted chart and the numbers don't match, this is why. The gap was never real. The adjustment that erases it isn't real either. Both are artefacts of an instrument that lives in dated pieces.
One clarification so you don't over-learn the lesson: contango is the normal state, but not the only one. In moments of acute physical demand gold can flip into backwardation — the nearer contract dearer than the further one — and a long roll then earns a small positive roll yield instead of paying. It's the exception, and you shouldn't build expectations on it, but it's worth knowing the curve has a direction and the direction can change. As ever: we're giving you the relationship — carry shows up in the curve, usually against a long — not a fixed number to expect at any given roll.
Carrying forward
That completes the machinery of Path B. B1 inverted the counterparty, B2 installed the institution that the inversion puts in its place, B3 counted what the new structure costs to transact, and B4 followed a position to the end of its dated life and through the roll. You now understand futures gold as a structure — who stands where, what it costs, and how a position lives and ends. The final module closes the path: it gathers what's durable, states plainly what we will not teach you, and points you to where the road goes next.