Take It or Leave It
The puzzle
You are given £100 to split with a stranger. You propose a division; they accept or refuse. If they accept, the money is split as you proposed. If they refuse, neither of you gets anything. One round only, no negotiation.
What should you offer, and what should they accept?
Stuck?
One at a time, smallest first. Open only as many as you need — the whole point of the puzzle is the part you work out yourself.
Hint 1
For the responder, is £1 better than £0?
Hint 2
Now ask whether real people behave that way - and whether the proposer should count on it.
Show the answer Hide the answer
Theory says offer £1 and they should take it. In practice offers below about 30% are usually refused, so the sound offer is close to a third or more.
How to get there
Working backwards: the responder choosing between £1 and nothing should take the £1. Knowing that, the proposer should offer the minimum - keeping £99.
Experiments say otherwise, consistently and across cultures. Offers below 20-30% are refused about half the time. Responders will pay real money to punish what they read as an insult, which is not irrational so much as a different objective - they are not maximising this round's cash.
So the proposer's real problem is not the theory but the refusal rate. Offering £40 almost guarantees acceptance; offering £5 is a coin toss on getting nothing. The expected value of a greedy offer is often lower than that of a fair one, which makes near-fair offers correct even on purely selfish grounds.
The technique worth keeping
Backward induction gives the equilibrium of a sequential game, but only if everyone's objective is what you assumed. When the other side may value something you have not modelled, weigh your options by their actual acceptance rate instead.