v2.0 — 2026-04-27
Headline: Six-axis formula re-architected to nine axes. Hard caps formalised. Segment normalisation introduced. NRI overlay added.
What changed
- Added Legal & Title Cleanliness (18%), Builder Financial Health (14%), and Governance & Approvals Depth (6%) as standalone axes. v1 buried these inside RERA Compliance and Builder Track Record.
- Reweighted Verified Buyer Sentiment from 20% to 10% with mandatory bot filtering and possession-cohort follow-ups at 6/12/24 months. Coached-review noise was the single largest source of v1 score drift.
- Split Construction Progress into Construction & Delivery Risk (10%) and Bank Loan-Approval Depth (8%). Bank APF presence is an independent legal proxy and deserved its own axis.
- Formalised eight hard caps with explicit ceilings — no valid RERA caps the score at 4.9, NCLT/IBC admission at 3.9, OC missing on RTM at 5.4, promoter conviction at 3.5.
- Introduced five-tier segment normalisation (Affordable, Mid, Premium, Luxury, Ultra) with per-axis weight modifiers, and the additive NRI-Safe flag.
Why
One year of running v1 across 1,721 projects exposed three Indian-specific failure modes that v1 under-weighted. The Amrapali, Jaypee and Unitech failures were title and financial-health failures masquerading as construction risk — visible in MCA filings and EC searches years before any QPR red-flagged them. v1 read those signals into a single Builder Track Record axis at 15% weight; v2 isolates them. The Supertech Emerald Court demolition was a governance-and-approvals failure that v1 treated as a sub-question; v2 makes it an axis. And the chronic gap between Google Reviews sentiment and possession-cohort reality — well documented across Pune and outer Hyderabad — forced the sentiment reweighting.
Affected projects
All projects in the Brickplot universe were re-scored under v2 the day this rubric went live. Projects whose v2 score moves more than one point from v1 are surfaced on the Avoid List and Wait List with the delta annotated.