ECSS System Engineering — Trade-off Analysis (space-systems/ecss/e10-tradeoffs)
Use when the task is choosing between candidate solutions to a system
engineering decision under ECSS-E-ST-10C clause 5.3.3 -- running a
documented trade-off analysis against a fixed set of weighted
criteria and recording the outcome as an Annex L trade-off report.
Domain quick reference
- A trade-off analysis compares two or more candidate solutions to
the same decision (a design option, a technology, a supplier, an
architecture) against a fixed set of evaluation criteria agreed
before scoring starts. Each criterion carries a weight expressing
its relative importance; weights are normalized to a common total
so studies with different numbers of criteria stay comparable.
- Some criteria are mandatory (pass/fail) rather than weighted: a
safety, regulatory, or interface-compatibility criterion below its
pass threshold disqualifies the candidate outright, no matter how
well it scores elsewhere. Weighted scoring only ranks candidates
that clear every mandatory criterion.
- Every candidate is scored against every criterion on the same fixed
scale so scores are comparable across candidates; a candidate
missing a score for a criterion, or scored on a criterion outside
the agreed set, is a broken trade-off study and is rejected before
ranking.
- The trade-off report (Annex L) does not just declare a winner: it
records the ranking, the rationale (criteria, weights, scores), and
whether the outcome is a close call -- a thin margin or a tie at
the top means the recommendation is sensitive to the weighting
assumptions and must be reported as such, not smoothed into a false
single winner.
Workflow
- Define the criteria set and each criterion's weight before scoring
any candidate; mark any pass/fail criterion as mandatory with its
pass threshold. Reject a criteria set that is empty, carries a
non-positive weight, or marks a criterion mandatory without a pass
threshold -- the study cannot proceed.
- Score every candidate against every criterion on the fixed scale.
Reject a candidate whose scores do not cover exactly the agreed
criteria set (missing or extra criterion) or that carries a score
outside the scale -- fix the input before ranking.
- For each candidate, check every mandatory criterion's pass
threshold. A candidate below any mandatory threshold is
disqualified and is not weighted-scored.
- Weighted-score every qualified candidate: normalize the criteria
weights to sum to one, then sum weight x score per criterion.
- Rank the qualified candidates by descending weighted score, using
competition ranking so tied scores share a rank and the next
distinct score resumes at its true position (not the next integer).
- Check whether the top of the ranking is a close call: the top two
qualified candidates are tied, or the leader's score is within a
small margin of the runner-up's. Report a single winner only when
the ranking is not a close call and exactly one candidate holds
rank 1; otherwise report the ranking with the close-call flag set
and withhold a single recommendation.
Pitfalls
- Scoring candidates before the criteria and weights are agreed --
criteria chosen or reweighted after seeing the scores bias the
outcome toward whichever candidate was already favored.
- Letting a high weighted score on other criteria compensate for a
failed mandatory criterion -- mandatory criteria are a gate, not
another weighted input; a disqualified candidate stays disqualified
regardless of its other scores.
- Declaring a single winner when the leader and runner-up are tied or
within a thin margin -- a razor-thin or tied result is itself a
finding (the decision is sensitive to the input assumptions) and
belongs in the trade-off report, not hidden behind a confident
recommendation.
- Comparing candidates whose scores were assessed against different
criteria sets or scales -- a candidate that skipped a criterion, or
was scored on a scale the others weren't, produces a ranking that
looks complete but is not actually comparable.
Behavior contract (gate 3)
The criteria/weight validation, candidate-score validation, mandatory
pass/fail gating, weighted ranking, and close-call sensitivity logic
is exercised by the gate 3 contract test:
scripts/test_e10_tradeoffs.py against scripts/e10_tradeoffs_logic.py
(stdlib unittest, offline). Run:
python3 scripts/test_e10_tradeoffs.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and
paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
1---2name: e10-tradeoffs3description: "Use when comparing candidate solutions to a system engineering decision under ECSS-E-ST-10C clause 5.3.3 and Annex L: define the evaluation criteria and their weights, score each candidate against every criterion on a common scale, gate any candidate that fails a mandatory pass/fail criterion, rank the remaining candidates by weighted score, and flag when the recommendation is a close call sensitive to the weighting assumptions rather than a clear-cut winner -- then record the result as a trade-off report. Trigger: ecss, e-st-10-system-scope, trade-off analysis, tradeoff, evaluation criteria, weighted scoring, candidate solution, annex l, trade-off report."4license: Apache-2.05compliance: STANDARDS-REF6standards:7 - id: ecss8 reference-only: true9gated: false10domain: space-systems11pack: space-systems12compatibility: "agentskills.io SKILL.md; any SKILL.md host (Claude Code, Hermes, OpenClaw)"13metadata:14 source: https://skillmd.com/skills/ashfordeou/e10-tradeoffs15 domain: space-systems16 subdomain: ecss17 tags: [ecss, e-st-10-system-scope, tradeoffs, trade-off-analysis, weighted-scoring, annex-l, decision-analysis]18 version: 0.1.019 author: Aero Agent Skills20---2122# ECSS System Engineering — Trade-off Analysis (space-systems/ecss/e10-tradeoffs)2324Use when the task is choosing between candidate solutions to a system25engineering decision under ECSS-E-ST-10C clause 5.3.3 -- running a26documented trade-off analysis against a fixed set of weighted27criteria and recording the outcome as an Annex L trade-off report.2829## Domain quick reference3031- A trade-off analysis compares two or more candidate solutions to32 the same decision (a design option, a technology, a supplier, an33 architecture) against a fixed set of evaluation criteria agreed34 before scoring starts. Each criterion carries a weight expressing35 its relative importance; weights are normalized to a common total36 so studies with different numbers of criteria stay comparable.37- Some criteria are mandatory (pass/fail) rather than weighted: a38 safety, regulatory, or interface-compatibility criterion below its39 pass threshold disqualifies the candidate outright, no matter how40 well it scores elsewhere. Weighted scoring only ranks candidates41 that clear every mandatory criterion.42- Every candidate is scored against every criterion on the same fixed43 scale so scores are comparable across candidates; a candidate44 missing a score for a criterion, or scored on a criterion outside45 the agreed set, is a broken trade-off study and is rejected before46 ranking.47- The trade-off report (Annex L) does not just declare a winner: it48 records the ranking, the rationale (criteria, weights, scores), and49 whether the outcome is a close call -- a thin margin or a tie at50 the top means the recommendation is sensitive to the weighting51 assumptions and must be reported as such, not smoothed into a false52 single winner.5354## Workflow55561. Define the criteria set and each criterion's weight before scoring57 any candidate; mark any pass/fail criterion as mandatory with its58 pass threshold. Reject a criteria set that is empty, carries a59 non-positive weight, or marks a criterion mandatory without a pass60 threshold -- the study cannot proceed.612. Score every candidate against every criterion on the fixed scale.62 Reject a candidate whose scores do not cover exactly the agreed63 criteria set (missing or extra criterion) or that carries a score64 outside the scale -- fix the input before ranking.653. For each candidate, check every mandatory criterion's pass66 threshold. A candidate below any mandatory threshold is67 disqualified and is not weighted-scored.684. Weighted-score every qualified candidate: normalize the criteria69 weights to sum to one, then sum weight x score per criterion.705. Rank the qualified candidates by descending weighted score, using71 competition ranking so tied scores share a rank and the next72 distinct score resumes at its true position (not the next integer).736. Check whether the top of the ranking is a close call: the top two74 qualified candidates are tied, or the leader's score is within a75 small margin of the runner-up's. Report a single winner only when76 the ranking is not a close call and exactly one candidate holds77 rank 1; otherwise report the ranking with the close-call flag set78 and withhold a single recommendation.7980## Pitfalls8182- Scoring candidates before the criteria and weights are agreed --83 criteria chosen or reweighted after seeing the scores bias the84 outcome toward whichever candidate was already favored.85- Letting a high weighted score on other criteria compensate for a86 failed mandatory criterion -- mandatory criteria are a gate, not87 another weighted input; a disqualified candidate stays disqualified88 regardless of its other scores.89- Declaring a single winner when the leader and runner-up are tied or90 within a thin margin -- a razor-thin or tied result is itself a91 finding (the decision is sensitive to the input assumptions) and92 belongs in the trade-off report, not hidden behind a confident93 recommendation.94- Comparing candidates whose scores were assessed against different95 criteria sets or scales -- a candidate that skipped a criterion, or96 was scored on a scale the others weren't, produces a ranking that97 looks complete but is not actually comparable.9899## Behavior contract (gate 3)100101The criteria/weight validation, candidate-score validation, mandatory102pass/fail gating, weighted ranking, and close-call sensitivity logic103is exercised by the gate 3 contract test:104scripts/test_e10_tradeoffs.py against scripts/e10_tradeoffs_logic.py105(stdlib unittest, offline). Run:106python3 scripts/test_e10_tradeoffs.py107108## Compliance109110- ECSS standards are freely downloadable (ESA); cite the source and111 paraphrase per standards-map.yaml.112- compliance: STANDARDS-REF, gated: false.