Rolling returns — what was the worst window, ever?
Aggregate CAGR hides the ride. A 14% long-term number can contain 5-year windows that returned -3% and 5-year windows that returned +28%. Which one you got depended entirely on when you started. This tool slides through every possible N-year window in the history and shows the worst, best, median, and percentile bounds — free. Freefincal has this as an Excel; no Indian web tool has it.
Rolling returns analysis
Slides a window across all of history, computes CAGR for each, aggregates best/worst/median. NIFTY data from 2000, Sensex from 1990. Stocks vary.
"20-year CAGR is 14%" hides the dispersion of outcomes you actually lived through. Rolling returns show: "what was the WORST 5-year window ever?" — the floor that matters for your actual retirement corpus.
Why rolling returns matter
Point-to-point returns ("NIFTY gave 14% from 2005 to 2024") depend on a single start and single end date. Rolling returns look at EVERY possible start-end combination of a given length. This matters because:
- • You almost never invest exactly at the start date used in press-release CAGR claims.
- • You'll typically hold for a fixed duration (3, 5, 10 years) — what matters is the DISTRIBUTION of outcomes at that duration.
- • The worst rolling window tells you your downside — the floor you need to be financially + emotionally prepared for.
The rule of thumb: NIFTY's 5-year rolling returns have NEVER been negative post-2005. 3-year rolling has been negative ~5% of the time. 1-year rolling has been negative ~20% of the time. Your holding period determines your probability of loss, not your skill.
Related backtests
Research tool · not investment advice.
Axel Markets is an information + analytics product. We are not a SEBI-registered Research Analyst (RA) or Investment Adviser (IA). Nothing on this page is a buy, sell, or hold recommendation. Past performance is not indicative of future returns. Verify all data against the authoritative source (NSE, BSE, AMFI, SEBI, company RHP / factsheet) before acting. Rolling return windows are computed on Stooq daily closes. Past rolling distributions don't predict future ones — regime shifts (rate cycle, valuations, earnings trajectory) change the underlying process.