ADR-0097: Portfolio vs market — live index context, not a fake benchmark

Source: section ### ADR-0097: Portfolio vs market — live index context, not a fake benchmark of the project decisions log (project-iris-prd/13-decisions-and-rationale.md), extracted on 2026-10-18 for the Evarness repository. It predates the rename from IRIS and is kept as written.

Decision: "How is my portfolio doing against the market" appends live market index context (Nifty 50 / Sensex / Bank Nifty, today's move) alongside the portfolio, fetched locally via fetch_indexes("IN") — never a web search. format_market_context (in finance/portfolio.py) renders the indexes with an explicit honest note, and the portfolio tool fetches it only when the query mentions market/index/benchmark/vs/against. It deliberately does not claim "you beat the market": a unit test forbids that language.

Rationale: The data was verified first (the rule): only 20 of 57 positions are intraday-priced — 2.4% of portfolio value (the rest are mutual funds at last statement NAV), and there's no historical index data locally. So a same-period return comparison ("you're up 39% vs the market's X%") is not honestly computable. Rather than fabricate one, the feature gives what's true and useful: the market's current state next to the portfolio, locally — which directly replaces the web-search-the-market behaviour that leaked the user's name (issue 0031), with framing that says plainly why the index's today move and a lifetime P&L aren't a like-for-like benchmark. A real over-time comparison is deferred to accumulating index history (as net worth now does, ADR-0096). Narrative docs/blog/2026-06-29-the-market-next-to-you.md.

---