Dollar Auction, explained.
The dollar auction is a contest where both the highest bidder and the runner-up pay their final bids, while only the highest bidder receives the prize.
Why it happens
A runner-up may compare a certain existing loss with the chance to regain the lead. That local comparison can encourage another bid even after the winning bid would exceed the prize. The payment rule creates the trap.
Paying the second-highest bid as well as the winning bid can create incentives to escalate even after the prize ceases to justify the total cost.
Read the result
Track your standing bid after the bot responds. Stopping costs that bid; winning costs the final winning bid minus the prize. The bot’s disclosed cutoff lets you inspect a finite escalation without guessing its intentions.
A worked example
Following a bid past the prize
The prize is 20. After your bid of 19, the bot bids 20.
You could stop and pay 19, or bid 21; the bot is programmed to continue up to 26.
Eventually bidding 27 wins the prize but nets minus seven, showing why winning a contest need not create a positive return.
OPTIONAL DEEPER DETAILGo deeper: inside the model
Inside this model
You and a disclosed bot alternate bids one token apart. The bot bids up to 26. Your standing bid is owed if you stop as runner-up; if the bot stops, you receive 20 minus your winning bid. No real money is involved.
Where this idea is useful
A practical use
A contest can make continuing appear attractive relative to accepting an existing loss, even while worsening total returns.
A common misconception
“Winning means the strategy was profitable.”
Net return includes what the winner paid. A prize can be smaller than its acquisition cost.
What this explanation leaves out
- The bot has a fixed cutoff, not human judgment. This is a finite classroom version of the auction.
Is the bot strategically intelligent?
No. Its fixed cutoff is visible so you can study the payment incentives without treating the outcome as human behavior.
Before entering a contest, what exit rule would limit exposure to escalation?
Associated thinkers
Further reading
Explore the original research or the teaching reference behind this experiment.