Expected monetary value converts a list of uncertain events into a single number by multiplying each one's probability by its financial impact and adding the results. It is the standard first-pass quantitative technique in project risk management, and its value is not that it predicts anything — it does not — but that it forces every risk on a register to carry an explicit probability and an explicit cost, which is where most risk registers quietly fail.
Arb Digital built this page for the register case specifically. Our live expected value calculator handles a single random variable whose outcomes are exhaustive and whose probabilities sum to one, and returns the mean, variance and standard deviation of that distribution. This page handles something different: many independent binary events, each with its own probability, none of which are mutually exclusive and which sum to nothing in particular. It separates threats from opportunities and produces a contingency figure, which the distribution tool does not attempt.
What This EMV Calculator Does
It reads a risk register, computes the expected monetary value of each line as probability times impact, and totals them. It reports the threat total and the opportunity total separately, because they are not interchangeable, and it derives a contingency reserve on whichever convention you select. It also expresses that reserve as a percentage of your base estimate, which is the form most useful in a budget conversation.
The bar breakdown ranks every risk by the absolute size of its EMV, which is usually more informative than the register itself. Risk registers are typically ordered by when someone thought of the risk, and the ordering by expected cost almost always differs from people's intuition about which risks matter.
It produces no verdict on whether the reserve is adequate. Adequacy is a function of risk appetite, the confidence level you are targeting and the correlation between risks, none of which a sum of products can see.
How to Use It
- Enter one risk per line as name, probability in per cent, and impact. Sign the impact: negative for something that costs money, positive for something that saves or earns it.
- Use only identified, discrete risks. Vague categories like “general uncertainty” have no meaningful probability and should be handled through estimate ranges instead.
- Choose the contingency convention. Threats-only is the more conservative practice and the more common one.
- Enter the base estimate so the reserve can be read as a percentage. That percentage is what a sponsor will actually respond to.
- Use the scaling factor to test sensitivity. If doubling every impact does not change the decision, the estimate is robust; if it does, the impact estimates deserve more work than the probabilities.
The Formula
For a single risk:
EMV = probability × impact
and across a register, EMV total is the sum of those products. A worked example on the default register. A supplier delay at 30 per cent costing 120,000 has an EMV of −36,000. A scope change at 45 per cent costing 60,000 gives −27,000. A permit refusal at 10 per cent costing 250,000 gives −25,000. Rework at 60 per cent costing 35,000 gives −21,000. Those four threats total −109,000. An early delivery bonus at 25 per cent worth 80,000 gives an opportunity EMV of +20,000. The net EMV is −89,000.
On the threats-only convention the contingency reserve is 109,000, which against a base estimate of 2,400,000 is 4.54 per cent. On the netting convention it would be 89,000, or 3.71 per cent. Same register, two defensible answers, and the difference is entirely the convention.
Notice which risk dominates. The permit refusal has by far the largest impact at 250,000, but its low probability puts it third by EMV. That reordering is the main thing EMV does, and it is also its main limitation, discussed below.
EMV Describes an Average You Will Never Experience
This is the point most misused in practice. An EMV of −36,000 for the supplier delay does not mean you will lose 36,000. You will lose either 120,000 or nothing. The 36,000 is the average over a very large number of identical projects, and you are running one.
For a register with many small, independent risks that averaging is reasonable, because the law of large numbers does some work for you and the total tends toward its expectation. For a register dominated by one or two large, low-probability items it is much weaker, because the actual outcome will be far from the average in either direction. A reserve of 109,000 does not cover the 250,000 permit refusal if it happens.
The practical implication is that EMV should never be the only quantitative technique on a project with material tail risk. Simulation over the whole register, which produces a distribution of outcomes rather than a point, is the standard answer, and it lets you reserve to a stated confidence level rather than to an average. The GAO covers risk and uncertainty analysis, and the relationship between it and contingency funding, in its Cost Estimating and Assessment Guide.
Where the Probabilities Come From
The uncomfortable answer is usually people's judgement, and judgement has known biases. Estimates cluster on round numbers, anchoring pulls a whole workshop toward the first figure spoken aloud, and the same event gets a different probability depending on whether it is framed as a risk of failure or a chance of success.
Two habits help. First, elicit probability and impact separately and from different people where possible, so that one does not contaminate the other. Second, keep the scale coarse and honest: 10, 30, 50, 70, 90 is a defensible set, and quoting a risk at 37 per cent implies a precision nobody has. Our probability calculator is useful for the combining arithmetic when a risk depends on several conditions.
The impact side needs the same discipline and gets less of it. An impact should be the full cost if the risk occurs, including schedule consequences converted to money, not the cost of the immediate remedy. Underestimating impacts is more common and more damaging than misjudging probabilities, which is why the scaling factor exists on this page. The Project Management Institute publishes practitioner material on quantitative risk methods, including EMV, in its library article on quantifying risk.
Independence, Correlation and Double Counting
Summing EMVs assumes the risks are separate events. They frequently are not. A supplier failure and a rework requirement may share a root cause, so they occur together or not at all, and treating them as independent understates the variance of the total even though it leaves the expected value unchanged.
Worse is double counting. If “supplier delay” and “schedule overrun” are both on the register and the second is largely caused by the first, their impacts overlap and the reserve is inflated. The test is whether the impacts could genuinely both be incurred in full. If not, merge them or reduce one impact to its incremental part.
Related risks with a common trigger should ideally be modelled as a single risk with several consequences. That keeps the arithmetic honest and, more usefully, points the response strategy at the trigger rather than at each symptom separately.
Threats and Opportunities Are Not Symmetric
Netting an opportunity against a threat looks like sound arithmetic and is usually poor risk management. If the opportunity does not materialise, you are left funding the threats from a reserve that assumed it would. The asymmetry is practical rather than mathematical: nobody is harmed by an unused reserve in the way they are harmed by an inadequate one.
Opportunities also tend to be softer. A cost saving from an unproven approach is frequently optimistic, whereas a threat impact is often understated for exactly the opposite reason. Keeping the two totals visible separately, which this page does by default, is more informative than a single net figure however you eventually decide to reserve.
The reserve itself is also not the whole picture. Contingency reserve covers identified risks. Management reserve, held above it, covers what nobody identified, and no register can size that by construction. Test the resulting number against the volume assumptions using our break-even calculator, and where risks affect a multi-year investment, the NPV calculator is the right frame.
What EMV Is Genuinely Good For
Three things. It ranks risks by expected cost rather than by how alarming they sound, which routinely changes where attention goes. It provides a defensible, auditable basis for a contingency number, which is far better than a percentage plucked from experience. And it prices risk responses: if mitigating a risk costs less than the reduction in its EMV, the mitigation pays for itself, and that comparison is the most useful single thing this arithmetic supports.
It also makes a register falsifiable. Once every line carries a probability and an impact, disagreement becomes specific and productive rather than a debate about whether something is “high” risk. For comparing structured alternatives rather than costing a register, our expected utility calculator handles the case where risk aversion matters, and the risk reward ratio calculator and value at risk calculator cover the trading and portfolio framings.
Arb Digital builds free tools like this one because useful pages earn attention. If you want tools, calculators or content built for your own audience, we can help.
Browse All Free Tools Talk to Arb DigitalCommon Mistakes to Avoid
- Reserving the EMV and calling it covered — a reserve equal to the expected value covers the average outcome, not the one you are about to have, and a single large risk can exceed it many times over.
- Putting certainties on the register — anything at 100 per cent is a cost, not a risk, and it belongs in the base estimate where it will not be double counted.
- Double counting related risks — if one risk largely causes another, their impacts overlap and the reserve is inflated. Merge them or reduce one to its incremental part.
- Netting opportunities against threats without saying so — it lowers the reserve on the strength of benefits that may never appear, and the convention has to be stated either way.
- Spending effort on probabilities and none on impacts — impacts are usually the weaker estimate and have exactly the same leverage on the result.
Related Free Tools From Arb Digital
Handle a single random variable's full distribution with the expected value calculator, bring risk aversion into a choice with the expected utility calculator, and combine conditions with the probability calculator. Frame downside in trading terms with the risk reward ratio calculator and the value at risk calculator, test volumes with the break-even calculator, discount multi-year outcomes with the NPV calculator, weight inputs with the weighted average calculator, and browse the full free online tools hub.
Frequently Asked Questions
Expected monetary value is probability multiplied by impact for a single risk, and the sum of those products across a register. Impacts are signed, so threats reduce the total and opportunities increase it, and the probabilities are per risk rather than a distribution that has to sum to one.
An expected value calculator takes one random variable whose outcomes are exhaustive and mutually exclusive, with probabilities summing to one, and returns the mean, variance and standard deviation. A risk register is many independent binary events that are not mutually exclusive, so it needs separate threat and opportunity totals and a contingency figure instead.
Only as a starting point. The EMV is an average, so reserving exactly that amount leaves roughly even odds of being short. Where one or two large risks dominate, a simulation that produces a distribution lets you reserve to a stated confidence level, which is a far more defensible basis.
That is a convention you must state. Reserving against threats alone is more conservative and more common, because an opportunity that fails to appear leaves the threats unfunded. This page defaults to threats only and shows both totals so the choice is visible.
Usually structured judgement, which carries known biases. Elicit probability and impact separately, use a coarse scale rather than implying false precision, and record the reasoning. A probability quoted to the nearest per cent implies confidence that almost never exists.
It is not a risk. Anything at 100 per cent probability is a known cost and belongs in the base estimate. Leaving it on the register either double counts it or hides a real cost inside a contingency line where nobody expects to find it.
No. Summing expected values is valid whether or not risks are correlated, but the spread of possible outcomes is not, and correlation is exactly what determines whether a reserve is likely to hold. Risks with a shared root cause are better modelled as one risk with several consequences.
That is one of its best uses. Compare the cost of a response with the reduction in EMV it produces, remembering that a response usually lowers probability, impact or both rather than eliminating the risk. The comparison is a genuine input to a decision, though it does not account for how much variance you are willing to carry.
This tool is provided for informational and educational use only. It is not financial, insurance or project management advice, and it does not determine whether any contingency reserve is adequate. Expected monetary value describes an average across many hypothetical outcomes and cannot predict a single project. Consult a qualified adviser before relying on any figure produced here.
