Reference
Templates and Checklists
Part of the War Gaming skill · loaded on demand from SKILL.md
Copy-ready structures for each deliverable in the workflow. Placeholders in angle brackets are allowed here and nowhere in a finished client document.
1. Focal decision sheet (Stage 1)
## Focal decision
Decision: <one sentence naming the options and the commit date>
Decision maker: <name, role>
Horizon year: <year>; reason: <why consequences have mostly played out by then>
Uncertainty level (Courtney, Kirkland, Viguerie 1997): <1-4>; reason: <one line>
Status quo option and its cost: <what happens if nothing is chosen>
| Option | Content | Investment | What would have to be true (from strategic-choice) |
|---|---|---|---|
| A | | | |
| B | | | |
| C | | | |
2. Predetermined elements and candidate uncertainties (Stage 2)
## Predetermined elements (appear in every scenario)
| Element | Evidence and source | Why it is not in doubt over the horizon |
|---|---|---|
## Candidate uncertainties
| Candidate (phrased as a question) | Range by <horizon year> | Importance 1-5 | Uncertainty 1-5 | Keep / fold into axis / sensitivity / drop |
|---|---|---|---|---|
3. Critical uncertainty definition (Stage 3)
## Axis <A/B>: <name>
Question: <the uncertainty as a question>
Low end: <label>; threshold: <number>; mechanism: <how the world gets there>; credible believer: <who>
High end: <label>; threshold: <number>; mechanism: <how the world gets there>; credible believer: <who>
Option favoured at low end: <option>; at high end: <option>
Independence from the other axis: <one sentence on differing drivers, and a route to each off-diagonal cell>
4. Scenario matrix and narrative (Stage 4)
## Scenario matrix
| | <Axis B low label> | <Axis B high label> |
|---|---|---|
| <Axis A high label> | <Scenario name> | <Scenario name> |
| <Axis A low label> | <Scenario name> | <Scenario name> |
Base case: <name>, because <reason>. No probabilities assigned to the others.
## Scenario: <Name>
Cell: <Axis A end>, <Axis B end>
Narrative (past tense, from the horizon year, 150-300 words): <what happened, in what order, why it made sense to the actors; the state of each rival by name; the predetermined elements in place>
Who wins in this world: <player and why>
Our problem in this world: <one sentence>
## Parameter table
| Parameter | <Scenario 1> | <Scenario 2> | <Scenario 3> | <Scenario 4> |
|---|---|---|---|---|
| Market growth | | | | |
| Price level vs base year | | | | |
| Channel mix | | | | |
| Margin change | | | | |
| Rival positions (branches, capacity, ownership) | | | | |
| <model-specific inputs> | | | | |
5. Signpost table (Stage 5)
| Signpost | Axis and direction | Source | Threshold | Review cadence and first check | Owner | Action if tripped |
|---|---|---|---|---|---|---|
Signpost acceptance checklist (delete any signpost that fails):
- Leading: moves before the outcome it indicates
- External: readable without the client's own results
- Cheap: a named public or routine source
- Thresholded: a number or a definite event
- Dated: a review date and cadence
- Owned: a named person
- Actionable: names the belief shift and the decision triggered
6. Value Net and response profile (Stage 6)
## Value Net
Customers: <segments; whose payoff changes under which option>
Suppliers: <key groups; whose payoff changes under which option>
Competitors: <named; test: customers value us less when they have them>
Complementors: <named; test: customers value us more when they have them>
Potential entrants: <named>
## Response profile: <Rival>
Objectives: <growth, share, margin, exit, succession, non-financial; what management is measured on>
Assumptions: <about the industry; about us; blind spots those create>
Current strategy: <what they are actually doing, by function>
Capabilities and constraints: <balance sheet, cost position, integration bandwidth, management attention, public commitments>
History of reaction: <how they responded to the last three comparable moves by anyone; time to respond; duration>
Most provoking move we could make: <one line>
Least noticed move we could make: <one line>
7. Competitor-response matrix (Stage 6)
| Rival | Our option | Will they react? (see, hurt, bandwidth, time) | Likely move (second most likely) | Lead time | Damage to us (metric, scenario) | Our counter | PARTS lever that changes their calculus |
|---|---|---|---|---|---|---|---|
Cell-writing rules:
- Answer "will they react" from their metrics, not ours.
- The likely move comes from their history unless a stated reason overrides it.
- Damage is a number in the horizon year, tied to a scenario.
- The counter must be something they cannot follow cheaply, or it is not a counter.
8. War-game design sheet (Stage 7)
## War game: <client>, <dates>
Setting scenario: <name>; injects move the world toward: <name>
Decision owner present: <name, role>
| Team | Members | Payoff metric and constraints |
|---|---|---|
| Home | | |
| <Rival 1> | | |
| <Rival 2> | | |
| Market | | |
| Control | | n/a |
Rounds: <n>, each representing <period>. Per round: <deliberation minutes> / <adjudication minutes> / <plenary minutes>.
Home team option played live: <option>; desk-game variants: <options>.
Legal moves: price (list, contract, promotional), capacity, product and service, channel, M&A, partnership, signalling, hiring.
Adjudication rule: <share response to price and service gaps, lags, caps, capacity limits, acquisition timing, synergy capture>
| Inject | Round and timing | What it tests | Source (signpost or scenario narrative) |
|---|---|---|---|
Briefing pack contents per team: the setting scenario one-pager; the team's own response profile; its payoff metric; the rules and legal moves; the adjudication rule; the move sheet; the public facts on every other player.
9. Round record and surprise log (Stage 7)
## Round <n> (<period>)
Inject: <event or none>
| Team | Moves submitted | Rationale in the team's words |
|---|---|---|
Adjudication: <what Control resolved and how the rule was applied>
Market team allocation: <share and service scores by player>
Scoreboard: <each team's payoff metric at end of round>
## Surprise log
| # | Round | What happened | Why nobody predicted it | Effect on option values |
|---|---|---|---|---|
10. Payoff table and verdict (Stage 8)
## Option-by-scenario payoff (<metric>, <horizon year>)
| Option | <Scenario 1> | <Scenario 2> | <Scenario 3> | <Scenario 4> | Worst | Max regret | Acceptable in (of 4) | Class |
|---|---|---|---|---|---|---|---|---|
Regret arithmetic: <best per scenario; each option's regrets; each maximum>
Downside against tolerance: <constraint; each option's worst case against it>
Classification rule: survivor = acceptable in all four; bet = best in one or two, unacceptable in at least one; dominated = never best and unacceptable in at least one.
## Verdict
Pursue <option> unless <signpost(s)> trip by <date>, in which case <trigger action>.
## Adaptive plan
No-regret moves now: <list>
Options to buy cheaply: <instrument, cost, what it converts>
| Trigger | Date checked | Decision |
|---|---|---|
11. Pre-delivery checklist
Scenarios:
- Two axes, each with numeric thresholds and a stated mechanism at both ends
- Independence argued in one sentence, with a route to each off-diagonal cell
- Each end of each axis favours a different option
- Four narratives in the past tense, each naming every rival's state
- Names of two to four words; no "optimistic", "pessimistic", "base", "downside"
- One base case named with a reason; no probabilities on the others
- Parameter tables consistent with narratives and with the model
Signposts:
- Three to five per axis, all passing the seven-point acceptance checklist
- Handed to
kpi-architecturewith owners and cadence
Competitor response:
- Value Net drawn with complementors named
- One profile per rival, with history of reaction
- Matrix complete for every rival-option pair; no cell reads "unlikely to react" without the three questions answered
- Each counter is something the rival cannot follow cheaply
War game:
- Decision owner in the room
- Strongest players on rival teams
- Adjudication rule published before play
- Injects drawn from signposts and narratives
- Home team lost at least one round, or a desk variant shows an option collapsing
- Surprise log written by Control before the debrief ends
Verdict:
- Payoff table reconciles with model and game record
- Three lenses applied and shown
- Verdict sentence contains "unless" and a date
- Adaptive plan has no-regret moves, bought options and dated triggers
- Internal failure modes handed to a separate pre-mortem, not duplicated here
- Output passed to
executive-communicationbefore any client sees it