Competitive Response Modeler

Competitive Response Modeler

> The move only matters if you correctly predict the reaction.

What this skill is

A workflow for simulating competitive interactions across multiple rounds: identify likely respondents, characterize each competitor's profile and capacity to respond, model the reaction function, compute the post-reaction payoff, and stress-test commitment moves (capacity, contracts, public statements) that change the game. Built on the Industrial Organization (IO) economics of strategic moves - entry deterrence, predation, accommodation, and signaling.

What it solves

• One-period payoff thinking ignoring the rival's response
• "We'll just match" assumptions that miss asymmetric capacity to respond
• Underestimating the value (or cost) of public commitment
• Mistaking accommodation for cooperation in repeated interactions
• Pricing or capacity moves that triggered a price war the analysis didn't foresee

When to invoke

• Pricing changes (cut, raise, repackage)
• New market entry - anticipating incumbent response
• Product launch in a category with strong rivals
• Capacity additions (manufacturing, sales force, channel)
• Major Mergers and Acquisitions (M&A) move that re-orders the competitive set
• Public strategic announcements where signaling effect matters

Phase 1: Identify the competitive set

List every competitor whose response could materially affect the move's payoff. For each:

• Market share
• Cost position (low-cost, mid, premium)
• Strategic focus (where this product or segment sits in their priorities)
• Recent moves (last 12 months)
• Public posture (aggressive, accommodating, niche)

Cut the list to 3-5 material respondents. Modeling all isn't useful.

Phase 2: Profile each competitor's response capacity

For each:

• Means - can they technically and financially respond? (capacity, capability, capital)
• Motive - do they care about this move? Is it a core market or peripheral?
• Opportunity - internal alignment, leadership focus, organizational bandwidth
• Speed - how fast can they move? days, weeks, quarters?
• History - what did they do the last 3 times something like this happened?

A competitor with means but no motive may not respond. A competitor with motive but no means is loud but harmless. The dangerous response comes from competitors with all three.

Phase 3: Build reaction functions

For each competitor, hypothesize the reaction function: action → reaction.

Examples:

• If we cut price 10% in segment X → competitor Y likely cuts 8% within 30 days because their Chief Executive Officer (CEO) has publicly committed to share leadership
• If we add capacity Z → competitor W likely waits and adds capacity only after observing demand absorption
• If we announce feature F → competitor V probably announces vaporware within 60 days

For each reaction:

• Probability (low / medium / high)
• Magnitude
• Timing
• Reasoning chain

Phase 4: Multi-round payoff matrix

Build a tree of plausible action-reaction sequences:

Round 0: We move
Round 1: Each competitor reacts (or not)
Round 2: We respond to their response
Round 3: New equilibrium or further escalation

At each terminal node, compute payoff (share × margin × volume, or Net Present Value (NPV)).

The headline question is not "is round 0 profitable?" but "is the terminal equilibrium profitable?" A 10% price cut that captures share but settles into a permanently lower-margin equilibrium can destroy value even if round 0 looks like a win.

Phase 5: Commitment moves

Standard game theory: making yourself less flexible can improve outcomes if it changes the rival's calculation.

Commitment types:

• Capacity commitment - invest in capacity that only pays back at high volume, signaling commitment to compete hard
• Long-term contracts - lock customers, raise rival's cost of poaching
• Public statements - CEO commitment publicly disclosed, harder to walk back
• Most-Favored-Nation (MFN) clauses - make discounting expensive (price war deterrent or cartel facilitator)
• Burning bridges - exit an alternative market to focus, signaling commitment

Commitment cuts both ways: it strengthens you when credible but boxes you in when wrong. Score each commitment for credibility (1-5) and reversibility cost.

Phase 6: Signaling

Moves carry information about your type, intentions, and capacity. Consider what each move signals:

• Tough type - willingness to absorb short-term loss to discipline rivals
• Accommodating type - willingness to share the market
• Limited capacity - easier to enter against
• Unlimited capacity - entry deterrence

Test the signal:

• What inference would a rational competitor draw from this move?
• Could you signal the same thing more cheaply with a different move?
• Is there a counter-signal you should send to override an unwanted inference?

Phase 7: Game-theoretic check - repeated versus one-shot

Most real competitive interactions are repeated. Apply the folk-theorem intuition:

| Interaction | Likely equilibrium | |-------------|--------------------| | One-shot, no future | Prisoner's-dilemma defection - undercut, take share | | Repeated, infinite horizon, observable actions | Cooperation sustainable via grim trigger or tit-for-tat | | Repeated, finite horizon | Unravels to defection (unless reputation effects with asymmetric information) | | Many small rivals | Coordination collapses; act like one-shot | | Few large rivals | Implicit coordination is possible (legal and dangerous areas - flag) |

The recommendation must be consistent with the time horizon of the interaction.

Phase 8: Recommendation

Synthesize:

• The move
• Expected reactions per competitor with probability and magnitude
• Terminal equilibrium with payoff versus status quo
• Commitment moves to deploy (or avoid)
• Signals being sent (intended and inadvertent)
• Triggers that would cause us to retreat or escalate
• One unilateral move that improves our position even if no one reacts (insurance)

Output

• Competitor profile sheet (means, motive, opportunity, speed, history)
• Reaction function per competitor with probability and reasoning
• Multi-round action-reaction tree with payoffs
• Commitment-moves shortlist with credibility scoring
• Signaling matrix (signal sent → likely competitor inference)
• Equilibrium projection with payoff versus status quo
• Trigger conditions for retreat or escalation
• One robust move that improves position regardless of reaction

Operating rules

Always

• Identify 3-5 material respondents, not the whole market
• Score each on means, motive, opportunity, and speed
• Compute payoff at terminal equilibrium, not just round 0
• Match the equilibrium concept to the time horizon
• Test commitment moves for credibility

Never

• Assume rivals will accommodate
• Use one-shot logic in a clearly repeated interaction
• Make commitments you can't sustain
• Move first without a contingency for retaliation
• Skip the signal check on any public action

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: varunk130
• Source: varunk130/claude-code-skills
• License: MIT

Install and usage instructions live in the source repository linked above.