Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 9-12

Decision Under Uncertainty

Decision under uncertainty is choosing between options whose results aren't known for sure, usually by comparing expected values or risk profiles.

Venn diagram of Event A and Event BA two-event view of decision under uncertainty.ABA ∩ B
A: Event A
B: Event B
A two-event view of decision under uncertainty.

Quick Answer

Decision under uncertainty is choosing between options whose results aren't known for sure, usually by comparing expected values or risk profiles. List each option's possible outcomes with their probabilities, compute each option's expected value ∑pixi, then look at how widely the outcomes spread and how bad the worst one is. You use it for insurance, warranties, game strategy and business bets: any choice where a big possible payoff competes with a safer one.

Why This Matters

Without a method, uncertain choices get made by the most vivid outcome: the jackpot, the disaster, the story a friend told. Expected value puts every outcome on one scale, weighted by how likely it is, and a look at risk shows when a higher average isn't worth it, for instance when the worst case is something you can't afford. That is how insurers set prices, and why people still buy insurance.

Intuitive Explanation

Offer someone a choice: $100 for sure, or a 1% chance at $5,000. The $5,000 grabs attention, but imagine making this choice 100 times. The sure option pays $100 each time, $10,000 in all. The gamble wins about once, $5,000 in all. Per choice that is 0.01×5000=50 dollars, the gamble's expected value. On average the boring option is twice as good.

Expected value is a long-run average, so it is the right guide for choices repeated many times, like a store setting prices across thousands of sales. For a choice made once, also look at risk: how spread out the outcomes are and how bad the worst one is. Option A pays $500 for sure; option B pays $1,200 with probability 0.5 and loses $100 otherwise. B's expected value is 0.5(1200)+0.5(−100)=550 dollars, more than A's 500, yet someone who cannot afford to lose $100 this month may sensibly take A.

Insurance shows both ideas. Suppose a phone plan costs $8 a month and 3% of phones need a $200 repair each month. The insurer's expected cost per customer is 0.03×200=6 dollars, so it gains about $2 a month per customer. For the customer, buying has a lower expected value than not buying, yet it can still be sensible if a sudden $200 bill would be a real hardship.

Computing one expected value is not yet a decision. "What is this ticket's expected value?" has a single numeric answer; "Should you buy it?" needs that number for every option, plus a judgment about risk.

Think of it like this. Choosing a supermarket checkout line is a small version of this. You can't know which line will move fastest, so you weigh what you see (how many shoppers, how full their carts) and pick the best average bet, avoiding the line where one price check could stall everything. Many real decisions happen only once, though, so no long run averages away bad luck.

Core idea. Decision under uncertainty means choosing the option with the best probability-weighted outcome, not the one with the flashiest single result.


When to Use

Use it when you must choose between options and at least one leads to uncertain outcomes with known or estimated probabilities: buy insurance or not, take a sure prize or a gamble, which strategy to play. Compute each option's expected value, then compare risk and the worst case. If a question asks only for one option's expected value, that is a calculation, not a decision. If no probabilities can be estimated at all, compare worst cases instead.

Tip. Ask: Am I choosing an action by weighing each option's outcomes against their probabilities?

How to Recognize It

Before computing anything, list the options you can actually choose between, the outcomes each one can lead to, and whether you know or can estimate the probability of each outcome.

  1. Does the problem end by asking which option to choose, not just what one option is worth?

    Yes means compute the expected value of every option and compare them. If it asks only for one game's or one ticket's expected value, that is an expected-value calculation, and the work stops at one number.

  2. Do two options have similar expected values but very different worst cases?

    Yes means expected value alone will not decide it: a sure $50 and a coin flip for $100 or nothing both average $50. Look at the risk, meaning how far outcomes spread and how bad the worst is, and whether you can afford it.

  3. Will you face this choice only once, rather than many times?

    Yes means a positive expected value does not protect you from one bad outcome, so give the worst case more weight. Expected value describes the average over many repeats, which is why casinos and insurers, who repeat the bet millions of times, rely on it.

  4. Can you put a probability, even a rough one, on each outcome?

    No means expected value cannot be computed, so compare the options by their worst or best cases instead. If you are only describing how likely the outcomes are, with no choice to make, that is probabilistic thinking.

Decision Under Uncertainty compared with Expected value, Probabilistic thinking, and Risk

All four rows involve outcomes you cannot predict for sure. Ask what the question wants at the end: one probability-weighted number (expected value), how widely the outcomes vary (risk), a way of reasoning about chance in general (probabilistic thinking), or an actual choice between options (decision under uncertainty).

Decision Under Uncertainty

Meaning
Use this when you must choose among options with uncertain outcomes by comparing their expected values or risk.
Key test
Am I choosing an action by weighing each option's outcomes against their probabilities?
Example
Option A: win $40 for sure. Option B: 50% chance of $100, 50% chance of $0. By expected value, which is better?

Expected value

Meaning
Computes one option's probability-weighted average; the input to the decision.
Key test
Use when you need the number for a single option, not the final choice between options.
Formula
E(X)=∑xiP(xi)
Example
EV of a single bet

Probabilistic thinking

Meaning
The broad mindset of reasoning in likelihoods, without necessarily choosing.
Key test
Use when reasoning about uncertainty in general, not committing to one option.
Example
Thinking 'rain is 40% likely'

Risk

Meaning
Describes the spread/variability of outcomes, one factor a decision weighs.
Key test
Use when characterizing how variable an option is, not the act of choosing.
Example
A bet that could win big or lose all

Formula & Notation

Worked Examples

Example 1 — Choose the better bet

Easy

Option A: win $40 for sure. Option B: 50% chance of $100, 50% chance of $0. By expected value, which is better?
Show the steps and the answer
  1. List each option's outcomes: A gives $40 with probability 1; B gives $100 or $0, each with probability 0.5.

    Expected value needs every outcome paired with its probability.

  2. A's expected value: 1×40=40 dollars.

    A sure outcome is its own expected value.

  3. B's expected value: 0.5(100)+0.5(0)=50 dollars.

    Weight each payoff by its chance and add the results.

  4. Compare and weigh the risk: 50>40, but B pays nothing half the time.

    Over many plays B earns $10 more per play; played once, B carries a 50% chance of leaving with nothing, which not everyone can accept.

Answer: Option B (EV $50 > $40), if you can accept the risk

Takeaway: Expected value ranks options by long-run average; risk tells you whether you can live with the bad branch.

Example 2 — Just compute, don't choose

Standard

What is the expected value of Option B above?
Show the steps and the answer
  1. Notice that the question asks about one option, with no choice to make.

    This is an expected-value calculation; no comparison or risk judgment is needed.

  2. Multiply each payoff by its probability: 0.5×100=50 and 0.5×0=0.

    Each product is that outcome's contribution to the long-run average.

  3. Add: 50+0=50 dollars.

    On average B pays $50 per play, although no single play ever pays exactly $50.

Answer: EV=0.5(100)+0.5(0)=$50

Takeaway: An expected value can be an amount that never actually happens; it is an average, not a prediction for one play.

Example 3 — Find the mistake: chasing the biggest prize

Application

A school fundraiser sells $5 tickets for two raffles. Raffle 1 sells 1,000 tickets and has one prize of $2,000. Raffle 2 sells 100 tickets and has five prizes of $50. Alex writes: "Raffle 1 is the better deal because $2,000 is a much bigger prize." Find the mistake and correct it.
Show the steps and the answer
  1. Find each ticket's chance of winning: Raffle 1: 11000; Raffle 2: 5100=0.05.

    Alex compared prize sizes only; each prize must be weighted by its probability.

  2. Expected winnings per ticket: Raffle 1: 11000×2000=2 dollars; Raffle 2: 0.05×50=2.50 dollars.

    This is what a ticket returns on average over many draws.

  3. Subtract the $5 price: Raffle 1: 2−5=−3 dollars; Raffle 2: 2.50−5=−2.50 dollars.

    Both lose money on average, as a fundraiser should, but Raffle 2 loses 50 cents less per ticket.

  4. Name Alex's error: the biggest prize is not the best bet.

    A $2,000 prize at 1-in-1,000 odds is worth less per ticket than a $50 prize at 1-in-20 odds.

Answer: Raffle 2 is the better deal: an expected loss of $2.50 per ticket, against $3 for Raffle 1.

Takeaway: Weight every prize by its probability before comparing; the most dramatic payoff often has the lower expected value.

Where You'll See It

  • Extended warranties on a tablet

    A store offers a 2-year warranty for $60 on a $400 tablet. If about 5% of tablets need a $300 repair in that time, the expected repair cost is 0.05×300=15 dollars. The warranty costs four times its expected value, which is why stores push it; it makes sense only if a $300 surprise would be a real hardship.

  • Fourth-down calls in football

    Suppose a successful fourth-down attempt leads to a drive worth about 4 expected points, a failure hands the opponent field position worth about −2, and a punt is worth about 0.5. With a 50% success rate, going for it is worth 0.5(4)+0.5(−2)=1 point, double the punt. Analysts make this comparison instead of trusting a fear of the dramatic failure.

  • Pricing car insurance

    If 1 driver in 50 in a group files a $10,000 claim each year, the insurer's expected cost per driver is 1000050=200 dollars. It charges more, say $260, to cover costs and profit. Drivers pay above their expected loss to trade a small chance of a ruinous bill for a certain, affordable one.

Common Mistakes

  1. The slip: Choosing the option with the largest possible payoff

    Instead: weight it by probability; a big prize at low odds may have low expected value.

  2. The slip: Ignoring risk and using expected value alone

    Instead: two options with equal EV can differ hugely in how risky they are.

  3. The slip: Letting one dramatic outcome dominate the choice

    Instead: account for ALL outcomes and their chances, not just the memorable one.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. Option A pays $30 for sure. Option B gives a 40% chance of $100 and otherwise $0. Which has the higher expected value?

    Hint: Multiply each payoff by its chance and add.

  2. There is a 30% chance of rain. Carrying an umbrella costs 2 annoyance points whatever happens. Getting wet without one costs 10 points. Should you carry it?

    Hint: Find the expected cost of carrying and of not carrying.

  3. A bakery's cakes cost $5 to make and sell for $12. Demand is 10 cakes or 20 cakes, each with probability 0.5. Unsold cakes are thrown away. Compare the expected profit of making 10 versus 20.

    Hint: For 20 cakes, find the profit in each demand case first.

  4. Option C pays $50 for sure. Option D pays $0 or $100 with equal chance. Find each expected value. Which option is riskier?

    Hint: Same average does not mean same spread of outcomes.

  5. You have $1,000 you need for rent next week. A bet offers a 50% chance of turning it into $2,200 and a 50% chance of losing it all. Its expected value is $1,100. Should expected value alone decide?

    Hint: What happens to you in the bad outcome?

  6. Lottery A: $1 ticket, a 1 in 1,000,000 chance of $500,000. Raffle B: $1 ticket, a 1 in 100 chance of $150. Tom picks A because its prize is bigger. Correct his choice.

    Hint: Weight each prize by its chance, then subtract the ticket price.

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

If insurance has a negative expected value for the buyer, why does anyone buy it?

Because the average is not the only thing that matters. An insurer must collect more in premiums than it pays out, so buyers lose a little on average. In exchange, they swap a small chance of a ruinous loss, such as a $300,000 house burning down, for a certain, affordable payment. When the worst outcome would be unbearable, paying to remove it can be the wise choice.

Would you take a sure $1 million or a 50% chance at $3 million?

The gamble has the higher expected value, 0.5×3=1.50.5\times3=1.5 million dollars, yet most people take the sure million, and that is reasonable. The first million changes a life far more than the next two would, so losing the sure thing would hurt more than the extra money would help. Economists call this diminishing utility: what money is worth to a person grows more slowly than the amount.

What if you cannot estimate any probabilities at all?

You can still compare options by their extremes. The maximin rule picks the option whose worst outcome is least bad, a cautious choice when you could be ruined. The maximax rule picks the option with the best possible outcome, an optimist's choice. A third rule picks the option you would regret least, whichever outcome turns up. None of these needs a probability.

Is a decision bad if it turns out badly?

Not necessarily. A good decision is one that made sense with the information and probabilities you had at the time. Taking an umbrella on a day with a 70% chance of rain was wise even if it stayed dry, and riding without a seatbelt was unwise even if you arrived safely. Judging decisions only by their results rewards luck and punishes sound reasoning.

Learning Path

Decision Under Uncertainty

You are here

Next →

Nothing after this yet
You need Expected Value (the probability-weighted average), Risk (how spread out the outcomes are) and Probabilistic Thinking first. This concept adds a full decision method: list outcomes, compute each option's expected value, then weigh risk and the worst case. It prepares you for insurance, finance and game-strategy problems, and for hypothesis tests, where you choose an action knowing that either kind of error is possible.

See Also