05 / Products
Wargame/Agentic Strategy
Model the actors. Simulate the outcomes. Improve the strategy.
Wargame builds an explicit model of a strategic environment, simulates possible interactions and helps identify actions that may improve the likelihood of the outcome you want.
Simulation
Actors, incentives and responses are read as a distribution of outcomes.
Strategic environments are interactive
Wargame is an agentic strategy and scenario simulation system. It creates a model of the relevant actors in a strategic environment — competitors, customers, regulators, partners, political actors, market participants, internal stakeholders — and the incentives, goals, constraints and relationships that shape what they may do.
Strategy is a system of interacting actors. Wargame attempts to model those actors and explore how their decisions may interact. It can then run large numbers of possible scenarios, using Monte Carlo simulation where appropriate, to estimate a distribution of outcomes rather than a single forecast.
The aim is not to predict the future with certainty. It is to make assumptions explicit, explore their consequences at scale, and help you see risks, opportunities, influential variables and actions that may shift the distribution.
Traditional analysis often treats variables independently. Real strategic situations contain actors responding to one another. A competitor reacts. A regulator intervenes. A customer changes behaviour. A partner changes incentives. Wargame is designed around those interactions, so you can inspect them before you have to act.
Build the actor model
Make the assumptions explicit before you explore them.
The system constructs an explicit actor model: objectives, incentives, capabilities, constraints, relationships, likely behaviours, decision criteria and possible actions. Human experts should review and refine those assumptions. That review is part of the method, not an optional courtesy.
Figure · Actor model
01
You
Decision
02
Competitor
Response
03
Customer
Demand
04
Regulator
Constraint
05
Partner
Incentive
Each actor carries objectives, capabilities, constraints and likely behaviours. Experts can review and refine those assumptions before any simulation runs.
Scenario generation
Explore what happens next — before you have to act.
Wargame explores possible actions and responses across the actor model. Models of incentives and constraints produce possible moves; those moves produce responses; the interactions become scenarios; the scenarios are simulated.
Figure · From environment to distribution
- 01
Environment
- 02
Actors
- 03
Incentives + constraints
- 04
Possible actions
- 05
Responses
- 06
Scenarios
- 07
Simulation
- 08
Outcome distribution
Monte Carlo simulation
Many runs. A distribution. Not a single answer.
Wargame can run many simulations with variation in uncertain assumptions. Rather than producing one forecast, it produces a distribution of outcomes. That may allow a reading such as: scenario A occurred in 52% of simulations, scenario B in 31%, scenario C in 17%. Those figures are illustrative. They are not a result from a customer engagement.
The quality of the result depends on the assumptions, the actor model, the data, the uncertainty ranges and the simulation design.
Figure · Outcome distribution
Illustrative distribution — not a real simulation
Scenario A
52%
Scenario B
31%
Scenario C
17%
Rather than a single answer, repeated simulation produces a distribution. The quality of that distribution depends on the assumptions, the actor model, the data and the design of the run.
Strategy optimisation
After you have a distribution, ask what would shift it.
If the desired outcome is X, which actions appear to increase its probability? Which reduce a particular risk? Which actors matter most? Which assumptions drive the model? Where are the leverage points you can actually influence?
The output should include recommended strategic actions, likely consequences, uncertainty, alternative scenarios and the assumptions they rest on. None of that is a guarantee. It is a structured way to reason about a complex environment under uncertainty.
Example scenarios
Corporate questions that involve interacting actors.
01
Competitive response
“What happens if we reduce our price by 10%?”
- Competitor response
- Customer migration
- Margin impact
- Market share
- Possible follow-on pricing behaviour
02
Market entry
“What is likely to happen if we enter this market?”
- Incumbent reaction
- Customer response
- Regulatory conditions
- Partner incentives
- Alternative strategies
03
Negotiation
“What strategy is most likely to produce an acceptable agreement?”
- Stakeholder incentives
- Fallback positions
- Concessions
- Reactions
- Negotiation paths
04
Strategic policy
“What actions increase the probability that this programme succeeds?”
- Multiple organisations
- Stakeholder responses
- Incentives and constraints
- Coordinating actions
- Points of failure
Capabilities
- 01
Actor modelling
Make objectives, incentives, capabilities, constraints and relationships explicit enough to inspect.
- 02
Assumption review
Let human experts examine and refine the model before it is used to explore decisions.
- 03
Scenario generation
Explore possible actions and responses across the actor model, not a single planned path.
- 04
Monte Carlo simulation
Run many simulations with variation in uncertain assumptions, and read a distribution of outcomes rather than a single point forecast.
- 05
Outcome analysis
Estimate which classes of outcome appear more or less often under a given strategy.
- 06
Intervention search
Ask which actions appear to increase the chance of a desired outcome, or reduce a particular risk.
- 07
Sensitivity
Show which assumptions and actors drive the result, and where better evidence would matter most.
- 08
Decision support
Return recommended actions with likely consequences, alternatives, uncertainty and the assumptions they rest on.
Use cases
Environments Wargame is built to model.
01
Competitive response
The problem
A price, product or channel move is rarely answered by the market as a static equation. Competitors react. Customers migrate. Margins can move more than once.
How it is addressed
Model the relevant competitors and customers, then simulate responses to a proposed action — for example a price reduction — including follow-on pricing behaviour and share effects. The output is a distribution of outcomes, not a promise.
02
Market entry
The problem
Entering a market is a decision under interaction: incumbents, customers, regulators and partners all have a move.
How it is addressed
Build the actor model for that market, explore alternative entry strategies, and inspect which assumptions most affect the result.
03
Negotiation
The problem
Negotiations fail when fallback positions, concessions and reactions stay implicit until they are tested in a live discussion.
How it is addressed
Represent the stakeholders, their incentives and their likely responses, then explore negotiation paths and the conditions under which an acceptable agreement appears more often.
04
Strategic policy
The problem
Programmes that depend on several organisations rarely fail for one reason. They fail because the actors do not move together.
How it is addressed
Model the organisations and stakeholder responses around a programme, then ask which actions appear to increase the probability that it succeeds — and which assumptions that reading depends on.
Sensitivity and uncertainty
Which assumptions matter most?
The aim is not simply to output a probability. It is to help you understand what drives the result, where uncertainty lies, which variables can be influenced, and what evidence would improve the model. Sensitivity, where it can be exposed, is how the system earns trust.
Human strategy + machine simulation
Judgement defines the problem. Simulation explores the consequences.
Wargame is not intended to replace human strategic judgement. Your judgement defines the problem. Wargame makes the assumptions explicit and explores their consequences at scale. The name refers to strategic simulation and competitive modelling — not to a military aesthetic, and not to a deterministic oracle.
Limits worth being clear about
A model under uncertainty, not a prediction of the future.
- 01
Simulations depend on model assumptions. A refined actor model is more useful than a precise-looking number.
- 02
Probabilities are conditional estimates, not forecasts of what will happen.
- 03
Actor models need review. Human experts should be able to inspect and correct the incentives and constraints.
- 04
Uncertainty should be visible. Sensitivity is part of the result, not a footnote.
Enterprise deployment
These products are delivered as engagements rather than self-serve licences. We start with your problem, your data, and the systems you already run.