The concepts behind the curve

Term structure momentum is a precise idea, not a loose metaphor. Before we discuss how we build analytical frameworks around it, it is worth defining what the term actually means — and what it does not mean. A futures curve is the set of prices for a given commodity, rate, or volatility instrument across successive delivery dates. The shape of that curve — whether near-dated contracts trade above or below deferred contracts, and by how much — changes over time as market participants revise their expectations about supply, demand, storage costs, and risk. Term structure momentum is the observation that changes in that curve shape carry forward-looking information that the direction of the front-month price alone does not. We develop analytical frameworks that make that logic explicit, document its assumptions, and map how the signal behaves across different market regimes and across the three connected futures market areas we cover. Past performance of any analytical framework does not guarantee future results, and we are clear about that boundary throughout our work.
futures term structure curve shape analysis

Roll dynamics, contango, and backwardation — the building blocks of term structure signal construction

Before a practitioner can evaluate a term structure momentum framework, they need a working understanding of the structural mechanics that generate the signals in the first place. Here is how we think about the foundational concepts across the three futures market areas we cover most closely.
The vocabulary of futures term structure is specific. Using it loosely produces analytical frameworks that sound coherent but break down under scrutiny.

Contango describes a market structure in which deferred futures contracts trade at a premium to near-dated contracts. In commodity markets, this typically reflects the cost of carrying physical inventory — storage, insurance, and financing — embedded in the price structure across time. When a market is in contango, a long position in futures experiences negative roll yield as the position is rolled from an expiring near-dated contract into a more expensive deferred contract. The steepness of the contango — how much the curve slopes upward across successive contracts — is itself a signal-bearing variable that changes over time as storage economics and forward demand expectations shift.

Backwardation describes the inverse structure, in which near-dated contracts trade at a premium to deferred contracts. This typically reflects tight near-term supply conditions or a high convenience yield — the implicit benefit of holding physical inventory when supply is scarce. In backwardated markets, a long position benefits from positive roll yield, and the steepness of the backwardation carries information about how acute the near-term supply tightness is perceived to be. Transitions between contango and backwardation — and the momentum of the curve slope through those transitions — are among the most analytically significant events in commodity futures term structure analysis.

In interest rate futures, the equivalent concepts operate differently. The term structure of interest rate futures reflects market expectations about the path of short-term rates across future dates. The slope between near-dated and deferred rate futures contracts embeds a forward rate expectation, and persistent directional shifts in that slope carry momentum signals that operate on different horizons than commodity roll dynamics. The rate curve can steepen, flatten, or invert across different policy regimes, and each of those structural states changes how momentum signals should be constructed and interpreted.

Volatility futures add a third distinct set of term structure mechanics. In calm market periods, volatility term structure typically exhibits persistent contango — near-dated implied volatility sits below longer-dated implied volatility — because buyers of future variance pay a premium for protection against uncertainty further out in time. During periods of acute market stress, this structure can invert sharply as near-dated volatility spikes. That mean-reverting character of volatility term structure interacts with momentum signals in ways that differ substantially from commodity or rate dynamics, and our analytical frameworks treat those differences explicitly rather than applying a single construction logic across all three areas.

Momentum lives in the curve shape

Most analytical approaches to futures momentum start and end with front-month price direction. We start with the curve itself — its slope, curvature, and how those properties shift over time — because that is where the more durable and distinctive signals tend to be found.
The difference is not cosmetic. It changes which contracts are analytically relevant, which roll windows carry signal, and how cross-market comparisons should be constructed. That is the working definition we apply to every framework we publish.
futures curve momentum signal construction review

How we construct term structure momentum signals — the three-layer approach we apply to every framework we publish

Signal identification from curve data, cross-market plausibility review, and stress-testing against historical structural dislocations — each layer has a distinct purpose and each one catches different categories of analytical error.

A momentum signal that survives only the first layer of review is not a framework. It is a data artefact waiting for a live market to disprove it.

The first layer is signal identification from curve data. This involves specifying the exact curve inputs — which contract months, which roll window, which measure of slope or curvature — and constructing the momentum signal from those inputs with explicit lookback window assumptions. The lookback window choice is not arbitrary; it reflects a view about the persistence horizon of the signal being measured, and that view needs to be stated and justified rather than arrived at by fitting to historical returns. We document this layer in full because it is where most analytical frameworks are weakest — the inputs and construction logic are often left implicit.

The second layer is cross-market plausibility review. A term structure momentum signal that appears robust in isolation but contradicts the signal dynamics in structurally related markets is a warning sign, not a confirmation. We examine how the signal behaves relative to term structure dynamics in adjacent futures markets — whether it tends to lead, lag, or invert relative to signals in markets with established supply or risk linkages. A signal that makes sense in isolation and also makes sense in the context of cross-market dynamics is a more credible analytical finding than one that requires ignoring the broader connected system.

The third layer is scenario and stress analysis against historical structural dislocations. The periods that matter most for evaluating a term structure momentum framework are not the quiet, trend-following periods — they are the supply shocks, the rate regime transitions, and the volatility spikes. These are the conditions under which simplified signal construction logic tends to fail in the most consequential ways. We build our stress analysis around those specific historical episodes and document what the framework would have implied during those periods, including where it would have produced misleading signals. That documentation is, in our view, one of the more useful things we publish.

Cross-market signal dynamics — why we treat commodity, rate, and volatility futures as a connected analytical system rather than separate asset classes

Momentum signals in one futures market do not develop in isolation. The sequencing of term structure shifts across commodity, rate, and volatility markets follows patterns that single-market analysis cannot detect — and that cross-market perspective is where our analytical work is most distinctive.

Treating futures markets as separate silos is analytically convenient and practically expensive. Signal sequencing across markets is where the most useful information often lives.

Energy commodity futures term structure is connected to interest rate futures dynamics through multiple transmission channels. Supply-driven backwardation in crude oil markets has historically coincided with, or preceded, specific patterns in inflation expectations that are visible in short-term rate futures curve slope. That connection is not always present — the strength of the relationship varies across economic regimes — and documenting when the connection is reliable versus when it weakens is as important as identifying that the connection exists at all. We map both the signal and its conditionality.

Volatility futures term structure interacts with both commodity and rate curve dynamics through the risk appetite channel. When volatility term structure inverts sharply — near-dated implied volatility spikes above deferred — this typically reflects acute near-term uncertainty that also tends to show up in commodity curve flattening and rate curve dynamics. The sequence and magnitude of those cross-market reactions varies, and the analytical question is whether the sequencing itself carries predictive information about subsequent term structure momentum in each individual market. That is the kind of question our cross-market signal review is designed to address.

Agricultural commodity futures term structure operates on supply cycles that are structurally distinct from energy markets — seasonal planting and harvest patterns create predictable roll dynamic windows that have different signal properties than the storage-driven dynamics in energy. Mapping how agricultural term structure momentum relates to broader commodity curve dynamics, and whether those relationships transmit into rate or volatility futures in observable ways, is an area of active analytical development for us. We are candid about where our coverage is still developing rather than presenting incomplete analysis as fully formed conclusions.

Commodity futures

Commodity futures term structure is the area where our analytical frameworks are most fully developed. Energy markets — crude oil, natural gas, refined products — are the primary focus, with agricultural markets representing a secondary coverage area. The signal construction logic in commodity term structure centres on roll yield dynamics, contango and backwardation transitions, and the behaviour of curve slope across different supply-demand regimes. We map how those dynamics have behaved historically across structural dislocations — major supply events, demand shocks, storage capacity constraints — because those are the periods that most stress-test the reliability of any momentum signal derived from the curve shape. Our three-layer review process applies in full to every commodity term structure framework we publish: signal identification from curve data, cross-market plausibility review, and scenario stress-testing. We document the conditions under which the signal has historically been less reliable because practitioners deserve that information before they can evaluate whether the framework is applicable to their current analytical context. Past performance does not guarantee future results — we say that clearly and mean it analytically, not just as a disclaimer.

energy commodity futures term structure analysis
interest rate and volatility futures curve analysis

Rate and vol futures

Interest rate futures term structure and volatility futures dynamics represent the second and third strands of our analytical coverage. Rate futures curve slope and curvature carry momentum signals that operate on different time horizons than commodity roll dynamics — the persistence of rate curve slope shifts tends to be driven by monetary policy regime transitions rather than physical supply-demand mechanics. Our analytical frameworks for rate futures specify the lookback window assumptions appropriate to rate regime persistence rather than applying commodity-derived construction logic. Volatility futures add a third distinct layer — the structural contango in VIX-related instruments during calm periods, and the sharp inversions during stress events, create momentum dynamics that require their own signal construction approach. We map how rate and volatility term structure dynamics interact with commodity curve signals across the connected system we cover, identifying where sequencing patterns have historically carried cross-market information. These are the analytical questions we find most interesting, and they are the ones where we think our cross-market framework produces the most distinctive analytical perspective compared to single-market approaches.
practitioners discussing futures term structure frameworks

Ready to go deeper?

Our team discusses the specific markets and frameworks most relevant to your analytical work

If the concepts on this page connect with questions you have been working through in commodity, rate, or volatility futures markets, we are straightforward people and we are happy to talk through the specifics. We do not provide personalized financial advice or trade recommendations — that boundary is firm and we respect it clearly. What we do is discuss analytical frameworks, signal construction logic, and cross-market dynamics with practitioners who are working through the same kinds of questions we find most interesting. Reach out through the contact page and tell us what you are working on. We will give you a direct, honest response about whether and how our analytical work might be relevant to your situation.

Say hello

Five things practitioners find most useful to know about our analytical approach

These are the clarifications that tend to make subsequent conversations more productive — the points where our approach differs from what practitioners typically encounter in standard term structure research, and where those differences have the most practical significance for how the frameworks should be read and evaluated.

We specify the failure conditions first
Every framework we publish includes an explicit account of the market regimes and structural conditions under which the signal has historically been less reliable. We present this before the headline findings, not after them, because knowing where a framework breaks down is essential for evaluating whether the period you are applying it to resembles the conditions under which it worked.
Lookback windows are justified, not fitted

The choice of lookback window in a term structure momentum signal is one of the most consequential methodological decisions in signal construction. We specify the lookback window based on a reasoned view about the persistence horizon of the relevant curve dynamic, and we document that reasoning explicitly so it can be evaluated and challenged independently rather than accepted as a given.

We distinguish signal from noise by market regime

A term structure momentum signal that is robust across all market regimes is almost certainly too good to be true. We segment our signal analysis by identifiable market regimes — contango versus backwardation, rising versus falling rate environments, calm versus stressed volatility surfaces — and report how the signal behaves differently across those states rather than averaging over them.

Cross-market comparisons require matched construction logic
Comparing term structure momentum signals across commodity, rate, and volatility futures requires careful attention to the differences in how those curves are constructed, what drives their dynamics, and what roll windows are analytically meaningful in each market. We do not apply a single signal construction template across all three areas — we adjust the construction logic to reflect the specific mechanics of each curve type.
Hedging claims is a feature, not a limitation

Every claim we make about term structure momentum comes with explicit boundary conditions. Past performance of any analytical framework does not guarantee future results. Signals degrade. Market structures evolve. We treat those caveats as analytically informative rather than as legal formalities to be minimised, because the conditions under which a signal stops working are often as instructive as the conditions under which it works.