Sleeping Beauty Problem Explained: 1/2 or 1/3?
Quick answer: In the Sleeping Beauty problem, Beauty is put to sleep and a fair coin is tossed. If heads, she is woken once (Monday); if tails, she is woken twice (Monday and Tuesday) with her memory of Monday erased. Each time she wakes, what probability should she give to heads? “Thirders” say 1/3, because two of the three possible awakenings follow tails. “Halfers” say 1/2, because she learns nothing new about a fair coin. Philosophers and mathematicians still disagree.
Few puzzles divide smart people as sharply as this one. It needs no advanced math, just a fair coin and a sleeping pill, yet after more than 25 years of debate there is still no consensus. Named by Robert Stalnaker and brought to wide attention by philosopher Adam Elga in 2000, it is a test case for how to reason about probability when you might be in more than one situation.
The Setup
- On Sunday, Sleeping Beauty is told the rules and put to sleep.
- A fair coin is tossed.
- Heads: she is woken on Monday, interviewed, and the experiment ends.
- Tails: she is woken on Monday and interviewed, given a drug that erases that memory, then woken again on Tuesday and interviewed.
- Each time she wakes, she cannot tell which day it is. She is asked: “What is your credence (degree of belief) that the coin landed heads?”
The Thirder Answer: 1/3
There are three possible awakenings, and Beauty cannot tell them apart:
| Coin | Day | Long-run share of awakenings |
|---|---|---|
| Heads | Monday | \(\frac13\) |
| Tails | Monday | \(\frac13\) |
| Tails | Tuesday | \(\frac13\) |
If the experiment is repeated many times, only one third of all awakenings follow heads. Elga also argued: if she were told “it is Monday,” she should give heads and tails equal credence; working backward with conditional probability gives \(P(\text{heads}) = \frac13\) before she is told. Betting supports this too: if she bets on tails at each awakening, she wins twice as often as she loses. (Brush up on basic probability.)
The Halfer Answer: 1/2
David Lewis (2001) replied that Beauty learns nothing new when she wakes: she knew in advance she would be woken at least once, and the coin is fair. Before sleeping, her credence in heads was \(\frac12\); since waking provides no new evidence, it should stay \(\frac12\). Halfers say the thirder count gives tails extra weight just because it produces more interviews, not because it is more likely.
Why the Disagreement Persists
- Different questions. “What fraction of awakenings follow heads?” is clearly 1/3. “What is the chance the coin landed heads?” sounds like 1/2. The debate is about which question Beauty’s credence should answer.
- Self-locating belief. Beauty is uncertain not only about the world (the coin) but about where she is in it (which day). Standard probability theory was not built for that second kind of uncertainty.
- Bigger stakes. The same reasoning appears in arguments about the anthropic principle, the “doomsday argument,” and cosmology.
In the 2020 PhilPapers Survey of philosophers, 27.7% accepted or leaned toward 1/3, 18.7% toward 1/2, and 40.3% were undecided: more thirders than halfers, but no majority. Research continues on both sides: a March 2026 preprint by Jiaxuan Zhang defends the halfer view, while a forthcoming chapter by Tim Button, Daniel Rothschild and Levi Spectre (“Too Clever by Halfing”) challenges halfer approaches. Some philosophers propose “double halfer” and other variants.
Common Misconceptions
“It is just the Monty Hall problem again.” No. In Monty Hall, the host’s action gives you information and there is one right answer. Here the issue is how to count yourself across memory-erased repeats.
“One side has made an arithmetic error.” No. Each side’s arithmetic follows from its premises; the disagreement is about which premises are right. Betting arguments and the “different questions” reading are proposed interpretations, not accepted resolutions.
Watch: The Sleeping Beauty Problem in Video
If you learn better by watching, these videos are worth your time.
Check Your Understanding
Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.
Answer each question and press Submit to see your result right away.
Frequently Asked Questions
What is the Sleeping Beauty problem?
A probability puzzle in which Beauty is woken once if a coin lands heads and twice (with memory erased) if tails, and must state her credence that the coin landed heads each time she wakes.
Is the answer to the Sleeping Beauty problem 1/2 or 1/3?
There is no consensus. Thirders (like Adam Elga) say 1/3; halfers (like David Lewis) say 1/2. The disagreement is about what credence should measure.
Who created the Sleeping Beauty problem?
Robert Stalnaker gave it its name. Adam Elga’s 2000 paper in Analysis launched the modern debate, drawing on earlier examples by Arnold Zuboff and by Michele Piccione and Ariel Rubinstein.
Why does the Sleeping Beauty problem matter?
It tests how to reason about self-locating uncertainty, which matters in debates about the anthropic principle and cosmology.
Keep Exploring
- Monty Hall Problem: Why Switching Wins 2/3 of the Time
- Two Envelopes Paradox Explained: Should You Switch?
- Simpson’s Paradox Explained: 3 Real Examples
Sources and Further Reading
Related to This Article
More math articles
- Benefits of Driving Electric Cars: Is It Cheaper?
- 8th Grade Common Core Math Practice Test Questions
- How To Get A Perfect Score Of 800 On SAT Math?
- How to Graph Absolute Value Function?
- Money Market, Loanable Funds, and Interest Rates
- Surface Area to Volume Ratio: Cell Size Worked Examples
- Hawaii Algebra 1 Free Worksheets: Printable Algebra 1 Practice Worksheets with Worked Solutions
- How to Decode Data Types
- Distributions in Line Plot
- Two-Column Proofs: How to Write Statements and Reasons


What people say about "Sleeping Beauty Problem Explained: 1/2 or 1/3? - Effortless Math"?
No one replied yet.