UR pools & migration
Known relaxed paper scores no longer enter the ordinary UR assumption. Non-fee concession history is unknown, not confirmed eligibility. The shared correction applies to sample, forms and reporting-bias/community calculations.
Software tests do not establish every recruitment rule or candidate document. This audit corrects identifiable logic gaps and separates conditional models from verified published results. It does not certify final selection.
Published result records, cutoff notices, source PDFs and frozen historical model comparisons are unchanged. Model outputs may differ because of the documented rule correction—not because scores, coverage or multipliers were tuned to a desired cutoff.
Official Results →Known relaxed paper scores no longer enter the ordinary UR assumption. Non-fee concession history is unknown, not confirmed eligibility. The shared correction applies to sample, forms and reporting-bias/community calculations.
Total potentially eligible category population is kept separate from the smaller ordinary-UR pool. Removing a relaxed candidate from UR does not erase them from their own-category population.
A special tag no longer excludes a candidate from ordinary merit or automatically deducts a seat. Candidate eligibility still needs verification.
Same-category horizontal minimum adjustment, widow↔divorce fallback, then other women, then men if women unavailable. BCI printed subquota counts are retained in the sandbox inputs. This is not the full multi-quota/department roster.
Removed unsupported default 56-seat subtraction. Area-wide PwD seats cannot produce a category-only cutoff. Zero sandbox holdout is not a finding of zero PwD reservation.
Rule-complete allocation is unavailable with missing documents/rosters/preferences. Rank sandbox OFF by default. Thin TSP pools are marked non-estimable; missing samples do not become allocation results.
Non-finite standings cannot create fake ties. Invalid/mismatched papers cannot prove qualification. Invalid/fractional top-seat averages are rejected. Score bands exclude invalid, hidden and repeated-ID entries.
Separate typed policy routes distinguish EWS/Sports normal process, women same-category fallback, SC/ST/OBC/MBC recruitment-count carry-forward, EXM next-year carry/lapse and PwD carry/interchange/employer decision. Missing histories do not trigger an invented release. These routes are not a final roster allocator.
Therefore no rule-complete final selection or appointment probability is produced. Rank Scenarios defaults to descriptive lookup; the incomplete-data sandbox requires a separate explicit opt-in. Vacancy policy routes do not automatically allocate or release seats.
The reservation review uses the user-supplied 26-page advertisement07/2026 copy (RajKaj22397996), especially pages8–12 and department preference on page1. Its authenticity/currentness and later amendments have not been independently established.
Open supplied reservation notice PDF ↗Automated regression tests, independent offline numerical checks and local/live browser checks support only the behaviours exercised. They do not authenticate candidate certificates or certify every legal/administrative edge case. No candidate or settings records are edited by this audit.