How it works
Strategy is a system of people. So model the people.
01
Environment
The decision, the objective, the time horizon and the outcomes that matter.
02
Actors
Everyone who can move the outcome, found and researched with sources.
03
Incentives and constraints
What each actor wants, what it can do and what holds it back.
04
Possible actions
The moves open to each actor, including you.
05
Responses
How each actor reacts to the others, compiled into a policy or reasoned move by move.
06
Scenarios
Uncertain inputs become ranges. Events sit on the real calendar.
07
Simulation
Thousands of seeded runs, each one reproducible.
08
Outcome distribution
How often each outcome occurs, what drives it and which interventions shift it.
Three ways to simulate
Breadth when you need the odds. Depth when you need the story.
Monte Carlo
Each actor’s policy is compiled from its goals, constraints and relationships, so the scenario can run thousands of times, up to 200,000 runs. Use it for the shape of the distribution.
Deep
Each agent reasons through every move, in a smaller number of runs. Use it to see how a future unfolds, step by step, and why.
Hybrid
The actors you pick, such as the acquirer or the swing vote, reason fully. The rest use their compiled policy. Spend the thinking where it matters.
Check its working
Nothing runs that a person hasn’t been able to read.
Every fact in the model is marked as evidenced, inferred or user-defined, with its source. Uncertain inputs are ranges, not guesses dressed as facts.
Your experts review the actors and assumptions before a run. Wargame can challenge its own model and point to what looks weak.
Every run records its seed, its settings and the model version it ran on, so anyone can reproduce it. Open any single future and follow how it got there.
Roles keep it orderly: modellers build, reviewers correct, viewers read.
Limits
A distribution, not an oracle.
- It is not an oracle. It tells you what usually happens under the model, not what will happen.
- Every probability is conditional on the model and its assumptions. We show them all.
- A carefully reviewed model is worth more than a precise-looking number.
- Sensitivity is part of the answer, not a footnote. If the result hangs on one guess, you will know.
Have a decision with several actors and an uncertain outcome?
See Wargame build a model from one sentence, play it out thousands of times, and find the moves that change the odds.