Near-Miss Distance Scorecard¶
Assessment artifact — instantiates Counterfactual Proximity Signal Calibration
Scores how close an actual case came to a value-changing alternative across named proximity dimensions, anchored to the factual outcome record.
A Near-Miss Distance Scorecard answers exactly one question about an event: how close did it come? It measures the distance between what actually happened and a value-changing alternative along explicitly named dimensions — seconds, meters, procedural barriers still standing, causal steps remaining — and records each as a score attached to the factual outcome. Its defining discipline is that every closeness claim must cite a named dimension; "that was close" is not admissible until it reads as "close on this axis, by this much." It deliberately does not price the alternative's value or decide the response; it is a measuring instrument for nearness, calibrated against the actual record so the distance can never drift free of the fact.
Example¶
On a construction site a wrench slips from a worker on the eighth-floor deck and falls to the ground below, landing three meters from a laborer. No one is hurt. A bare log records "dropped object, no injury." The Near-Miss Distance Scorecard measures the near-ness instead, along named axes. Spatial distance: 3 m from a person — close. Temporal distance: the laborer had stepped into that zone eight seconds earlier — very close. Barrier distance: the toe-board that should have stopped the wrench was missing, and the tethering policy was not followed — zero barriers held; the only thing between the tool and a person was luck. Each score is written beside the unchanged factual anchor (object fell, no injury, at this time and place). The scorecard's output is a profile — spatially close, temporally very close, zero barriers — that says this "no injury" was far nearer to serious harm than the label suggests, without ever converting it into an injury record.
How it works¶
The scorecard fixes the factual anchor first — the observed outcome, timing, and conditions — so the measurement is pinned to what happened. It then measures proximity on each named dimension: pick the axes on which "close" is meaningful for this domain, define the unit for each (metric, ordinal, or count of barriers/steps), and score the actual case's distance to the value-changing alternative on each. Scores are kept per-dimension rather than mashed into one number, because a case can be spatially far but temporally razor-thin, and collapsing that hides the lesson. The result is a distance profile, not a verdict.
Tuning parameters¶
- Dimension set — which axes of closeness are measured (distance, time, barriers, causal steps). Adding axes captures more near-ness but dilutes attention and effort.
- Scale type per axis — continuous metric, ordinal band, or barrier/step count. Metric scales are precise where a real unit exists; ordinal scales are honest where "close" is inherently judgmental.
- Aggregation stance — keep dimensions separate or roll up to a summary closeness. Rolling up eases triage but can mask a single lethal axis.
- Anchor strictness — how firmly the factual record is quarantined from the distance scores, so nearness never edits the outcome.
When it helps, and when it misleads¶
Its strength is making "how close" a measured, comparable quantity instead of a shudder — two events with the same "no harm" label become distinguishable by their distance profiles. It gives teeth to the old safety intuition that abundant near-misses sit beneath every rare serious event, so the near-misses are worth counting.[1] Its failure mode is dimension blindness: if the axis that actually mattered is not on the scorecard, a genuinely close call scores as distant and gets ignored, and teams then "learn" the wrong lesson from a mis-measured profile. The guarding discipline is to choose dimensions from how harm or value actually changes in the domain — not from what is easiest to measure — and to keep a residual "other" axis open for closeness the fixed dimensions miss.
How it implements the components¶
proximity_metric_or_ordering— it defines the named dimensions of closeness and scores the actual case's distance to the value-changing alternative on each.factual_outcome_anchor— every distance score is pinned to the recorded actual outcome, timing, and conditions, which the scorecard keeps unaltered.
It does NOT implement threshold_band_partition — cutting the measured distance into named bands (far miss, close miss, crossing) is Threshold Band Map's job; this scorecard produces the raw distance, not the bands. Nor does it implement value_delta_frame, the worth-of-the-alternative pricing owned by Counterfactual Value-Delta Table.
Related¶
- Instantiates: Counterfactual Proximity Signal Calibration — supplies the proximity measurement the archetype weights against value delta.
- Sibling mechanisms: Threshold Band Map · Counterfactual Value-Delta Table · Almost-Reward Annotation · Close-Call Review Protocol · Counterfactual Plausibility Filter · Near-Miss Response Tier · Proximity Signal Backtest · Regret-Weighted Decision Log · Salience Overweighting Check
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: The scorecard defines proximity dimensions and computes the actual case's distance from a value-changing alternative while preserving the factual anchor.
Nearest alternative: Representation, Specification & Plan — A scorecard presents the values, but the operative work is multidimensional measurement and comparison.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Industrial safety engineering made near-miss severity and proximity measurable inputs to prevention rather than lucky non-events.
Related originating lineages:
- Law & Governance — Legal counterfactual reasoning contributes the idea of proximity to a value-changing alternative.
- Statistics & Experimental Design — Statistical risk measurement contributed calibrated dimensions, bands, and the disciplined separation of observed outcome from counterfactual closeness.
Review resolution: Both independent reviews agree on primary origin engineering_design; reconciliation resolves reported_ambiguity, alternate_origin_disagreement. Formative alternate lineages retained: statistics_experimental_design, law_governance. The broader reach of later applications is kept separate as domain_reach=multi_domain; origin_mode=cross_disciplinary_synthesis describes the historical relationship among lineages. Confidence is conservatively reconciled to medium, and encyclopedia_synthesis=true preserves the reviewers' boundary judgment.
Attribution caveat: The exact multidimensional scorecard appears to synthesize several safety metrics rather than reproduce a single canonical scale. Near-miss proximity is used in several risk traditions without a single canonical scorecard lineage.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
References¶
[1] Heinrich's safety triangle (H.W. Heinrich, 1931) posited a fixed ratio — roughly 300 no-injury near-misses to 29 minor injuries to 1 major — beneath every serious event. The exact numbers are much disputed, but the surviving insight is sound: near-misses are the abundant, countable early evidence of the rare severe outcome, which is why measuring their distance is worthwhile. withdrawn registry ↩