This scene shows why honest bidding wins in a second-price auction. There is one opponent, and they bid honestly.
Your bid does not set the price you pay. Winning means paying the opponent's bid. So bidding above your value gains nothing. It only creates wins that cost too much. Bidding lower throws away wins you could afford. As the formula below shows, the loss grows with the square of the gap.
For example, your class auctions off one ⚽ ball. You value it at 6,000 won. A friend bids 4,000 won. Bid your honest 6,000 and you win. You pay only the friend's 4,000 💰 won. Bid 8,000 and the price stays the same. But if the friend had bid 7,000, you would overpay.
My value is v=60 and the opponent bids truthfully.
If I overbid with b=80, compute my expected payoff, win probability, and expected payment, and compare with truthful bidding (b=60).
U(b;v)=∫0b100v−vodvo=100vb−b2/2
U(80;60)=1004800−3200=16<18=U(60;60)
v=60, b=80 — overbidding loses 2.Show proof
Expected payoff is U(b;60)=(60b−b2/2)/100. At b=80, U=(4800−3200)/100=16; at truthful b=60, U=(3600−1800)/100=18 — overbidding loses 18−16=2.
The source of the loss: when vo∈(60,80] you win but pay vo>60=v, earning negative surplus.
The win probability rises to 80/100=0.8, but expected payment rises with it to 802/200=32, and the net effect is negative.
Since dU/db=(60−b)/100<0 for b>60, raising the bid past the value only shrinks the payoff.
dbdU=100v−b,dbdUb=80=−0.2
In the same situation, what is the expected payoff of underbidding with b=40?
Compare with truthful bidding and prove in general that no bid can beat the truth.
U(40;60)=1002400−800=16<18=U(60;60)
v=60, b=40 — underbidding also loses 2.Show proof
U(40;60)=(2400−800)/100=16 — lower than truthful bidding by 2.
Whenever vo∈(40,60] you lose, yet winning there would have earned the positive surplus 60−vo>0.
The loss from underbidding is the opportunity cost of forfeited wins.
In general, U(v;v)−U(b;v)=200v2−100vb−b2/2=200(v−b)2≥0 — no bid beats truthful bidding. This is the computational proof that truth-telling is a weakly dominant strategy.