Modeling Photographer-Model Interactions as a Signaling Game

1. The Essence of a Signaling Game

A signaling game models situations where one player has private information that another player lacks. The informed player (the "Sender") takes an action (a "signal") that is observed by the uninformed player (the "Receiver"). The Receiver then uses this signal to update their beliefs about the Sender's private information and chooses their own action.

The key components are:

• Sender (Player 1): The player with private information.
• Type: The Sender's private information. Nature determines the Sender's type.
• Signal (or Message): An action taken by the Sender, which the Receiver observes.
• Receiver (Player 2): The player who observes the signal and then acts.
• Beliefs: The Receiver's probabilities about the Sender's type, which are updated after observing the signal.
• Payoffs: Outcomes for both players, which depend on the Sender's type, the signal sent, and the Receiver's action.

The following game tree illustrates this sequence. Nature first chooses Player 1's type. Player 1, knowing their type, sends a signal (A or B). Player 2 observes the signal (but not the type) and chooses a response (C or D).

The solution concept for these games is a Perfect Bayesian Equilibrium (PBE). A PBE consists of:

1. A strategy for the Sender (what signal to send for each type).
2. A strategy for the Receiver (what action to take for each signal).
3. A system of beliefs for the Receiver (a probability distribution over the Sender's types for each possible signal).

These three components must satisfy two conditions:

• Sequential Rationality: Each player's strategy must be a best response to the other's strategy, given their beliefs.
• Belief Consistency: The Receiver's beliefs must be updated from their prior beliefs using Bayes' Rule whenever possible.

For a great introduction to the structure of these games and the PBE concept, the following videos are highly recommended.

There are two primary types of equilibria we look for:

• Separating Equilibrium: Each type of Sender sends a different signal. The signal perfectly reveals the Sender's type.
• Pooling Equilibrium: All types of Senders send the same signal. The signal provides no new information about the Sender's type.

Our goal today is to find a separating equilibrium in our boudoir photoshoot scenario.

2. Modeling the Photoshoot as a Signaling Game

Let's frame the photographer-model interaction using the components of a signaling game.

• Sender: The Model (Player 1)
• Receiver: The Photographer (Player 2)
• Model's Types (Private Intent): Nature endows the model with one of two artistic intents, each with a 50% prior probability.
  • Type A (Artistic): The model's goal is to create sophisticated, high-fashion images that are elegant and sensual.
  • Type P (Provocative): The model's goal is to create bold, edgy images that are overtly sexual and daring.
• Model's Signals (Outfit Choice):
  • Signal E (Elegant): A classic lace and silk lingerie set.
  • Signal B (Bold): A strappy leather or vinyl lingerie set.
• Photographer's Actions (Shooting Style):
  • Action FA (Fine Art): Soft lighting, classic composition, emphasizing form and grace.
  • Action ED (Edgy): Dramatic hard lighting, unconventional angles, emphasizing intensity and power.

Payoffs
The payoffs represent the satisfaction or utility for both the model and the photographer. A successful shoot occurs when the intent, outfit, and style are aligned. A mismatch leads to lower payoffs. Let's define the payoffs (Model, Photographer) in a game matrix.

If the Model chooses Elegant Lingerie (E):

Interpretation:

• (Artistic, Elegant, Fine Art) -> (10, 5): Perfect match. The model gets her ideal shots, and the photographer produces great portfolio work.
• (Provocative, Elegant, Edgy) -> (4, 4): The model is Provocative but chose an Elegant outfit. The Edgy style helps her achieve some of her goal, and the photographer creates an interesting, tense image.

If the Model chooses Bold Lingerie (B):

Interpretation:

• (Provocative, Bold, Edgy) -> (10, 5): Perfect match for the provocative intent.
• (Artistic, Bold, Fine Art) -> (5, 3): The model is Artistic but chose a Bold outfit. The Fine Art style tones it down, resulting in a decent but not ideal outcome.

3. Finding a Separating Equilibrium

We will now systematically check for a separating equilibrium where the model's outfit choice perfectly reveals her intent. The following handout provides the formal steps for this process. Your background in operations research should make this structured approach feel familiar.

Let's apply these steps to our photoshoot game.

Step 1: Propose a Separating Strategy for the Model
Let's hypothesize the most intuitive separating strategy:

• An Artistic model chooses Elegant lingerie (E).
• A Provocative model chooses Bold lingerie (B).

Step 2: Derive the Photographer's Beliefs
If the model follows this strategy, the photographer can update their beliefs using Bayes' Rule.

• If the photographer observes Elegant (E), they know it must have come from the Artistic type. Their belief is: P(Artistic | Elegant) = 1.
• If the photographer observes Bold (B), they know it must have come from the Provocative type. Their belief is: P(Provocative | Bold) = 1.

Step 3: Determine the Photographer's Best Response
Given these beliefs, what is the photographer's optimal action for each signal?

• If they see Elegant (E): They believe the model is Artistic. They compare their payoffs from shooting Fine Art (5) vs. Edgy (2). They will choose Fine Art (FA).
• If they see Bold (B): They believe the model is Provocative. They compare their payoffs from shooting Fine Art (1) vs. Edgy (5). They will choose Edgy (ED).
  So, the photographer's strategy is: (FA if E, ED if B).

Step 4: Check for the Model's Incentive to Deviate
Now, the crucial check. Given the photographer's strategy, would either type of model benefit from sending the "wrong" signal to fool the photographer?

• Consider the Artistic Model:
  • Equilibrium path: She chooses Elegant (E). The photographer responds with Fine Art (FA). Her payoff is 10.
  • Deviation path: If she chooses Bold (B) instead, the photographer will respond with Edgy (ED). Her payoff would be 8.
  • Since 10 > 8, the Artistic model has no incentive to deviate.
• Consider the Provocative Model:
  • Equilibrium path: She chooses Bold (B). The photographer responds with Edgy (ED). Her payoff is 10.
  • Deviation path: If she chooses Elegant (E) instead, the photographer will respond with Fine Art (FA). Her payoff would be 2.
  • Since 10 > 2, the Provocative model has no incentive to deviate.

Conclusion: A Separating Equilibrium Exists!
Since neither player has an incentive to deviate and the beliefs are consistent, we have found a Perfect Bayesian Separating Equilibrium. The full equilibrium is:

• Model's Strategy: If Artistic, choose Elegant. If Provocative, choose Bold.
• Photographer's Strategy: If they see Elegant, shoot Fine Art. If they see Bold, shoot Edgy.
• Photographer's Beliefs: If they see Elegant, P(Artistic)=1. If they see Bold, P(Provocative)=1.

In this equilibrium, the model's outfit choice is a credible signal of her private intent, allowing for a perfectly coordinated and mutually beneficial artistic outcome.

4. An Extension: Poses, Confidence, and Multi-Dimensional Signals

The learning outcome also mentions "poses signal confidence." This adds another layer. So far, we've used a single signal (outfit). But in reality, communication is multi-dimensional. A model signals with her outfit, her pose, her expression, and more.

This is where things get really interesting. Imagine a truly confident model wants to distinguish herself not just from other intents but from less confident models who might try to imitate her. This is a problem of "countersignaling" or using alternative signals for distinction.

A fascinating paper, "Trickle-Round Signals," explores this very dynamic. It argues that high-status individuals sometimes adopt low-status signals to distinguish themselves from middle-status imitators. This is only possible when multiple signaling dimensions exist, allowing for a "mix-and-match" strategy.

Let's apply this "mix-and-match" concept to the photoshoot:

• High-Status Signal: An elegant, expensive lingerie set.
• Low-Status Signal: A very provocative, "risky" pose that could look awkward or cheap if not executed perfectly.
• High-Status Actor: A truly confident, skilled model.
• Middle-Status Actor: A less confident model who tries to appear confident by copying signals.

A less confident model might choose an elegant outfit and stick to safe, classic poses. She avoids the risky pose because the "misidentification cost" is too high—if she fails, she looks amateurish.

A truly confident model, however, can afford to take that risk. By pairing the high-status outfit with the risky, provocative pose, she sends a complex signal that the "middle-status" model wouldn't dare to copy. The signal isn't just "I'm artistic" or "I'm provocative"; it's "I am so confident and skilled that I can make even this risky pose look high-fashion." This is a powerful way to distinguish herself, enabled by the use of multiple signaling dimensions.

Conclusion

In this lesson, we have moved from simple information gathering to the strategic world of signaling games. You have learned how to formally model an interaction with incomplete information and find a separating equilibrium where actions credibly reveal private information.

Key Takeaways:

• Signaling games model strategic communication where a Sender's actions (signals) are interpreted by a Receiver to infer the Sender's private information (type).
• A Perfect Bayesian Equilibrium (PBE) is the solution concept, requiring players to act optimally given their beliefs, and for beliefs to be updated via Bayes' Rule.
• In a separating equilibrium, different types send different signals, allowing the Receiver to perfectly deduce the Sender's type. We found such an equilibrium in our photoshoot model based on outfit choice.
• Real-world signaling is often multi-dimensional. A sophisticated sender can "mix and match" signals (like an outfit and a pose) to distinguish themselves from imitators in ways a single signal cannot achieve.

In our next lesson, we will shift from these one-shot communication games to long-term strategic relationships. We will "Analyze the hotwife-bull dynamic as a repeated game," exploring strategies like tit-for-tat that are designed to build cooperation and satisfaction over time.