Stress-testing

Take a commitment the school has already made, or is about to make, and run it through each scenario in turn. Where does it hold, where does it bend, where does it break, and what would it have to become? It is the least glamorous futuring move and the one most likely to change something.

Futuring: the four moves

Move 4 of 4 · Finds the plan’s single point of failure

Run a commitment through four futures and see where it breaks.

Time: 2 hours
People: 4–10, including whoever owns the commitment
You need: Four scenarios, your own or borrowed, the commitment in one sentence, the two worksheets
You leave with: Load-bearing assumptions, each with a signpost and a hedge, and usually one change to the commitment

Start here

The idea

What the move is, and why it works

What the move is

Every plan rests on assumptions: that funding holds, that staff stay, that a tool stays affordable, that parents agree. Most plans have eight to fifteen. Stress-testing lists them, then runs the plan through each of four futures and asks which assumptions fail where. An assumption that fails in more than one world is load-bearing.

For each load-bearing assumption you write a signpost and a hedge. The signpost is the first thing you would see if it began to fail. The hedge is what you can do now so the plan depends on it less.

The idea behind it

A plan tuned to one future is fragile. Better to choose a plan that does acceptably across many futures than one that does best in the future you expect. This is the core of decision-making under deep uncertainty, and the scenarios are simply a disciplined way of supplying the futures.

It works because it starts from something real. Scanning and scenarios can feel abstract to a busy school. A commitment already in the budget is not abstract, and watching it break in a world the group has just written is persuasive in a way a scan never is.

Why it is the place to start

It needs scenarios, but they can be borrowed. It takes two hours. It touches a decision that is live now. And it nearly always produces a change, often small, to the wording, the order or the hedges of the plan.

The condition is that the commitment can still change. Run it on a plan nobody is allowed to alter and the group learns that foresight is theatre.

When

When to reach for it, and when not to

Use it when

A plan is about to be approved, and nobody has asked what would break it.
A commitment is already running and you suspect it rests on one supplier, one person or one policy.
The group is split on a decision, and you want the disagreement on specific assumptions, not on taste.
You have scenarios and want them to touch a real decision.

Skip it when

The plan cannot change, whatever the test finds. Say so, and do something else with the time.
The commitment is too vague to fail. Sharpen it into one sentence first.
The horizon is under a year. A plain risk register will do.

Before

Before anyone is in the room

Write the commitment in one sentence, with who does what by when.
Confirm with the owner that the commitment can still change, and what kind of change is possible.
Have four scenarios ready, with a one-paragraph summary of each for people who have not read them.
Invite at least one sceptic of the plan, and one person who will have to carry it out.
Print the four-world test and the assumptions grid, one set per group.
Book a follow-up with the owner within two weeks, to decide on any changes.

Steps

How to run it

7 steps, about 2 hours in all. The questions in quotation marks are meant to be read out as they are.

01
State the commitment in one sentence
10 min
Read the commitment aloud and write it at the top of the four-world test. Check that it says who does what, for whom and by when. If the group cannot agree on the sentence, that is the first finding.
Ask the room
- What exactly have we committed to?
- What would count as it working, in three years?
Watch for A commitment so vague it cannot fail, such as “embrace digital learning”. Sharpen it before going on.
02
List the assumptions
20 min
Ask what must stay true for the commitment to work. Funding, staffing, tools, regulation, parental consent, student behaviour, other schools’ behaviour. Aim for eight to fifteen, each one sentence, on the assumptions grid.
Ask the room
- What must stay true for this to work?
- What are we assuming about other people: parents, other schools, suppliers?
- What are we assuming about students that we have not checked?
Watch for Only listing resources. The deepest assumptions are usually about trust and behaviour.
03
Run it through each world
40 min
Take the worlds one at a time, ten minutes each. Read the world’s summary aloud. Ask what that world does to the commitment, whether it holds, bends or breaks, and which assumption fails. Write each answer on the four-world test.
Ask the room
- In this world, what happens to our commitment in its second year?
- Does it hold, bend or break?
- Which assumption failed?
- What would it have to become to work here?
Watch for A commitment that “holds” in a world because it still runs. Ask whether it still does what it was for. Some plans hold technically and fail in purpose.
04
Mark the load-bearing assumptions
15 min
Look across the four answers. Any assumption that fails in two or more worlds is load-bearing. Mark it. There are usually two to four.
Ask the room
- Which assumptions failed in more than one world?
- Which one, if it failed, takes the whole commitment with it?
Watch for The group defending an assumption because it is inconvenient to lose. Ask what evidence would change their mind.
05
Write a signpost and a hedge for each
20 min
For each load-bearing assumption, write a signpost: the first concrete thing you would see if it started to fail. Then a hedge: what you can do now so the commitment leans on it less. Give each signpost a person who will watch for it.
Ask the room
- What would we see first if this began to fail?
- Who would notice, and who would they tell?
- What can we do now so we depend on this less?
Watch for Signposts nobody will look at. A signpost needs a person and a date.
06
Run a premortem
10 min
Tell the group it is two years from now and the commitment has failed. Each person writes, silently, the most likely reason. Read them out. Add any new assumption to the grid.
Ask the room
- It is two years from now and this has failed. What happened?
- What did we see coming and ignore?
Watch for Everyone writing the same safe answer. Ask for the reason they would only say privately.
07
Decide what changes
5 min
Rewrite the commitment sentence with the hedges in it. Agree which changes go to the owner and when. If nothing changes, write down why.
Ask the room
- What is the commitment now?
- Who takes this to the person who can change it, and by when?
Watch for A good session and no change. If the test finds a load-bearing assumption and the plan stays the same, record the decision and who made it.

Worksheets

Two sheets to print and fill in

Each one prints on its own sheet of A4. The grey line under each is an example from the worked situation.

Futuring · Stress-testing · Worksheet

Four-world test

Runs one commitment through four futures and records where it holds, bends or breaks.

Commitment

Owner

Date

Review on

	What this world does to the commitment	Holds, bends or breaks	Which assumption fails	What it would have to become
World 1
World 2
World 3
World 4

Example Division Economy | Nobody trusts work done out of sight, so the next school ignores the portfolio | Breaks | The work in it is the student’s own | Include pieces witnessed in person or signed off by a mentor

geraldajam.com/work/toolkit

Futuring · Stress-testing · Worksheet

Assumptions, signposts and hedges

Lists what must stay true, marks the load-bearing assumptions and gives each a signpost to watch and a hedge to act on now.

Commitment

Owner

Date

AssumptionWhat must stay true	Worlds where it fails	Load-bearing?Fails in two or more	SignpostWhat you would see first, and who watches	HedgeWhat you do now

Example The student, not the platform, controls what travels | Velvet Cage, Division Economy | Yes | A receiving school asks for the analytics, not the pieces (Year Head) | Student-owned portfolio, access that expires after each transfer

geraldajam.com/work/toolkit

Worked example

Stress-testing a portable portfolio

A horizon scan for Singapore’s schools

The group picked a commitment a Singapore secondary school could plausibly make: every student keeps a digital portfolio that follows them to the next school. Assessment had been the sharpest discriminator between the four worlds, so a commitment about assessment was the obvious one to test. We ran it through all four.

Flourishing Techno-human City: holds, with a wobble. This is the world the commitment was written for. Portfolios are learner-owned and credentials are portable. Even here the scenario shows the problem of portfolio inflation, students curating for show, and its answer: slow portfolios, where each piece carries a note from someone the work affected.
Velvet Cage: holds technically, breaks in purpose. The portfolio runs perfectly. It also becomes part of continuous tracking, and assessment becomes the prediction of conformity. Students curate to please the system. On every indicator it looks like success, which is what makes this world dangerous. The group argued for twenty minutes over whether to call this “holds” or “breaks”. That argument was the most useful part of the session.
Division Economy: breaks. Credential credibility collapses under AI-generated work and distrust of data. The next school does not believe the portfolio, so it does nothing. Oddly, the hedge came from inside the scenario: a policy officer in that world proposes portable portfolios verified by local mentors.
Gilded Enclaves: bends into a class marker. Well-resourced schools produce rich portfolios with advanced micro-credentials. Others produce thin ones or none. The portfolio still travels, but it carries privilege with it and widens the gap it was meant to narrow.
Three load-bearing assumptions. That the receiving school trusts the work is the student’s own, which fails in two worlds. That the student, not the platform, controls what travels, which fails in two. That every school has the time and tools to help students build a good one, which fails in one and wobbles in another, so we kept it.
The awkward part. Two of the hedges pull against each other. Witnessed pieces build trust but add teacher time, which is exactly what is scarce in the enclave world. We did not resolve it. We wrote it down as the thing a school would have to decide.

What came out of it

The commitment was rewritten: every student keeps a small portfolio they own, with at least two pieces witnessed in person, and chooses what goes to the next school. Signposts: a receiving school asking for analytics rather than pieces, a supplier offering an authenticity score, and portfolio quality tracking family income in a sample. The word “follows” became “goes with them when they choose”.

Signs

How to tell it is working, and when it has gone hollow

Working

The commitment sentence changes, even by a few words.
An assumption nobody had said aloud ends up marked load-bearing.
Each signpost has a person and a date against it.
The group argues about whether a world breaks the plan, not about whether the world is likely.

Gone hollow

The commitment holds in all four worlds. Either it is very well designed or the worlds are too alike.
The plan cannot change, and everyone knew that before they started.
The hedges are all “monitor the situation”.
The worksheets are filed and the review date passes unread.

Variations

Other ways to run it

In 45 minutes

Skip the full assumptions list. Read each of four borrowed scenarios aloud and ask one question per world: what breaks first? Then run the ten-minute premortem. You will usually find the one load-bearing assumption that matters most.

On an assessment change

Test a change in assessment rather than a tool: moving coursework to in-class writing, say, or adding a spoken defence to project work. In the Singapore worlds, assessment was where the futures differed most, so it is where a stress test earns its time.

With students

Give a class the commitment and the four worlds and ask them where it breaks. Students are usually quicker to spot the assumptions about their own behaviour, such as whether they will curate honestly, that staff take for granted.

What to do with what you made

Take the rewritten commitment to its owner within two weeks, with the load-bearing assumptions and hedges attached. Put each signpost on a named person’s calendar with a check date. If the plan does not change, record who decided that and why, so the decision can be revisited.

Compare the futures Read the four Singapore worlds on assessment, side by side.Probing with options open Tests the plan against the present by acting, not against futures.Scanning Signposts need watching. A light, regular scan is how.

Tools

Tools, templates and guides worth having

Published elsewhere, and worth the time.

Method
Performing a project premortem (Klein, Harvard Business Review, 2007)
Two pages, and the cheapest stress test there is.
Toolkit
The Futures Toolkit (Government Office for Science, 2017)
Includes a policy stress-testing tool that runs a set of objectives through scenarios.
Routine
TRIZ (Liberating Structures)
Asks a group to design the worst possible result, then spot what they already do that resembles it. From Lipmanowicz and McCandless (2013).
Book
Decisive (Heath & Heath, 2013)
Its chapter on tripwires is a practical guide to setting signposts that someone actually checks.

Reading

Where it comes from

Start with the first on the list.

Dewar, J. A. (2002). Assumption-Based Planning: A Tool for Reducing Avoidable Surprises. Cambridge University Press.
The method for finding load-bearing assumptions, with signposts and hedges.
Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18–19.
The premortem in two pages.
Lempert, R. J., Popper, S. W., & Bankes, S. C. (2003). Shaping the Next One Hundred Years: New Methods for Quantitative, Long-Term Policy Analysis. RAND.
Choose the plan that does acceptably in many futures, not best in one.
Walker, W. E., Haasnoot, M., & Kwakkel, J. H. (2013). Adapt or perish: a review of planning approaches for adaptation under deep uncertainty. Sustainability, 5(3), 955–979.
A map of the methods, for whoever wants to go further.

Singapore