Game Theory 4: Bayesian Games, Harsanyi Transformation, Bayesian Nash Equilibrium, and Core Applications

A single-file study note on Bayesian games, written in the same style as the dynamic-games page above. The focus is conceptual completeness, readable mathematics, and worked examples that connect directly to the cases discussed earlier: Harsanyi transformation, discrete Bayesian reasoning, first-price auctions, the signaling extension, and why electricity markets are often modeled with Bayesian games.

Incomplete information Types and beliefs Common prior Bayesian Nash equilibrium Signaling extension First-price auction Electricity markets

Table of contents

1. What Bayesian games are incomplete information

A Bayesian game is a game in which at least one payoff-relevant feature is not publicly known. The unknown object is encoded as a type. Each player knows their own type, but may not know the types of the others.

Core picture: a Bayesian game is an ordinary strategic game plus private payoff-relevant information and beliefs about that information.
Main intuition

In an ordinary Nash game, player i compares actions given what others do. In a Bayesian game, player i compares actions given what others do and what others might privately be.

\[ \text{Ordinary game: } u_i(a_i,a_{-i}) \qquad \text{Bayesian game: } u_i(a_i,a_{-i};t_i,t_{-i}). \]

Complete information, incomplete information, perfect information, imperfect information

Concept Meaning Why it matters
Complete information Everyone knows the game structure and every player's payoff function. This is the benchmark in ordinary normal-form analysis.
Incomplete information At least one payoff-relevant variable is privately known or unknown to some players. This is exactly the setting modeled by Bayesian games.
Perfect information Whenever a player moves, they observe the entire past history of play. This is about observed actions, not private types.
Imperfect information At some decision points, a player does not observe all earlier actions. A game can be Bayesian and still be imperfect-information or perfect-information depending on the timing.
Key distinction. Incomplete information is about unknown characteristics, such as costs, valuations, or risk types. Imperfect information is about unknown past actions. The two ideas are different and often coexist.

What counts as a type?

A type can represent any privately known payoff determinant: a buyer's valuation, a firm's marginal cost, a worker's productivity, an insurer's risk class, or a generator's true operating cost.

Why Bayesian?

Players do not know the true type profile, so they attach probabilities to possible types and choose actions that maximize expected payoff under those beliefs.

Mathematical punchline. The jump from Nash to Bayesian Nash is structurally small: add a type variable, add a belief distribution, and optimize expected payoff conditional on your own type.

2. Formal model and notation types, actions, beliefs, strategies

Static Bayesian game

A standard finite Bayesian game can be written as

\[ G=\bigl(N,(A_i)_{i\in N},(T_i)_{i\in N},p,(u_i)_{i\in N}\bigr). \]

The ingredients are:

  • Players. \(N=\{1,2,\dots,n\}\).
  • Action sets. Each player \(i\) chooses an action from \(A_i\).
  • Type sets. Each player has a type \(t_i\in T_i\).
  • Common prior. \(p(t_1,\dots,t_n)\) is a probability distribution over type profiles.
  • Payoff functions. For each player \(i\), \(u_i(a_1,\dots,a_n;t_1,\dots,t_n)\).

Pure and mixed Bayesian strategies

Pure strategy

A pure Bayesian strategy for player \(i\) is a function

\[ s_i:T_i\to A_i. \]

Interpretation: for each possible type, the strategy specifies what action that type takes.

Mixed strategy

A mixed Bayesian strategy assigns a distribution over actions to each type:

\[ \sigma_i(\cdot\mid t_i)\in \Delta(A_i). \]

Interpretation: type \(t_i\) may randomize over actions.

Beliefs induced by the common prior

If types can be correlated, player \(i\)'s belief about the other players' types after observing their own type \(t_i\) is obtained by conditioning the common prior:

Conditional belief
\[ \mu_i(t_{-i}\mid t_i) = \frac{p(t_i,t_{-i})}{\sum_{\tau_{-i}\in T_{-i}}p(t_i,\tau_{-i})}. \]

When types are independent, this simplifies substantially because knowing your own type does not change the distribution of the others' types.

Type

A privately known characteristic, such as cost or valuation.

Belief

A probability assessment over the unknown types of the other players.

Strategy

A full plan that tells each type what action to choose.

Why strategies are functions. In a Bayesian game, the player does not choose one action once and for all. The player chooses a rule: what to do if my type is high, what to do if my type is low, and so on.

3. Harsanyi transformation Nature draws types first

Harsanyi's idea is to convert incomplete information into an ordinary extensive-form game by adding a move by Nature at the beginning. Nature draws the type profile, each player observes their own type, and play then continues with those types fixed in the background.

Transformation idea

Instead of saying, "player i does not know player j's true cost," we say:

  1. Nature draws a type profile \(t=(t_1,\dots,t_n)\sim p\).
  2. Player \(i\) privately observes \(t_i\).
  3. Players choose actions based on their observed types and beliefs.
Nature type H with prob p type L with prob 1-p player sees H player sees L action a action b terminal payoffs
The unknown type is treated as an initial random move by Nature. After that move, each player knows their own type but not necessarily the others' types.

Why this is useful

1
It standardizes incomplete information. Private characteristics become explicit objects in the model.
2
It clarifies beliefs. The common prior tells us exactly how uncertainty is represented.
3
It clarifies strategies. A player's decision rule becomes a type-contingent strategy.
4
It lets us use standard equilibrium ideas. Once the transformation is made, Bayesian Nash equilibrium becomes the natural solution concept for the resulting strategic problem.
Short memory aid. Harsanyi transformation means: Nature draws the types first, players observe what they are allowed to observe, and the rest of the game is solved normally.

4. Bayesian Nash equilibrium expected-payoff optimality by type

A Bayesian Nash equilibrium (BNE) is a strategy profile in which every type of every player is choosing a best response, given beliefs about other players' types and given the strategy functions of the other players.

Expected payoff for finite type spaces

Given a pure-strategy profile \(s=(s_1,\dots,s_n)\), player \(i\)'s expected payoff conditional on type \(t_i\) is

\[ U_i(s\mid t_i) = \sum_{t_{-i}\in T_{-i}} \mu_i(t_{-i}\mid t_i) \,u_i\bigl(s_i(t_i),s_{-i}(t_{-i});t_i,t_{-i}\bigr). \]
Expected payoff for continuous type spaces

If types are continuous, the sum becomes an integral:

\[ U_i(s\mid t_i) = \int_{T_{-i}} u_i\bigl(s_i(t_i),s_{-i}(t_{-i});t_i,t_{-i}\bigr) \,d\mu_i(t_{-i}\mid t_i). \]
Definition of BNE

A strategy profile \(s^*\) is a Bayesian Nash equilibrium if for every player \(i\), every type \(t_i\), and every feasible alternative action \(a_i\in A_i\),

\[ U_i(s^*\mid t_i) \ge \sum_{t_{-i}\in T_{-i}} \mu_i(t_{-i}\mid t_i) \,u_i\bigl(a_i,s^*_{-i}(t_{-i});t_i,t_{-i}\bigr). \]

So each type must be best responding in expected-value terms.

How BNE differs from ordinary Nash equilibrium

Ordinary Nash equilibrium asks whether an action is a best response to others' actions. BNE asks whether a type-contingent action is a best response to others' type-contingent strategies under beliefs about the unknown type profile.

What is the main extra object?

The extra object is not a new optimization method. It is the player's belief about hidden types. Once the belief is specified, the computation is expected-utility maximization.

Practical solution checklist

1
Identify the types. Determine what information is privately known and how it affects payoff.
2
Write the strategy space. A strategy should specify an action for every type, not just one realized action.
3
Condition on your own type. Fix \(t_i\) and compute the expected payoff from each feasible action using beliefs over \(t_{-i}\).
4
Find best responses type by type. Different types may optimally choose different actions.
5
Look for consistency. A BNE is a fixed point of type-contingent best responses.
Common beginner mistake. Treating a strategy as one action. In Bayesian games, the strategy is a map from type to action. That is the whole point of the model.

5. Worked discrete example: posted pricing with private buyer valuation compact BNE computation

This example supports the earlier discussion of Harsanyi transformation and Bayesian reasoning. A seller posts a price and a buyer privately knows whether they are a high-value or low-value type.

Environment
  • The seller chooses a price \(P\in\{5,9\}\).
  • The buyer's type is \(H\) or \(L\).
  • Valuations are \(v_H=10\) and \(v_L=4\).
  • Beliefs: \(\Pr(H)=0.6\), \(\Pr(L)=0.4\).
  • After observing the price, the buyer chooses \(\text{Buy}\) or \(\text{Not Buy}\).

To be formally complete, the buyer's strategy is a mapping from type and observed price to a response:

\[ s_B:\{H,L\}\times\{5,9\}\to\{\text{Buy},\text{Not Buy}\}. \]
Buyer type If P = 5 If P = 9 Best response pattern
H with value 10 Buy payoff = 10 - 5 = 5 Buy payoff = 10 - 9 = 1 Buy at both prices
L with value 4 Buy payoff = 4 - 5 = -1 Buy payoff = 4 - 9 = -5 Reject at both prices

Buyer's best responses

Type-contingent behavior
\[ s_B(H,5)=\text{Buy}, \qquad s_B(H,9)=\text{Buy}, \] \[ s_B(L,5)=\text{Not Buy}, \qquad s_B(L,9)=\text{Not Buy}. \]

Seller's expected payoff

The seller knows that only type \(H\) buys. Therefore the sale probability is \(0.6\) under either feasible price, and expected profit is:

Seller's comparison
\[ \Pi_S(5)=0.6\times 5=3, \qquad \Pi_S(9)=0.6\times 9=5.4. \]
Bayesian equilibrium outcome. The seller chooses \(P=9\). The high-value buyer buys and the low-value buyer rejects. This is a Bayesian best-response configuration.
Technical note. Because this example is sequential, one can also formulate it in extensive form and refine behavior with sequential rationality. In this simple case the Bayesian best-response calculation and the extensive-form reasoning give the same outcome.

6. Continuous types: first-price auction with n bidders general formula

This is the most useful continuous-type Bayesian example for developing technique. Each bidder privately knows their valuation and submits a bid simultaneously. The highest bidder wins and pays their own bid.

Auction environment
  • There are \(n\) risk-neutral bidders.
  • Each valuation \(v_i\) is independently drawn from distribution \(F\) on \([0,\bar v]\) with density \(f\).
  • Each bidder uses the same strictly increasing bid function \(b(v)\).
  • We seek a symmetric pure-strategy Bayesian Nash equilibrium.

Step 1: Write the payoff from mimicking another type

Fix a bidder with true valuation \(v\). If they bid as if they were type \(z\), they submit bid \(b(z)\). Since the bid function is strictly increasing, they win exactly when every rival valuation is below \(z\). Thus the winning probability is \(F(z)^{n-1}\), so expected profit is

Expected profit from bidding like type z
\[ \Pi(z;v)=\bigl(v-b(z)\bigr)F(z)^{n-1}. \]

Step 2: First-order condition at the symmetric optimum

In equilibrium, type \(v\) should choose \(z=v\). Differentiate \(\Pi(z;v)\) with respect to \(z\) and set \(z=v\):

Differentiation
\[ \frac{\partial \Pi}{\partial z} = -b'(z)F(z)^{n-1} + \bigl(v-b(z)\bigr)(n-1)F(z)^{n-2}f(z). \]

At \(z=v\), the equilibrium condition becomes

\[ -b'(v)F(v)^{n-1} + \bigl(v-b(v)\bigr)(n-1)F(v)^{n-2}f(v)=0. \]

Step 3: Rearrange into a linear differential equation

ODE for the equilibrium bid function
\[ b'(v)+\frac{(n-1)f(v)}{F(v)}b(v) = \frac{(n-1)f(v)}{F(v)}v. \]

Using the integrating factor \(F(v)^{n-1}\), one obtains

\[ \frac{d}{dv}\Bigl(F(v)^{n-1}b(v)\Bigr) = (n-1)vf(v)F(v)^{n-2}. \]

Step 4: Solve the equation

Imposing the natural boundary condition \(b(0)=0\), the symmetric equilibrium bid function is

General n-bidder formula
\[ b(v) = \frac{\int_0^v (n-1)tf(t)F(t)^{n-2}\,dt}{F(v)^{n-1}} = v-\frac{\int_0^v F(t)^{n-1}\,dt}{F(v)^{n-1}}. \]

Uniform special case

If valuations are uniform on \([0,1]\), then \(F(v)=v\) and \(f(v)=1\). Substituting into the formula yields

Uniform distribution result
\[ b(v) = v-\frac{\int_0^v t^{n-1}\,dt}{v^{n-1}} = v-\frac{v^n/n}{v^{n-1}} = \frac{n-1}{n}v. \]
valuation v bid truthful line b(v) = v first-price equilibrium b(v) = (n-1)/n v bid shading
In a first-price auction, equilibrium bids lie below valuation. The gap reflects the tradeoff between increasing winning probability and lowering payment conditional on winning.
Interpretation. As \(n\) increases, \((n-1)/n\) moves closer to \(1\), so bids become more aggressive and approach truthful bidding. Stronger competition reduces bid shading.

7. From Bayesian games to signaling games why signaling is deeper

Static Bayesian games treat types as hidden and actions as direct responses to those hidden types. Signaling games add an extra strategic layer: actions themselves can reveal, conceal, or distort beliefs about type.

Object Static Bayesian game Signaling game
Unknown variable Type is privately known Type is privately known
Timing Usually simultaneous or one-stage incomplete information A sender moves first, then a receiver updates beliefs and responds
Main equilibrium object Type-contingent strategies Type-contingent signals, posterior beliefs, and receiver actions
New issue Expected-payoff maximization under hidden types Actions change beliefs, and beliefs change later actions
Posterior belief after a signal

If the sender sends signal \(m\), the receiver forms a posterior

\[ \mu(t\mid m). \]

Now the strategic chain is

\[ \text{signal } m \to \text{belief } \mu(\cdot\mid m) \to \text{receiver action} \to \text{sender payoff}. \]
Why signaling is more subtle.

A player's action now has two effects at once: a direct payoff effect and an informational effect. The sender asks not only "what do I gain from this move?" but also "what will the receiver infer from this move?"

Canonical distinction.

Separating equilibrium: different types choose different signals. Pooling equilibrium: different types choose the same signal. The distinction is about whether information is revealed endogenously.

Good bridge intuition. Static Bayesian games are about acting under hidden information. Signaling games are about acting in ways that actively shape what others believe about hidden information.

8. Why electricity markets often use Bayesian games application mapping

Electricity markets are a natural environment for Bayesian modeling because they combine private information, strategic bids, and market-clearing rules under physical network constraints.

Private information

Generators know more than rivals do about marginal costs, outage risk, start-up constraints, fuel positions, ramping capability, or the true flexibility of their assets.

Strategic interaction

Participants submit offers and quantities while anticipating the bids of competitors and the clearing rule used by the market operator.

Bayesian-game mapping for electricity markets
  • Type. Cost, available capacity, outage probability, flexibility, or demand elasticity.
  • Action. Bid price, offered quantity, block structure, reserve offer, or demand response commitment.
  • Belief. Distribution over rivals' costs, capacity states, and likely bidding behavior.
  • Payoff. Revenue net of production cost, adjusted by dispatch, congestion, reserve payments, penalties, and market rules.
private type cost, capacity, risk flexibility, demand state strategic bid price, quantity, blocks reserve offer market clearing dispatch plus network constraints payoff profit or loss
The Bayesian logic is natural: hidden operational conditions shape bids, bids feed into constrained market clearing, and the resulting prices and quantities determine profit.
Why ordinary complete-information models are often too strong.

Rivals do not observe each other's true marginal costs, start-up costs, or real-time capability with full precision. Modeling everyone as fully informed can erase exactly the strategic uncertainty that matters.

Where Bayesian games are especially useful.

They are most useful when the research question is strategic bidding under private information. They are not the only tool in electricity economics; optimization, stochastic programming, and equilibrium-with-network models remain central for many other questions.

One-sentence summary. Electricity markets often call for Bayesian games because participants submit strategic offers while holding private information that materially affects clearing outcomes and profit under network constraints.

9. Summary and checklist what to remember

Concept What it adds What to compute
Bayesian game Private payoff-relevant information Type spaces, beliefs, and type-contingent strategies
Harsanyi transformation Turns hidden characteristics into Nature's initial move The induced type structure and information pattern
Bayesian Nash equilibrium Best-response optimality for every type Expected payoff conditional on own type
Continuous-type auction analysis Functional strategy rather than a single action The equilibrium bid function \(b(v)\)
Signaling extension Actions affect beliefs and later responses Signals, posterior beliefs, and receiver best responses
Bayesian games add types and beliefs; Bayesian Nash equilibrium says each type must be optimizing in expectation; signaling games go one layer deeper by making actions informative.
Checklist 1. Ask first: what information is private, and whose payoff does it change?
Checklist 2. Write the strategy correctly as a map from type to action, or from type and history to action in dynamic settings.
Checklist 3. Condition on own type and compute expected payoff under beliefs about the unknown types of others.
Checklist 4. For continuous types, expect to solve for a function, not a single number.
Checklist 5. If actions are observed before others respond, ask whether the richer signaling or sequential-belief framework is needed.
Final takeaway. The mathematical core of Bayesian games is not mysterious: add types, add beliefs, take conditional expectations, and solve best responses. The real richness comes from what those types represent in economic environments such as auctions, signaling, labor markets, insurance, and electricity markets.