How this instrument works
Expected monetary value takes every possible outcome of an uncertain choice, multiplies its dollar result by how likely it is, and adds the pieces together. The output is not a prediction of what will happen — it is a weighted average of what could happen, built so the more likely branch pulls the number toward itself and the less likely branch pulls less. A coin flip worth $100 or nothing has an EMV of $50, even though $50 itself can never be the actual result of a single flip.
The technique comes from decision analysis, the discipline built around decision trees: branch out every choice and chance point, work backward from the payoffs, and at each chance point replace the branch with its EMV so the tree collapses into a single comparable number. A procurement officer weighing two supplier bids, a project manager choosing between contracts with different failure risks, and an underwriter pricing a policy all lean on the same arithmetic — probability times payoff, summed across the branches.
EMV treats a dollar the same no matter who is holding it, which is its biggest limitation. Losing $20,000 barely dents a company sitting on $50 million in reserves but can sink a small contractor outright, yet the formula scores both cases identically. It also says nothing about how spread out the outcomes are — a bet with a huge win and a huge loss can carry the same EMV as a mild bet with two modest outcomes. Use it to rank options on average return, then look at the worst case on its own before committing money to any of them.
- Enter Probability of outcome 1 (0-1) and Payoff of outcome 1, $ for the first branch of the decision.
- Enter Probability of outcome 2 (0-1) and Payoff of outcome 2, $ for the second branch — enter a loss as a negative number.
- Check that the two probabilities describe the full set of branches you are modeling; the formula will not warn you if they miss a case.
- Read Expected monetary value — the single weighted figure to compare against another option's EMV before choosing between them.
Worked example — a 60/40 bid decision
A contractor is deciding whether to chase a bid that has a 60% chance of landing a job worth $50,000 in profit and a 40% chance of losing the bid after $20,000 already spent on the proposal. That is p1 = 0.6, payoff1 = $50,000, p2 = 0.4, payoff2 = -$20,000 entered into the instrument exactly as the decision is framed.
The instrument multiplies each branch on its own — 0.6 times $50,000 equals $30,000 for the win, and 0.4 times -$20,000 equals -$8,000 for the loss — then adds them: $30,000 plus -$8,000 equals $22,000. That $22,000 is not a guaranteed payday; it is the average result if this exact bid were chased many times over, and it is the number the contractor sets beside a second bid's EMV before deciding where to spend proposal time.
Questions
What does a negative expected monetary value mean?
It means the weighted-average outcome favors a loss even though a win is still possible. If a bet's EMV is -$3,000, chasing it many times over would lose money on average, though any single attempt could still land on the winning branch. A negative EMV is a signal to compare it against the EMV of walking away, which is $0.
Can EMV handle more than two outcomes?
Yes — this instrument covers the common two-branch case, but the same arithmetic extends to any number of outcomes: multiply each outcome's probability by its payoff and add every term. Decision trees with three or four branches at a node use the identical rule; you would just sum more products.
Why doesn't EMV account for how painful the worst case would be?
Because it only weights outcomes by probability and dollar size, never by how much a loss would hurt the specific decision-maker. Losing $20,000 is a rounding error for a large firm and a crisis for a solo contractor, but EMV scores both the same. Analysts who care about that gap use expected utility instead, which reshapes payoffs by risk tolerance before averaging them.
Do the two probabilities have to add up to 1?
They should, if the two branches are meant to cover every possibility a 0.6 and a 0.4 together account for the whole decision. The instrument still computes an answer if they do not sum to 1, but the result then describes an incomplete or overlapping set of outcomes rather than the full picture.
Is EMV the same thing as ROI?
No. Return on investment compares a gain to the cost that produced it, as a ratio. EMV compares outcomes to each other in raw dollars, weighted by how likely each one is. A high-EMV choice can still carry a lower ROI than a low-EMV one if it needs far more capital to attempt.
How is EMV used inside a full decision tree?
At each chance point in the tree, EMV replaces the branching outcomes with one number — exactly what this instrument computes for a single point. Working backward from the tips of the tree, each chance point collapses to its EMV, decision points then keep whichever branch has the higher EMV, and the tree resolves into one recommended path.
References
- Open University OpenLearn — Decision trees and expected value
- MIT OpenCourseWare — Risk and Decision Analysis (IDS.333)
Read this first: This instrument shows arithmetic, not advice. Real offers add fees, taxes and terms that vary by lender and place — verify the figures against your actual paperwork before deciding anything.