GFT instruments / Risk and resilience
Put exposure and assumed controls on the same field. This illustrative register separates the risks that reshape a plan from those that need monitoring.
An instrument for the decision
Change the inputs. Read the result. Keep the assumptions visible. Every number below is an illustrative assumption or arithmetic from your inputs. The roles are proposed responsibilities, not GFT staff.
Sum of six residual indices, one per risk
Keep the actions with the scores
Read both columns before choosing a response. A low residual score depends on the entered control assumptions. A high-impact flag remains a separate reason to reconsider the plan. Sort the columns, move the threshold, and watch the matrix beside the table answer.
| Owner and action | Assessed against residual | Edit | ||||
|---|---|---|---|---|---|---|
| 01Power interruption | Operations leadTest backup supply and rehearse a controlled restart. | 59.5 | 25.0 | Reshape | Edit ↑ | |
| 02Demand shortfall | Commercial leadValidate customer commitments and stage capacity against orders. | 44.0 | 29.9 | Reshape | Edit ↑ | |
| 03Input disruption | Supply leadQualify alternate supplies and define material-cover triggers. | 29.3 | 15.2 | Review | Edit ↑ | |
| 04Supervisor capacity | Production leadDemonstrate supervisor competence and cross-train relief cover. | 36.0 | 16.2 | Review | Edit ↑ | |
| 05Water constraint | Engineering leadTest supply quality and buffer essential process demand. | 31.5 | 16.5 | Review | Edit ↑ | |
| 06Data integrity | Data leadReconcile shift records and test access and recovery controls. | 25.0 | 8.1 | Review | Edit ↑ |
AssessedConditional residualThreshold
A decision needs a counterfactual
Save the current valid register as A, adjust the live register to make B, then drag the scrubber from A to B and watch every dot travel. The snapshot includes the review window, owners, actions and control assumptions. It stays in this page session and is cleared on reload. Saved A also returns as dashed rings on the matrix above and on the closing field.
A is preloaded with the illustrative pressure case: likelihood plus 15 points, impact plus 10, both control reductions halved. Save the current register to replace it.
| Risk | A | B | Change | Direction |
|---|---|---|---|---|
| 01Power interruption | 54.9 | 25.0 | -29.9 | |
| 02Demand shortfall | 52.4 | 29.9 | -22.5 | |
| 03Input disruption | 33.4 | 15.2 | -18.2 | |
| 04Supervisor capacity | 36.8 | 16.2 | -20.6 | |
| 05Water constraint | 37.2 | 16.5 | -20.7 | |
| 06Data integrity | 24.1 | 8.1 | -16.0 |
A: 238.8 (6 reshape). B: 111.0 (2 reshape). B is 127.9 index units lower. Same review window. Control credit, A: assumed; B: assumed.
Reproduce the arithmetic
Step 1 of 4
Define one adverse event per row and use the same review window for all likelihood estimates. Likelihood means the chance of at least one occurrence in that window.
Impact is a continuous disruption index that your team must calibrate before comparing risks. It is not a monetary amount, a percentage of revenue or a five-point category. Multiplying ordinal category labels would give a misleading impression of precision.
Assessed index = L × I / 100
Step 2 of 4
Prevention reduces likelihood relative to its assessed value. A relative reduction scales the number; it does not subtract percentage points.
Changing the review window does not automatically rescale likelihood: return to the evidence and re-estimate each event. A constant event rate is not assumed.
Residual L = L × (1 − P / 100)
Step 3 of 4
Protection reduces impact relative to its assessed value. Prevention and protection act on different dimensions, so their effects combine multiplicatively.
Residual I = I × (1 − C / 100)
Step 4 of 4
No mitigation credit ignores both reductions. Displayed scores are rounded to one decimal; decision bands use unrounded values. CSV retains the calculated precision.
Residual index = residual L × residual I / 100
Assessed = residual + reduction
Worked illustrative example: power likelihood is 70%, impact is 85 index units, prevention is 30% relative and protection is 40% relative. The assessed index is 59.5. Conditional likelihood is 49%, impact is 51, and the residual index is 24.99. The reduction is 34.51 index units. These values demonstrate the method; they are not a GFT operating result.
L: likelihood (%). I: disruption index. P: prevention (% relative). C: protection (% relative). For no credit, P and C are treated as zero.All numerical defaults are illustrative workshop assumptions, chosen to exercise different exposure and control combinations. They are not estimated from factory data. Each cell carries a bar for its default and a tick for the live value. Hover or focus a row to find its dot, and restore any risk to its default.
| Risk | Likelihood | Impact | Prevention | Protection | Live register | Residual drift | Restore |
|---|---|---|---|---|---|---|---|
| 01Power interruptionIllustrative | 70 | 85 | 30 | 40 | Default | ||
| 02Demand shortfallIllustrative | 55 | 80 | 20 | 15 | Default | ||
| 03Input disruptionIllustrative | 45 | 65 | 35 | 20 | Default | ||
| 04Supervisor capacityIllustrative | 60 | 60 | 40 | 25 | Default | ||
| 05Water constraintIllustrative | 35 | 90 | 25 | 30 | Default | ||
| 06Data integrityIllustrative | 50 | 50 | 50 | 35 | Default |
Illustrative policy: reshape at a residual index of 20 or more, or a residual impact of 80 or more with positive likelihood; review at 8 or more; otherwise monitor. The two contours on the matrix are these thresholds. These are GFT tool conventions, not externally prescribed acceptance limits.
Illustrative stress case: add 15 percentage points to each likelihood and 10 impact units, each capped at 100. Halve both assumed control reductions. The arrows on the matrix show where each residual dot travels. This is a deliberate challenge, not a forecast.
Illustrative default: 12 months, a workshop planning interval. Allowed range: 1 to 60 whole months. No inference about annual event rates. If the window changes, return to the evidence and re-estimate each likelihood.
Live review window: months
Illustrative role labels and suggested actions. Replace them with responsibilities agreed by your team. No risk defaults are sourced programme measurements, so a GFT programme defaults preset would be inappropriate.
Keep the decision wider than the matrix
04Treat an attractive residual position as a hypothesis to verify. Record the evidence behind the likelihood and the action that would falsify the assumed reduction. Revisit owners and triggers when the production process, order book or supply conditions change.
Provenance stays with the instrument
The NIST CSRC residual-risk glossary describes remaining risk after responses. Its context is security risk. This instrument borrows the distinction between assessed and remaining exposure; the probability-index arithmetic, risk selection, thresholds and reduction assumptions are GFT's illustrative design. NIST does not validate this model. The NIST risk glossary provides terminology context. Both pages were opened on 30 October 2026; the source receipt is available in the lane package.
These are the research programme's paper titles, not a claim that the papers are released.
Find the production loss with the OEE instrument ↗
Read the measurement approach ↗
Only available routes are linked. Export this register before changing the scope of the wider operating model.
Sharing places the full register in the URL, including owners and actions. Use role names and non-sensitive wording. No register data is submitted by this tool. CSV contains your current valid inputs and outputs. This is a planning aid, not investment, legal or tax advice.
Use the result in its boundary
The field shows what would follow if the entered controls worked as assumed. It cannot show whether they will work. Bring outage logs, order evidence, supplier tests and an accountable owner to the next review.
The register slip: what the file will hold
Sum, B now
Reshape
Saved A