Skip to content
Omnikalc

Two envelopes paradox calculator: should you switch, given what you see?

Enter the amount you see and the prior ratio to get the chance of holding the smaller envelope and the expected value of switching.

Updated on
Your inputs
$

Value out of range

Value out of range

Results

Expected value if you switch

$37.14

Keeping has a higher expected value
Chance that you hold the smaller envelope
28.6 %
Expected gain from switching
-$2.86
Expected gain as a share of your amount
-7.14 %
Margin 1 − 2 × ratio (positive means a finite expected prize)
0.20

The two envelopes problem is one of the best-known puzzles in probability. The two envelopes paradox calculator takes the amount you see and a ratio describing how quickly large amounts become less likely, and gives the probability that you hold the smaller envelope, the expected value of switching, the expected gain and the margin that shows if the expected prize is finite.

How the calculation works

  • Enter the amount in the envelope you opened.
  • Enter the ratio of the prior: the pair of amounts (x, 2x) at each step up the doubling ladder is rr times as likely as the previous pair. A ratio near 0 means small amounts are very likely, and near 1 means large amounts are almost as likely as small ones.
  • The calculator gives the probability that you hold the smaller envelope, which is ratio ÷ (1 + ratio), and the expected value of switching: P(smaller) × 2x + P(larger) × x ÷ 2.
  • Compare the expected value with your amount. The break-even ratio is 1/2.

Example

You open an envelope with $40. The prior has a ratio of 0.4, so the chance that you hold the smaller envelope is 0.4 ÷ 1.4 = 28.6 %. If you switch, you expect 0.286 × 80 + 0.714 × 20 = $37.14, a loss of $2.86 or 7.1 %. At a ratio of exactly 0.5, the expected value of switching equals $40. Above 0.5, switching wins on average, but the expected prize is then infinite.

Why it matters

Working through the numbers shows why the naive argument fails: it ignores that a prior must make large amounts less likely. Changing the ratio shows when switching gains, when it loses, and why an infinite expected prize breaks the argument.

Good to know

The model uses a geometric prior on a ladder of doubling amounts, which is only one choice of prior. Other priors give other answers, and in real life no one knows the prior. The calculator is an illustration of the paradox and not a strategy for a game with money.

The Monty Hall calculator is another famous puzzle about switching, the Bertrand box and boy or girl paradoxes involve conditional probability, and the expected value calculator handles general cases.

Frequently asked questions

What is the two envelopes paradox?

You are given two envelopes, one holding twice as much as the other. You open one and see an amount. A naive argument says that the other envelope holds either half or double with equal chance, so its expected value is 1.25 times your amount, and you should always switch, even before looking, which is absurd.

What is wrong with the naive argument?

It assumes that the other envelope is equally likely to hold half or double whatever amount you see. That is impossible for every amount at once: a proper probability distribution of amounts makes small amounts more likely than large ones when the expected prize is finite. Then the chance of holding the larger envelope is greater than one half, and the gain from doubling is balanced by the loss from halving.

What does the ratio mean?

The calculator uses amounts that double at each step and a prior where each step up is a fixed factor less likely than the one before. The factor is the ratio. With a ratio below 1/2 the expected prize is finite, and switching has a lower expected value than keeping for any amount other than the smallest.

Is there a case where switching is right?

Yes. If you see the smallest possible amount, you know you hold the smaller envelope and switching doubles your money. With a ratio above 1/2, the expected prize is infinite, so comparing expectations loses its meaning, which is another way out of the paradox.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.