Journey 12 — Rank verification

solvjump verify-ranking checks that finalists ordering holds under tightened SRK thresholds — without re-sweeping the full grid.

Prerequisites: Journey 07 or 13 — session with ranked finalists exists. Scenario: Prove jump order vs full SolvSRK — do not re-run the grid

Key APIs

• solvjump verify-ranking <session_id>
• ranking thresholds
• satellite_comm quick session (alternative prerequisite)

Concepts to internalize

• Verification ≠ resweep — spot-checks ordering; does not rerun 256+ point grid.
• Threshold knobs — tighten rtol/atol or survival gates to stress the ranking.
• Satellite sessions — Journey 13 quick tier also produces verifiable rankings.
• Compliance use — attach verify-ranking log to study sign-off packages.

Important code

The guided tour script wraps this pattern. Copy the production shape into your pipeline:

solvjump verify-ranking <session_id>

Act-by-act walkthrough

Act 1 — The client question

Jump-ranked finalists_topN.csv is an extrapolation. verify-ranking re-scores a stratified sample with SolvSRK. Thresholds: --srk-sample-n, --min-spearman, --min-top-n-overlap.

• Using session <session_id> (<product-root>\python\examples\brusselator_model.py:make_rhs)

Act 2 — Run the gate

Same as: solvjump verify-ranking <id> --srk-sample-n 20

Rank verification (jump vs full SRK)

• top-50 overlap: skipped (sample size 9 < top_n 50)

• top-100 overlap: skipped (sample size 9 < top_n 100)

• steady_ok agreement (jump vs SRK): 78%

• Report: <session-dir>\rank_verification_report.json

• SRK shortlist (verified ranking for HPC handoff): <session-dir>\srk_shortlist.csv

Act 3 — Read the report

Check rank_verification_report.json before citing 'best' rows.

• top-10 overlap: skipped (sample size 9 < top_n 10)

• top-50 overlap: skipped (sample size 9 < top_n 50)

• top-100 overlap: skipped (sample size 9 < top_n 100)

Act 4 — Next

Journey 13 scales a comm-window search and shows why HPC still exists.

Guided tour output

Verified console capture from 12_rank_verification.py:

Prove jump order vs full SolvSRK — do not re-run the grid

-- Act 1 - The client question --
  -> Jump-ranked finalists_topN.csv is an extrapolation. verify-ranking re-scores a stratified sample with SolvSRK. Thresholds: --srk-sample-n, --min-spearman, --min-top-n-overlap.
  log: Using session <session_id> (<product-root>\python\examples\brusselator_model.py:make_rhs)

-- Act 2 - Run the gate --
  -> Same as: solvjump verify-ranking <id> --srk-sample-n 20
  _section: Rank verification (jump vs full SRK)
  verified_rows: 9/9 jump-ranked rows
  spearman_rho: 0.817 (gate >= 0.85)
  srk_verify_time: 0.8 s
  best_srk_jump_rank: #4 (regret 3)
  verdict: FAIL — do not cite ranking to clients
  -> top-10 overlap: skipped (sample size 9 < top_n 10)
  log: top-50 overlap: skipped (sample size 9 < top_n 50)
  log: top-100 overlap: skipped (sample size 9 < top_n 100)
  log: steady_ok agreement (jump vs SRK): 78%
  log: Report: <session-dir>\rank_verification_report.json
  log: SRK shortlist (verified ranking for HPC handoff): <session-dir>\srk_shortlist.csv

-- Act 3 - Read the report --
  -> Check rank_verification_report.json before citing 'best' rows.
  passed: False
  spearman_rho: 0.8166666666666667
  rank_regret: 3
  srk_verified_rows: 9
  verdict: FAIL — do not cite ranking
  -> Full report: <session-dir>\rank_verification_report.json
  log: top-10 overlap: skipped (sample size 9 < top_n 10)
  log: top-50 overlap: skipped (sample size 9 < top_n 50)
  log: top-100 overlap: skipped (sample size 9 < top_n 100)

-- Act 4 - Next --
  -> Journey 13 scales a comm-window search and shows why HPC still exists.

-- Run complete --
  session_id: <session_id>
  results_folder: <session-dir>

Run it

cd python
# Guided tour — narrated scenario walkthrough (recommended first run):
python examples/journeys/12_rank_verification.py
 
# Production one-liner — same API call you ship:
solvjump verify-ranking <session_id>

What comes next

Journey 13 runs satellite comm HPC tiers at operations cadence (72 × 300 s steps).

Navigation

← Journey 11 — Golden proof · Journey 13 — Scale and HPC gate →