P&L Explained for Commodity Traders

Executive summary

Profit and loss is the number a trading floor lives by, and yet it is one of the most misunderstood figures in the business. Most published material treats P&L as an accounting concept, revenue recognised under reporting standards, but a commodity trading desk needs something different: economic P&L, calculated continuously, that reflects the real, current value of the book. This guide is about that number, how a modern ETRM calculates it, not just what it means.

Trading P&L is the product of a chain: trade capture, market data, forward curves, positions, and valuation all feed into it, and settlement and accounting flow out of it. An error anywhere in that chain shows up in the P&L, which is why P&L is best understood as the visible output of the whole platform working, or failing, together. The desks that trust their P&L are the ones whose capture, curves, and valuation all sit on one governed model.

This is a complete, practitioner-level treatment: the types of P&L (realized, unrealized, intraday, forecast), how mark-to-market drives unrealized P&L, how physical and financial P&L differ, and, most valuable of all, how P&L attribution explains why the number moved. It is a pillar reference that builds on trade capture, position management, market data, and forward curves, and connects forward into the risk stack.

What trading P&L is

Profit and loss measures the financial performance of a trading portfolio, but in commodity markets it is shaped by far more than executed trades. P&L moves with changing market prices, forward curves, FX rates, transportation and storage costs, and settlement activity. A position untouched all day can still generate P&L simply because the curve that values it moved.

The critical distinction is between accounting profit and trading P&L. Accounting recognises revenue according to financial-reporting standards, on defined events and schedules. An ETRM calculates economic performance continuously, marking the book to market to show what it is worth right now. Both are legitimate and necessary, but they answer different questions, and a trader manages the economic number. Conflating the two is a common source of confusion between the front office and finance.

Why P&L is the most important trading metric

Traders watch P&L throughout the day because so many decisions hang on it. P&L measures trading performance and evaluates whether a strategy is working. It reveals exposures as they generate gains or losses. It triggers hedging decisions when the number moves beyond comfort. It drives management reporting and, often, trader compensation. And it underpins regulatory and audit requirements.

Because so much depends on it, delayed or inaccurate P&L is genuinely dangerous. A P&L that is hours old, or that cannot be trusted, leads to hedging the wrong exposure, misjudging a strategy, or reporting a number that later has to be restated. This is why a modern platform treats P&L as a real-time, explainable output rather than an overnight report, and why the integrity of the whole data chain matters so much.

How an ETRM calculates P&L

P&L is the end of a dependency chain, and understanding that chain is the key to understanding the number. Market data and forward curves provide the prices; trade capture records the deals; the position engine aggregates them into exposures; the valuation engine marks those positions to market; and the P&L engine combines the marks with realized cash flows to produce the number, which flows on into risk reports and dashboards.

The consequence of this chain is that P&L quality is data quality. A stale price, a wrong curve version, a mis-captured trade, or a mis-aggregated position each produces a wrong P&L. This is the strongest argument for a single governed model: when capture, curves, positions, and valuation all read the same governed data, the P&L is consistent and reproducible, because every input traces to one trusted source rather than several that disagree.

The types of P&L

P&L is not one number but several, each answering a different question. Precision about the types is what lets a desk and its finance team speak the same language.

The realized-versus-unrealized distinction is the most important. Realized P&L is locked in: a closed position or settled trade has produced actual cash that will not change. Unrealized P&L is mark-to-market on open positions, it fluctuates every time the curve moves and is only realized when the position closes. A desk with large unrealized gains has not yet banked them, and confusing the two flatters or alarms unnecessarily. Reporting them separately is a mark of a disciplined P&L process.

Mark-to-market valuation

Mark-to-market is the engine of unrealized P&L, and it is central to commodity trading because so much of the book is open positions with future delivery. Marking to market means valuing each open position against the current market, which for forward-dated energy means against the current forward curve.

Getting this right depends on several things done correctly: selecting the right forward curve for the commodity and location, using the correct curve version for the valuation date, applying appropriate discounting to future cash flows, and respecting delivery periods and valuation dates. Because the curve version matters, reproducibility is essential: a past mark must be rerunnable against the exact curve used at the time. The output of this process, the mark-to-market value, feeds directly into unrealized P&L, which is why the valuation engine and the P&L engine must share the same governed curves.

Physical vs financial P&L

Physical and financial trades generate P&L differently, and a desk trading both needs to see them coherently on one book.

Physical P&L reflects the full economics of moving a real commodity: the purchase and sale price, but also transportation, storage, line losses, fuel costs, and imbalances. A physical power or gas deal that looks profitable on price alone can be marginal once delivery costs are included, so physical P&L must capture the whole cost stack, not just the price spread.

Financial P&L comes from instruments, futures, swaps, options, and FX, and is shaped by margining, premiums, and settlement mechanics rather than physical delivery. The two must net on one model, because a physical position is frequently hedged financially, and only a unified P&L shows the true combined result. Seeing physical and financial P&L separately and together, on the full commodity complex, is what a modern platform provides.

P&L attribution: explaining why

The single most valuable P&L capability is attribution: breaking a P&L movement into its explainable causes so a trader understands why the number changed, not just that it did. A P&L that moved by a large amount overnight is far more useful when it can be decomposed.

Attribution splits the movement into components: the price effect (how much came from price moves on existing positions), the volume effect (new trades), curve movement, FX movement, time decay (theta), carry, transportation cost, and fees and taxes. A well-attributed P&L reads like an explanation: so much from the gas curve steepening, so much from new trades, so much from the euro weakening, so much from a day of time decay. This is what turns P&L from a scoreboard into a diagnostic. Attribution also validates the P&L itself: if the components do not add up to the total, there is an unexplained residual to investigate, which is a key quality signal. It ties directly to the Greeks, which quantify the same sensitivities.

Real-time P&L

In a fast market, P&L has to be live, and that requires an event-driven architecture. A trade event or a market update flows through the position engine and valuation to the P&L engine, which incrementally recalculates only what changed and pushes the result to dashboards in near real time.

The traits that make this work are incremental recalculation rather than full rebuilds, low-latency updates, continuous intraday reporting, and alerting when P&L crosses a threshold. The contrast with an overnight P&L run is decisive: a live P&L lets a trader see the effect of a decision immediately and react to the market as it moves, while a batch P&L only tells the story after the close. Real-time P&L is a natural consequence of the same event-driven, single-model architecture that powers real-time positions.

Portfolio-level reporting

P&L is most useful when it can be viewed along every dimension a manager cares about. A modern P&L engine reports by trader, portfolio, commodity, strategy, region, delivery month, counterparty, and business unit, and lets a user combine them on demand.

This flexibility serves different audiences from the same underlying data. A trader wants their own P&L by strategy and delivery month; a desk head wants it by trader and commodity; an executive wants it rolled up by business unit with the ability to drill into any surprise. Because all of these views come from the same governed P&L calculation, they always reconcile, unlike a landscape where each report is built separately and the numbers mysteriously differ. Governed reporting and analytics over one model is what makes this coherence possible.

Integration with risk

P&L does not stand alone; it is tightly bound to the risk stack. P&L and its underlying valuations feed VaR, stress testing, scenario analysis, credit exposure, liquidity reporting, and hedge-effectiveness measurement.

The essential requirement is consistency of valuation assumptions across P&L and risk. If P&L marks the book on one set of curves and VaR uses another, the two tell contradictory stories and neither can be trusted. When P&L and risk share the same governed valuation, the risk numbers are measuring the same book, valued the same way, as the P&L, which is what lets a desk reconcile its performance and its risk into one coherent picture. This shared-valuation discipline is a defining feature of an integrated platform.

P&L best practices

The practices that produce trustworthy P&L are consistent across desks:

• Centralize market data so P&L is marked on one trusted source.
• Version forward curves so any mark is reproducible.
• Automate valuations to remove manual error.
• Standardize P&L definitions so everyone means the same thing by realized, unrealized, and total.
• Reconcile daily against settlements and independent sources.
• Explain P&L movements through attribution, with a small unexplained residual as a quality gate.
• Preserve historical valuations for audit and restatement.
• Align P&L with risk by sharing valuation assumptions.
• Separate realized and unrealized in every report.
• Monitor calculation latency so P&L stays genuinely real-time.

The thread is the same one that runs through this guide: P&L is only as good as the governed data and shared valuation beneath it, and explainability is what makes it trustworthy.

P&L KPIs

P&L quality can be measured, and a few KPIs keep the process honest.

The most telling of these is unexplained P&L: the portion of the movement that attribution cannot account for. A low unexplained residual means the P&L is understood and trustworthy; a rising one is an early warning that something in the data or valuation chain is wrong. Together with latency and reconciliation, these KPIs describe a P&L a desk can rely on.

Why the Gravitas P&L engine is different

Gravitas produces P&L as the unified output of one governed platform rather than a separately built report.

Because P&L is computed from the same governed market data, curves, positions, and valuation that feed risk, the performance number and the risk number always agree, and every movement is attributable and reproducible. And it is delivered at economics that suit desks the incumbents priced out. See who Gravitas is for or request a demo.