What the book covers

The book looks at rules and negotiation at six levels:

1. Mechanism design as “reverse game theory”. Game theory usually takes the rules as given and predicts the outcome; mechanism design starts from the outcome we want and searches for rules that deliver it. The core ideas are incentive compatibility and the revelation principle, illustrated by “I cut, you choose”, by free-riding on public goods and by the Clarke pivot mechanism.
2. Auctions. Why bidding one’s true value is weakly dominant in Vickrey’s second-price auction; why “shading” a bid in a first-price auction is an equilibrium strategy rather than a gamble; the revenue equivalence theorem; and the lovely but lonely side of the Vickrey–Clarke–Groves (VCG) mechanism.
3. The Myerson–Satterthwaite impossibility theorem. When buyer and seller both hold private information, no rule can satisfy four conditions at once; in the simplest uniform example, even the best rule realises only about 84.4 % of the potential gains from trade.
4. Bargaining. The Nash bargaining solution and Rubinstein’s alternating-offers model, and the sense in which they give the same answer; in which models “your leverage depends on your BATNA” holds, and in which it does not.
5. Lemons and signals. How asymmetric information can make a market shrink round after round; when a costly, verifiable commitment can separate the “good” from the “bad”, and the social cost that separation itself carries.
6. Goodhart’s law. Why a statistical regularity breaks down once it is used for control; how manipulable metrics, random audits and multidimensional checks can be combined.

Chapter 7 reads these modern results beside the Han Feizi on “holding names to account for reality” and “the two handles”, the Guanzi chapter “Making the Law Clear”, the Analects on “to govern is to correct” and “lead them with virtue”, and Mencius’s warning about “superiors and inferiors competing for profit”, and states that ancient Legalism served the centralised power of the ruler while modern mechanism design serves voluntary contracts between equal parties and public welfare, a fundamental difference of purpose. Chapter 8 offers a “four-step mechanism and bargaining checklist” and a pre-registered dyad-randomised controlled trial.

The book complements earlier volumes of 5D Library without repeating their models: Book 16 of the series dealt with multitask principal–agent problems, relational contracts and reputation, Book 19 with the Kelly criterion and cash buffers, and Book 24 with the crowding-out of intrinsic motivation by extrinsic rewards. Where those topics come up here, the book only cross-refers briefly.

Claims corrected after checking

Several points in the author’s original framework needed correcting or qualifying after checking:

1. Bid shading in a first-price auction is not “gambling” but an equilibrium strategy. In the standard model with independent private values uniform on [0, 1], risk-neutral bidders and $n$ bidders, the symmetric equilibrium bid is $b(v) = (n-1)v/n$: each bidder trades off “bid a little more, win a little more often” against “win, but earn a little less”. In experiments, bids are usually above this risk-neutral equilibrium.
2. The weak dominance of truthful bidding has preconditions. It holds only with private values, a single item (or the VCG setting) and no collusion among bidders; second-price auctions in the laboratory also show widespread overbidding. VCG can yield very low revenue, revenue that falls as bidders are added, and is exposed to coordination among losing bidders and to false identities (Ausubel and Milgrom 2006).
3. The exact statement of the Myerson–Satterthwaite theorem. The four conditions are Bayesian incentive compatibility, interim individual rationality, ex post efficiency and budget balance (no outside subsidy), and the supports of the two parties’ valuations must overlap; the conclusion is that they cannot all hold, not that “negotiation always fails”. In the uniform example, the efficiency of about 84.4 % (exactly 54/64 = 84.375 %) is the best achievable by any rule meeting the other conditions; the linear equilibrium of the Chatterjee–Samuelson double auction attains it, but that double auction has many other equilibria.
4. “Your leverage depends on your BATNA” holds only in one class of models. In Nash’s axiomatic bargaining solution, the fallback is the threat point, and a better fallback means a larger share. In alternating-offers models with outside options, however, the “outside option principle” applies: an outside option matters only if it is better than the share you would get anyway; otherwise it has no effect on the outcome (Sutton 1986; Binmore, Shaked and Sutton 1989). What persistently moves shares is payoffs during the negotiation (“inside options”) and the parties’ patience. The term BATNA comes from Fisher and Ury’s 1981 Getting to Yes.
5. “A costly verifiable commitment that a cheat cannot afford” must be written as the conditions for a separating equilibrium. A signal works only under a “single-crossing condition”: the same signal is more expensive for the type who would imitate. Its cost must be high enough that imitation does not pay, yet low enough that the genuine type still wants to pay it. A pure signal creates no value and is a social waste, and separating equilibria are usually not unique.
6. Goodhart’s original wording and the popular version come from different sources. Goodhart wrote in 1975 that “any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes”; “when a measure becomes a target, it ceases to be a good measure” is the anthropologist Marilyn Strathern’s 1997 generalisation in European Review, where she notes that Keith Hoskin calls it Goodhart’s law.
7. Sources of the classical texts. The story of Marquis Zhao of Han, who fell asleep drunk, was covered by the keeper of caps, and “punished both the keeper of robes and the keeper of caps”, comes from the Han Feizi chapter “The Two Handles” (Er bing), not “Defining the Law” (Ding fa); the latter contains “technique means granting office according to ability, holding names to account for reality …”. The received text of the Guanzi chapter “Making the Law Clear” (Ming fa) reads “Therefore when the former kings governed the state, they did not let their intentions roam outside the law, nor grant favours within the law”; “Explaining ‘Making the Law Clear’” (Ming fa jie) is a separate, line-by-line commentary on it.
8. The lemons spiral is a theoretical mechanism, not a universal fact. A direct test in the used pickup-truck market found no evidence that traded trucks were of lower quality (Bond 1982), suggesting that warranties, reputation and inspection often offset much of the information asymmetry.

Each chapter states these issues where they arise, and they are collected in the Editorial Notes published with the book.

How to read the boxes

As in the other books of 5D Library, different kinds of statement are marked separately.

About this edition

This is version 1.0.0, the first complete edition: a preface and eight chapters, in Chinese and English. The book contains 16 Python programs and 4 vector figures. The text, boxes and code were drafted with an AI writing tool from the author’s title, framework and notes; the author is responsible for reviewing and revising them and for checking every source. Corrections and suggestions of literature the author has missed are welcome through the 5D Library website.