Skip to content
Find your tool
Guides · 5 min read

Why the sum of EMVs is not a risk reserve

Probability times impact, added up, gives a total that almost never happens. Two reasons it makes your reserve too small, and what to do instead.

By Mike Bouman, Quantil · November 4, 2026

Many risk registers end with a total: the sum of probability times impact for every risk. It's tempting to treat that total as the risk reserve. It's not, for two reasons, and both make the reserve too small.

What EMV is, and what it is not

The expected monetary value (EMV) of a risk is its probability times its impact. A risk with a 25% chance of costing $2M has an EMV of $0.5M. That's a useful number for ranking and for judging whether a mitigation pays for itself.

But $0.5M is an average, not an outcome. In any real future the risk either happens and costs around $2M, or doesn't happen and costs nothing. Its $0.5M will never appear on an invoice. Add up the EMVs of a whole register and you get an amount that the project almost never actually spends.

Reason 1: the average says nothing about confidence

Take the five risks in the simulator on our homepage. With the most likely impact of each, their EMVs add up to $16.7M. We simulated just that risk layer 200,000 times, with a fixed seed.

Outcome of the five risksShare of the simulated futures
None of the risks occurs15%
Exactly one occurs35%
Two or more occur50%
Total risk cost within $2.5M of the EMV sum14%

The outcomes are spread out, and the EMV sum lands in a gap between them. A reserve set at the EMV sum covers the risks in about half of the futures. Covering them in 80% of the futures takes $30.6M, nearly twice the EMV total. For 90%, it takes $38.4M.

Reason 2: impacts are skewed upward

The EMV column normally uses the most likely impact. But impacts are ranges, and they run further up than down. A risk that most likely costs $18M might cost $10M at best and $40M at worst. Its average impact is higher than its most likely impact.

Across the five simulator risks, the true expected cost is $18.7M, not $16.7M, about 12% higher. In the fictional tunnel case of our QRCIM whitepaper the same effect is larger: the EMV column adds up to $29.8M, while the risks' expected contribution to the total is about $36M. Budgeting on the EMV sum means budgeting too low even before confidence comes into it.

Ranking on EMV hides the dangerous risks

The same problem shows up in top-ten lists. Two risks with the same EMV can have very different worst cases. In the tunnel case, a subcontractor insolvency ranks low on EMV, at $2.4M, but has a maximum impact of $48M. A list sorted on EMV alone would push it down the page. Look at the range and the contribution to the spread, not only at the EMV column.

Where EMV does belong. EMV is a fair way to compare risks of a similar shape, and to check whether a mitigation is worth its cost: if a $50,000 measure lowers a risk's EMV by $300,000, that's a strong case. Just don't add the column up and call it a reserve.
A practical tip. In some risk registers the EMV total is presented as the risk budget, simply because it’s the only total the register produces. The gap only shows when several large risks hit in the same year. Use the EMV total as a check on the register, never as the reserve. The reserve comes from the simulation.

What to do instead

Simulate the register. Each run decides for every risk whether it occurs and, if so, draws its impact from the full range. The result is a distribution of total risk cost, from which you read the reserve at the confidence level your board uses.

The Cost & Schedule Risk Analyzer does this straight from your register in Excel, for cost and for schedule, with correlation between risks. You get the reserve in money and in weeks at P50, P80 and P90, with a report and a deck. To see the difference first, run the simulator on our homepage. For why a flat percentage doesn't solve it either, read why a percentage on top is not a risk budget.