Choice Set Pruning Method¶
Method — instantiates Decision Load Management
A method for filtering or tiering options before detailed comparison.
A Choice Set Pruning Method attacks decision load at its most expensive point: the comparison itself. When a chooser faces thirty options, the cost is not making the final call — it is the combinatorial work of holding thirty candidates in mind and weighing each against the rest. This method interposes a cheap, criteria-based filtering pass before any detailed comparison, using coarse cutoffs to eliminate dominated or disqualified options and, where useful, to tier the survivors into "compare closely" and "keep in reserve." The defining idea is that pruning is a pre-comparison operation on the size of the option set — it shrinks how much deliberation the real decision will cost, without preselecting a winner and without scheduling, routing, or delegating anything.
Example¶
A hiring team receives 240 applications for a senior analyst role. Reading all 240 closely would consume a week of the panel's attention and, past the first few dozen, they would be comparing candidates while depleted — the later resumes getting shallower reads than the early ones. Instead they apply a pruning method. Pass one is a hard-requirement filter: candidates lacking the non-negotiable credential are eliminated (240 → ~90). Pass two applies two coarse cutoffs — a minimum years-in-role and evidence of a specific technical skill — knocking the field to ~30. Pass three tiers those 30 by a quick three-signal rubric into a "deep-review" dozen and an "if-needed" reserve.
Only now does the expensive work begin, and it begins on twelve candidates the panel can genuinely hold in mind at once. The illustrative counts matter less than the structure: three inexpensive passes, each transparent about its criteria, convert an impossible comparison into a tractable one. This is elimination by aspects[n1] used deliberately as a load tool rather than as an unconscious shortcut.
How it works¶
Pruning proceeds in ordered passes, cheapest and most decisive first. Each pass declares an explicit criterion and a cutoff, and the criterion should be conjunctive at the top — a disqualifier that removes options for failing a floor, not a weighted score that requires the very comparison we are trying to defer. Early passes are non-compensatory: a fatal flaw eliminates an option regardless of its other strengths. Only once the set is small does the method hand off to full, compensatory comparison. The tiering variant adds a final sort that ranks survivors into bands so the deep-comparison budget is spent on the contenders first. Throughout, every cut is recorded with its reason, so a pruned option can be reinstated if a criterion is later judged wrong — the transparency that keeps pruning from becoming quiet gatekeeping.
Tuning parameters¶
- Cutoff strictness — how aggressive each filter is. Tight cutoffs shrink the set fast but risk discarding a strong option that narrowly missed one floor; loose cutoffs are safer but leave more to compare.
- Pass ordering — which criterion runs first. Leading with the most decisive, cheapest filter removes the most options for the least effort; a poorly ordered sequence does expensive work on options that a later pass would have cut anyway.
- Compensatory vs. non-compensatory — whether early passes allow strengths to offset weaknesses. Non-compensatory floors are fast and defensible; compensatory scoring is fairer but reintroduces comparison cost.
- Tier count — whether survivors are simply "kept" or sorted into ranked bands. Tiering focuses the comparison budget but adds a sorting step.
- Reinstatement policy — how easily a pruned option can be recalled, trading protection against false cuts for the risk of re-bloating the set.
When it helps, and when it misleads¶
Pruning shines when the option set is large, the options are comparable on a few decisive attributes, and the true cost is comparison volume rather than the final judgment. It directly protects the chooser's limited deliberative energy by ensuring the deep work happens on a set small enough to be held in mind — and while attention is still fresh.
Its failure mode is premature or biased elimination. A single crude cutoff can discard the one unconventional option that would have won on the merits, and because early passes are non-compensatory, a candidate strong everywhere but weak on one floor vanishes silently. Worse, an opaque prune can hide a designer's preference behind a "criterion." The archetype warns that choice reduction fails when it removes options for the designer's convenience rather than the chooser's benefit. The guarding discipline is to keep every cutoff explicit and criterion-based, log the reason for each elimination, and periodically audit the discarded pile — an informal check that the floors are removing noise, not signal.
How it implements the components¶
choice_reduction— the method's whole output: a smaller, tiered option set produced by transparent, criteria-based filtering rather than arbitrary trimming.decision_energy_budget— by shrinking the set before detailed comparison, it caps the deliberative energy the decision will consume, spending the deep-comparison budget only on contenders.
It does not preselect a single winner with an override path — that default_or_delegation_rule design is Default Option Design — and it does not classify decisions by stakes to route them; that decision_stakes_classification routing is Decision Triage Queue.
Related¶
- Instantiates: Decision Load Management — pruning realizes the "reduce and tier the option set" load treatment.
- Sibling mechanisms: Default Option Design · Decision Triage Queue · Approval Threshold Matrix
Editorial Notes¶
Form Classification¶
Form family: Decision, Gate & Allocation
Rationale: A method for filtering or tiering options before detailed comparison, making its operative form a case-specific gate, selection, routing, prioritization, or disposition decision.
Independent corroboration: The frozen evidence defines Choice Set Pruning Method as 'A method for filtering or tiering options before detailed comparison', so its operative form is Decision, Gate & Allocation.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Operations Research
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Decision analysis established screening dominated or disqualified alternatives before costly detailed comparison.
Related originating lineages:
- Cognitive Science — Decision-load and working-memory research explains why reducing the candidate set lowers comparison burden.
Review resolution: Decision analysis is the agreed primary lineage: dominance and noncompensatory screening reduce a candidate set before expensive evaluation. Cognitive-science work on decision load materially explains the design, while behavioral economics is an application-adjacent framing rather than another origin.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Elimination by aspects, Amos Tversky's model of choice in which options are removed for failing successive attribute thresholds rather than scored all at once. As an unconscious heuristic it can misfire; used explicitly and transparently as a pruning method, its non-compensatory passes are exactly what make a large set tractable. ↩