Nothing behind a black box.

01 · Latest NAV & dates

Each fund ends on its latest valid saved NAV, including the latest report. Histories are date-sorted, duplicate dates collapse to one observation, and non-positive or invalid values are excluded. The latest-report value wins if the same date occurs in both sources.

Lookbacks use calendar months/years with month-end clamping and the latest NAV on or before the target. This implementation rejects a fallback more than 7 calendar days old, so missing imports do not silently become distant lookbacks.

02 · Risk window

Risk uses 36 consecutive completed calendar-month returns from 37 month-end NAVs. The current month is excluded unless the latest NAV is dated on its final calendar day. Month-end fallback is limited to 7 days. Missing endpoints make the risk metrics unavailable.

Standard deviation uses the sample denominator n − 1 and √12 annualisation. Rp is the geometric annual return between the same risk endpoints using actual days / 365.25. Annual Rf and MAR are 6.5%; monthly MAR = (1.065)^(1/12) − 1. No substitute 12% market return is used.

03 · Drawdown & frequency

Maximum drawdown, recovery and peak metrics use all saved NAV observations. Calmar divides available-history CAGR by the absolute drawdown over that same history. CAGR/Calmar require at least one year; a zero denominator gives N/A.

Successive NAV returns are used for daily statistics, with jumps over 14 calendar days excluded as a data-quality guard. Worst 5% Day is the linearly interpolated inclusive 5th percentile. Monthly Win Rate uses the same 36 monthly returns as risk. Gaps and incomplete backfills can understate observed drawdowns.

04 · Rolling years & SIP

Rolling 1Y evaluates every eligible saved ending date against its calendar-year lookback. Only endpoints with the stated 7-day fallback are included; missing windows are not filled.

SIP uses 60 monthly contributions of ₹10,000 (₹6,00,000 total), on the first day of each month ending in the latest NAV’s month. The purchase NAV is the latest observation on/before the first, within 7 days. Cash flows are dated on the scheduled first; redemption is valued at the latest NAV. This is a disclosed NAV simulation, not an execution-price guarantee.

XIRR solves dated cash flows with actual days / 365 using a bracketed numerical solver. Fees, taxes, loads, IDCW payouts and distribution reinvestment units are excluded. All 60 purchase observations are required.

05 · Plans & benchmarks

Direct/Regular and Growth/IDCW labels use explicit fields first, then recognisable scheme-name labels. Unlabelled schemes stay Unknown; missing plan labels are never assumed Regular. Distinct AMFI codes are never combined.

IDCW figures are NAV-only and exclude cash distributions and additional reinvestment units. They are not total returns and should not be treated as like-for-like Growth returns. Beta, Alpha and Treynor show N/A because benchmark TRI history and scheme-to-benchmark mapping are not connected.

06 · Freshness & speed

Only selected funds are loaded, up to six at once. A background worker calculates all 39 metrics once. Changing the ten visible metrics reuses the result immediately.

The browser caches derived results in IndexedDB (up to 256 funds). Each refresh checks a content version covering the whole saved NAV history, latest NAV and metadata. Older backfill additions and NAV corrections invalidate the cache even when the latest date stays unchanged. Failed checks are explicitly labelled; a cache entry is not proof of live market data.

Auto-check runs every 60 seconds while this page is visible. Values update when new records are committed to your database; this page does not fetch new NAVs from AMFI. Your import and any future scheduler remain separate.