The market has no single price for a commodity — it has a price for every future delivery date. The shape of those prices, not their average, is what a dated position is actually long or short.
The market does not have one price for a commodity. It has a price for the commodity delivered next month, a different price for it delivered six months out, and a different price for every date in between. Those prices, laid out by delivery date, are the forward curve — and the differences between them are not a detail laid on top of "the price." For a physical trader they are frequently the whole of the position.
This article sets out what the forward curve is and why it matters: why a trade's delivery date is part of its price rather than a label on it, why the shape of the curve is the market pricing time rather than noise, why a book that is flat on quantity can still be long or short the shape, and why a hedge only protects the date it sits on. It is written for the people who carry the consequences: traders whose hedges roll from month to month, the middle office who mark dated positions, and the risk function who report the result.
If the curve is new to you, start with A price for every date. If you already mark against a curve and want the part that moves quietly, start with How the shape moves.
There is no "the price" — the curve is a price for every future delivery date
A commodity has as many prices as it has delivery dates, and a trade is not complete until its date is attached.
The number a screen shows for "the price" of crude or gas or metal is usually the price for the nearest delivery, the prompt month. It is one point on a curve, not the curve. The same grade, at the same place, against the same counterparty, prices differently for delivery next month, the month after, and a year out. Those are not one instrument at several discounts; they are several instruments, each with its own bid, offer and history.
A trade therefore carries its date as part of its price. "Buying oil" is not a complete instruction until it says for which month, and the month selects which point on the curve the trade sits on. Two trades of the same grade, same volume, opposite direction, on different months, do not offset each other. They are long one point and short another, and the gap between those two points is the position.
The consequence is worth stating plainly: there is no number a position is up or down against until you say which date it lives on. The level of the market is an abstraction built from the curve. A position is built from specific points on it.
The shape is not noise — it is the market pricing time
The gaps between the points on the curve are prices, not a byproduct. Whether later delivery costs more or less than prompt is the market's answer to a real question: what is it worth to have the commodity now rather than later.
When each later month prices above the one before it, the curve is in contango — the market is saying that carrying the commodity forward costs something, in storage, financing and insurance. When prompt prices above the deferred months, the curve is in backwardation — the market is saying that having the commodity now is worth more than having it later, which is what scarcity and immediate demand look like.
| Contango | Backwardation | |
|---|---|---|
| The shape | Each later month prices above the one before it | Prompt prices above the deferred months |
| What the market is pricing | The cost of carrying the commodity forward: storage, financing, insurance | The value of having it now rather than later: scarcity, immediate demand |
| What it is not | A forecast that prices will rise | A forecast that prices will fall |
Neither shape is good or bad in itself. Both are ordinary states of a physical market, and the same commodity can be in one shape at the front of the curve and the other at the back. What matters is that the shape is a source of profit and loss that is separate from the level. A market can sit flat — the average of all its points going nowhere — while the shape swings between contango and backwardation, and that swing moves every calendar spread on the book.
This is the first place the distinction bites. Reading "the market" as one number hides exactly the part a dated position is exposed to. Two dates can move in opposite directions in the same day: prompt can rally while the back of the curve falls. Someone marked to "the market" sees nothing. Someone marked to those two dates sees the whole event.
A position is a point on the curve, and a spread is the gap between two points
Being flat is a statement about dates, not just about direction. Quantities that cancel can still leave you long or short the shape of the curve.
Take a purchase for delivery in December and a sale of the same volume for delivery in June. Same grade, same place, opposite direction. On quantity the book is flat. On the curve it is long December and short June — a calendar spread. If December rallies against June, the position makes money; if it falls against June, it loses. Neither outcome touches the flat quantity.
This is a trap that recurs throughout physical trading, wearing a different mask each time: aggregating across two different things and calling the result flat. Here the two different things are two delivery dates. The unit of a curve position is a date, or a pair of dates, and any report that sums across dates and prints one number has discarded the position while keeping the volume.
The useful test is the same one that works elsewhere: ask what single trade would close the reported position. If the answer is "buy the same grade for the same month," the netting was legitimate. If the answer is "several, across different months," the report has aggregated across a spread rather than netted within one.
Only a few points on the curve trade — the rest are filled in
The curve is assembled, not observed. A handful of liquid months trade; the points between and beyond them are placed by assumption.
A delivery month far out, or an odd month between two liquid ones, can pass a whole day with no transaction at all, yet the curve report shows a price for it. That price was not observed. It was constructed — interpolated between the months that did trade, extrapolated beyond the last liquid point, or carried over from a relationship that held the last time anyone checked.
This is not a criticism of the people who build curves; it is what a curve has to be in a market where only some dates ever trade. The thing worth knowing is what it makes the curve: a mixture of measurements and assumptions, in unknown proportion, wearing one shape. The liquid points are measurements. The gaps are models. Nothing about the line that connects them shows where one ends and the other begins.
The consequence appears when the shape moves. A steepening at the front of the curve can be a genuine re-pricing of scarcity. The same steepening three years out can be the interpolation quietly responding to one traded month. The two look identical on a chart and mean completely different things about how much of the exposure is real.
The curve moves by reshaping, not just by shifting
A curve does not move as one number. It shifts, steepens, flattens and twists, and different positions feel each of those differently.
A parallel shift — every point up by the same amount — is the only move a single "market level" can describe, and it is the least typical of the ways a curve actually changes. More often the near months move more than the far ones, or the front steepens while the back flattens, or the two ends move in opposite directions. Each is a different kind of event, and each hits a different part of the book.
Two books can be flat on the level and opposite on the shape. One is long prompt and short the back; the other is short prompt and long the back. If the curve flattens, one gains and the other loses while the average of the whole curve has barely moved. A report that marks both against "the market" shows both unchanged and both wrong.
That is why the shape has to be a first-class object rather than an afterthought. The level is what the screen reports. The shape is what a dated book actually holds, and the shape is where the quieter, more expensive surprises live.
A hedge protects the date it sits on, and rolling it is a trade on the shape
A futures hedge offsets the delivery month it is written for. Everything between that month and the position's own month is a spread the trader is still carrying.
Hedge a December physical position with a December future and the two sit on the same point: they offset. Hedge it with a January future — because the December contract is too thin, or because the hedge went on before the date was final — and the position has become long December and short January. The hedge removed the level and installed a calendar spread.
The same mismatch appears in crude before any roll: a cargo priced over a window that straddles two contract months is dated to two points of the curve at once, while its hedge is dated to one. Crude Oil Trading: Why the Barrel You Hedge Is Not the Barrel You Own
The same thing happens automatically when a hedge rolls. As the prompt contract approaches expiry, the hedge is moved out to the next month. That roll is not housekeeping; it is two trades that realise the spread between the old month and the new month. When the shape between those two months has moved, the roll turns a shape move into realised cash — which is why a strategy that looks flat on the level can still produce a profit or loss every month.
The practical point is not that rolling is wrong. It is that the roll is a position in the calendar spread and deserves to be seen as one. Left as "just rolling the hedge," it is the part of the book that changes every month with no one watching the shape it depends on.
A position report that shows one number has thrown away the date
A curve-aware position has to carry its dates, and a report of it has to carry the shape, or the hedge cannot be matched to the thing it hedges.
A single mark per grade answers "what is the level." It does not answer which point of the curve a given trade sits on, which point its hedge sits on, or how far apart the two are. Those are the facts the hedge is built from, and they are not recoverable from an average.
| Field | Why it is load-bearing |
|---|---|
| The delivery month of each physical trade | Selects which point on the curve the trade is on |
| The contract month of each hedge | Selects which point the hedge is on |
| The spread between each pair meant to offset | Shows whether a hedge is on the same point or a different one |
| Which points of the curve were observed and which were filled in | Distinguishes measurement from assumption before the shape moves |
| The roll date and the two months it connects | Makes the calendar-spread trade visible instead of routine |
Laid out by date, the same book that reads as "flat on the level" shows its actual contents: long here, short there, and a set of spreads whose behaviour is decided by the shape of the curve rather than by its average.
Questions people ask about this
What exactly is a forward curve?
It is the set of prices for the same commodity at a series of future delivery dates, ordered by date. Each point is the price of delivering the commodity at that date, and the points are not separate numbers so much as one structure — the line that connects them is the curve. When someone says "the market," they usually mean one point on it, most often the nearest delivery.
Is contango good or bad for a physical trader?
Neither, by itself. Contango means later delivery prices above prompt; backwardation means prompt prices above the deferred months. Each is the market pricing a real thing — the cost of carrying a commodity forward, or the value of having it now. What matters is not the shape but where your own position sits on it: a shape that costs one book money is making money for the book on the other side.
Why does my profit and loss move when the market looks flat?
Because "the market looks flat" means the average of the curve is flat, and your position is not on the average. It is on specific dates. When the near months move against the far ones, or the front steepens while the back does not, the average can stay put while your calendar spreads all move. The number that looks flat is simply not the number your book is exposed to.
We are flat in quantity — why do we still carry exposure?
Because quantity and dates are different dimensions, and flatness in one says nothing about the other. A purchase for one month against a sale of the same volume for another month is flat on quantity and long one month against the other. The gap between the two is a calendar spread, and it moves with the shape of the curve even when the total volume never changes.
When does a hedge actually offset a physical position?
When it sits on the same delivery month as the physical trade, so that both are the same point on the curve. The moment the hedge is written for a different month, the offset is incomplete: the hedge has removed the level and left behind the spread between the two months. That spread is a position, and it is managed or not depending on whether anyone can see it.
What is actually happening when we roll a hedge?
The hedge is being moved from one delivery month to a later one, which means closing the old month and opening the new. That is a trade in the spread between the two months, not a neutral bit of housekeeping. If the shape of the curve between those two months has moved, the roll realises that move as cash. A book that rolls every month is therefore taking a position in the calendar spread over and over, whether or not anyone is measuring it.
How much of the curve we mark against is actually observed?
Far fewer points than the curve shows. A handful of liquid months trade regularly; the months between and beyond them are placed by interpolation and extrapolation rather than by transactions. The curve is a mixture of measurements and assumptions, and nothing in the line tells you where one becomes the other. The useful question to ask of your own curve is which points were observed and which were filled in, because the filled-in ones inherit the health of the relationship behind them.
Where should we start if we have never looked at the curve?
Take the open book and attach a delivery month to every physical trade and a contract month to every hedge. Nothing else is needed for a first view. Laid out by date, that list shows which trades are meant to offset which and how far apart each pair actually is — and it usually shows at least one place where a hedge that looked matched is sitting on a different month from the thing it is supposed to protect.
Where this lands in a trading system
Nothing above is an argument for buying software. It is an argument about which facts a dated position is made of and where they are kept. Parts of it map onto system capabilities and parts of it do not.
The first conclusion is that a physical trade and the hedge against it have to be held together, with their dates, before anyone can see whether they actually offset. That is the territory of a CTRM/ETRM system providing complete trading lifecycle management for commodity traders, integrating physical trade, financial hedging, risk control, and settlement in one platform; Time Dynamics' Fusion is one, and the parts of it that bear on this article are that complete physical trade lifecycle management and execution and advanced trading management for futures and derivatives markets sit in the same place, so the December purchase and the December hedge can be compared as the same point rather than as two unrelated entries.
The second conclusion is that the curve itself, and the marks that place a trade on it, usually live outside the trading system — in a price feed, a curve built in a spreadsheet, a broker's indication. That is what X-Ray is for: a non-invasive data processing and analysis platform, whose collection toolkit XDK does non-invasive automated data collection from databases, Excel files, and web interfaces, and which is built to collect data without disrupting existing systems or workflows. The last property is the relevant one, because the curve somebody maintains in a spreadsheet is working, and a plan that begins by taking it away will not survive the first volatile month.
The third conclusion is that the same book has to be read by date and not just by quantity, and that the dated view is the one most often rebuilt by hand. X-Sheet does one-click generation of personalized reports with Excel-like interface.
And here is where it stops. No system on this page decides the shape of the curve. It will not tell you whether prompt will trade above or below the back of the curve, it will not pick the perfect month to hedge into, and it does not make two months trade when only one of them does. It has no view on whether contango is about to break, and it cannot turn an interpolated mark into an observed one. What a system can settle is narrower and duller than the problem in this article: whether the delivery date of every trade and the contract month of every hedge are stored where they can be compared, whether the spread between them is visible, and whether the person who needs the shape can see it without rebuilding it by hand. That is a real part of the gap. It is not the whole of it, and it is worth being clear about which part you are buying.
A note on the evidence in this article
This article contains no statistics, and that is deliberate. The observations behind it — which points on a curve actually trade, how the rest are filled in, and where a dated position's exposure lives — come from material we are not in a position to publish as a citable figure. A number nobody could check would make the argument look stronger and would make it worth less.
So the argument runs on mechanism, and each claim is written to be tested against your own book rather than taken on our authority. Three tests are proposed above and they are the intended use: attach a delivery month to every trade and a contract month to every hedge and see which pairs actually coincide; ask of any "flat" figure what single trade would close it; and ask of your own curve which points were observed and which were filled in. If your own book contradicts what is written here, your book is the better evidence.
The level is what the screen reports. The shape is what your book holds.