Last night’s VaR run was accurate. It reflected your positions at close, your market data at close, your risk parameters at close. The problem is that none of those things are the same this morning. The market moved overnight. A position got updated. A counterparty pushed back on a trade. Your VaR number is still sitting in the report from yesterday.
This is the quiet contradiction at the centre of commodity trading risk management for most active desks. VaR is treated as a monitoring tool. But when it is calculated once a day, at the end of the session, it is not monitoring anything. It is recording history with one data point.
Value at Risk (VaR) is a statistical measure that estimates the potential loss a trading desk could reasonably expect over a defined period, at a given confidence level, based on current positions and market conditions. That definition contains a critical word: current.
The moment VaR becomes a backward-looking calculation, run against yesterday’s closing positions and yesterday’s closing prices, it stops being a risk tool and becomes a record. A record of what the risk was. Not what it is.
Most desks know this. Most desks accept it anyway. The infrastructure to calculate VaR continuously does not exist in most spreadsheets, and it does not exist in most legacy systems either. So end-of-day VaR becomes the standard: not because it is best in class, but because it is the only thing available and therefore “good enough”.
The gap between what VaR reports and what risk actually is tends to be invisible most of the time. When markets are stable and positions are steady, yesterday’s VaR number is close enough to today’s that the difference rarely matters in practice.
Commodity markets are not always stable. And the sessions where the gap widens the most are precisely the sessions where the on-demand VaR calculation become critical.
Consider what happens to a VaR figure during a period of sharp intraday volatility. Oil prices move 4% during the session in response to a supply disruption. The desk has open positions across several books. The VaR calculated at last night’s close reflected a different price environment entirely. The positions themselves may have changed as traders responded to the move. But by the time the risk team has a clear picture of the desk’s current exposure, the session may already be over.
That is not a failure of the traders or the risk managers. It is a failure of the infrastructure. When VaR is calculated once a day, the risk team is always operating on a time delay. In stable conditions, that delay is tolerable. In volatile conditions, it is the thing that turns a manageable loss into a difficult conversation.
The same logic applies to margin. Margin calls do not wait for end-of-day. If your margin exposure is only visible once the session closes, you are always reacting. Never anticipating. The desk that gets a margin call it was not expecting is the desk whose infrastructure was giving it a picture of yesterday, not today.
There is a moment that most risk managers in commodity trading will recognise. The session is moving against several open positions. Someone asks for the current risk picture. The honest answer is: we have last night’s numbers, and we are piecing together what has changed since then.
That moment, the gap between what the risk report says and what the desk actually knows, is where on-demand VaR matters most. Not as a compliance exercise. As a decision-making tool.
The standard response to this problem has been to run more frequent batch calculations. Some desks run VaR at midday as well as end-of-day. Some run it hourly. Each additional batch calculation adds operational overhead and still leaves gaps. A batch calculation, however frequent, is always a snapshot of the past. The market does not pause between snapshots.
The practical difference between end-of-day VaR and on-demand VaR is not primarily about the number. It is about when the number is available and what the risk team can do with it.
When VaR, margin, mark-to-market (MTM), and exposure are calculated in seconds, against live positions, and visible alongside real-time P&L by book, the risk team’s relationship to volatility changes. A session that moves sharply does not produce a scramble to understand the current exposure. The current exposure is already visible. The decision is whether to act on it, not how to calculate it.
Loqsea’s Risk Manager calculates VaR, margin, MTM, and exposure in seconds, updated continuously as positions and prices move. The risk view is not a report produced at the end of the day. It is a live picture of where the desk stands at any point during the session, visible alongside real-time P&L tracked by book, strategy, and instrument.
That shift changes what a risk manager can do. It allows the desk to see margin exposure building before a call arrives. It allows a head of trading to see VaR moving in real time as a volatile session develops, rather than discovering what happened after the close. It removes the time delay that turns manageable risk into reactive damage control.
Most commodity trading desks are running VaR in a way that would not pass scrutiny if the question were asked directly. Not because the calculation is wrong. Because the calculation is old by the time anyone acts on it.
The question is not whether end-of-day VaR is better than nothing. It is. The question is whether a desk running active positions across multiple books, in markets that can move 5% in an hour, can genuinely call what it is doing risk management. Or whether it is, more accurately, risk recording.
The infrastructure to do better exists. The question is whether the desk is using it.
Loqsea is a cloud-native commodity trading and risk management platform built for active trading desks. No lengthy implementation. No legacy infrastructure. VaR, margin, MTM, and real-time exposure available from day one, alongside live P&L tracked by book, strategy, and instrument.
Built by traders who lived the problem. Used by CTAs, hedge funds, and trading desks who cannot afford to work from yesterday’s numbers.
Value at Risk (VaR) in commodity trading is a statistical measure that estimates the potential loss a trading desk could expect over a defined period, at a given confidence level, based on current positions and market conditions. It is used by risk managers and heads of trading operations to monitor exposure across books, instruments, and commodities.
End-of-day VaR reflects positions and market prices at the close of the previous session. For desks with active intraday positions, it does not account for trades executed during the current session, intraday price moves, or changes in market conditions since the close. In volatile commodity markets, the gap between end-of-day VaR and current risk can be significant and consequential.
When VaR is available on demand, risk managers can see exposure building in real time as positions and market conditions change. This allows proactive decisions around hedging, position sizing, and margin management, rather than reactive responses after the session has closed and the opportunity to act has passed.
VaR is a probabilistic measure: it estimates the range of potential losses under defined conditions. Real-time exposure is a direct measure of the desk’s current risk position, including open positions, mark-to-market values, and margin requirements, updated continuously. The two are complementary. VaR without real-time exposure context is incomplete, and real-time exposure without a VaR framework lacks statistical depth.
Loqsea’s Risk Manager calculates VaR for all major exchanges, covering energy, metals, agriculture, and freight, across futures, swaps, and options.
How is Var calculated:
Parametric VaR over a 180-day look-back period
95% and 99% VaR. As well as VaR as a percentage of your book value.
With the ability to chart your VaR over a timeseries.
