Alternating-offer Bargaining, explained.
Alternating-offer bargaining studies how proposals, counterproposals and the cost of delay shape an agreement.
Why it happens
A pie begins at 100 and shrinks by a common factor δ each round. A stationary bot requires δ/(1+δ) of the current pie. Rejection advances a round; accepting its next counteroffer gives you δ/(1+δ) of the next pie. The first-mover benchmark is 1/(1+δ).
Alternating-offer bargaining studies how proposals, counterproposals and the cost of delay shape an agreement.
Read the result
A percentage applies to the current pie, not the original hundred. Track absolute amounts as rejected demands consume value through delay.
A worked example
A larger demand and a smaller pie
At discount factor 0.8, the bot requires about 44.4% of the current pie, leaving the first proposer about 55.6%.
Request sixty percent. The bot rejects and the pie falls to eighty, even though no agreement has been reached.
Consider a smaller request or the bot's next counteroffer. Compare amounts after delay rather than celebrating a percentage alone.
OPTIONAL DEEPER DETAILGo deeper: inside the model
Inside this model
A pie begins at 100 and shrinks by a common factor δ each round. A stationary bot requires δ/(1+δ) of the current pie. Rejection advances a round; accepting its next counteroffer gives you δ/(1+δ) of the next pie. The first-mover benchmark is 1/(1+δ).
Where this idea is useful
A practical use
A negotiator can compare a larger requested share with the value destroyed while waiting for agreement.
A common misconception
“More patient bargaining always improves both parties' outcome.”
Patience affects strategic shares and the cost of delay. The model's common discounting and stationary policy do not imply that prolonging real negotiations helps everyone.
What this explanation leaves out
- The bot implements a stylised infinite-horizon stationary policy with common discounting, not a forecast of human behaviour. Outside options, deadlines, asymmetric information and unequal patience are omitted.
Is the stationary bot predicting a human negotiator?
No. It applies a transparent reservation-share rule from a stylised bargaining model, without psychology, outside options or incomplete information.
How much value would disappear while you wait for the additional share you want?
Associated thinkers
Associations marked provisional are awaiting source review.
Further reading
Explore the original research or the teaching reference behind this experiment.