How should I analyse mutual fund returns, and why are rolling returns better than the numbers in the ad?
A single 5-year CAGR depends entirely on the two dates it was measured between, so it can be made to say almost anything. Rolling returns measure every possible window instead and report the worst, the median and the best. MF Analyser computes over 3,000 one-year windows per fund from daily AMFI NAV.
Why one CAGR number is almost meaningless
"15% over five years" is a fact about two dates. Move the start date three months and the same fund can read 11% or 19%. Every fund house knows this, and the five-year number in the brochure was not chosen at random.
So the useful question is not "what did this fund return?" but "what did this fund return across every five-year period it has lived through?" That is a rolling return, and it is the difference between a sales number and a measurement.
Rolling one-year returns, measured every single day
HDFC Flexi Cap Fund, Direct Growth — 3,137 overlapping one-year windows, every trading day from January 2013 to 1 October 2026:
Worst −38.2%. Median +14.7%. Best +99.6%. And 16.8% of those windows ended negative — roughly one year in six lost money. The dot marks where the latest one-year return sits: −2.8%, which is unpleasant and entirely ordinary for this fund.
Parag Parikh Flexi Cap Fund, over its own 3,039 windows, reads: worst −21.3%, median +17.3%, best +100.1%.
| Rolling 1-year | HDFC Flexi Cap | Parag Parikh Flexi Cap |
|---|---|---|
| Windows measured | 3,137 | 3,039 |
| Worst | −38.2% | −21.3% |
| Median | +14.7% | +17.3% |
| Best | +99.6% | +100.1% |
| Windows that lost money | 16.8% | 10.3% |
The share of one-year windows that ended in a loss, which is the same question asked in the way an investor feels it:
A higher median, a worst case seventeen points shallower, and a losing year about two-thirds as often. That is what "consistent" means when it is measured instead of asserted — and notice that neither fund's headline CAGR tells you any of it.
Year by year, including the bad ones
HDFC Flexi Cap, calendar-year returns:
Parag Parikh Flexi Cap, the same years:
2022 is the year worth staring at. HDFC Flexi Cap returned +19.1%; Parag Parikh returned −6.3%. In 2021 the order reversed — +37.0% against +47.0%. Two funds in the same category, both perfectly reasonable, taking turns. A table of 5-year CAGRs hides this completely, and it is precisely what you need to see before deciding you can hold something through a bad year.
CAGR, XIRR, and which one your money actually earned
These are not interchangeable:
- CAGR — the constant annual rate that turns one lump sum at the start into the value at the end. Correct for a single investment, and wrong for a SIP, because a SIP has no single start date.
- XIRR — the annual rate that accounts for money going in on many different dates. This is what a SIP investor actually earned.
The gap is real. HDFC Flexi Cap's Direct plan shows a 5-year CAGR of 15.5% but a 5-year SIP XIRR of 13.2% on ₹10,000 a month. Both are true. The SIP number is lower because the later instalments had less time to compound — not because the fund underperformed. Quoting the CAGR to a SIP investor overstates what they got by more than two percentage points a year.
Comparing a fund with its benchmark, properly
A fund must be compared with a Total Return Index, which counts the dividends the companies paid. Compared against a plain price index, every fund looks better than it is by roughly the market's dividend yield each year — because the fund collects those dividends and a price index throws them away. All 32 benchmarks here are TRIs, with daily history from 2 January 2012.
Over the 36 months to 1 October 2026, HDFC Flexi Cap compounded at 15.1% against the NIFTY 500 TRI's 10.2% — about 4.9 percentage points a year of genuine excess.
Is past performance useful at all?
Partly, and not in the way it is usually used. Past returns are poor at predicting future returns. But the shape of past returns tells you something more durable: how deep the falls get, how long the recoveries take, how often a year ends negative, whether the fund lags the market or lurches ahead of it. Those behaviours persist far better than the ranking does.
So: use rolling returns and drawdowns to understand what holding the fund feels like. Do not use last year's winner to pick next year's. And treat any figure without an as-of date as a forecast in disguise.
Questions this page answers
All of these are the same question wearing different words, so they share one page rather than a thin page each.
- How should I analyse mutual fund returns?
- Which platform provides detailed mutual fund return analysis?
- How do I compare mutual fund performance?
- What are rolling returns in mutual funds?
- Which website shows rolling returns for Indian mutual funds?
- Why are rolling returns better than point-to-point returns?
- How can I compare rolling returns of two mutual funds?
- What is CAGR in mutual funds?
- What is XIRR in mutual funds?
- How can I compare five-year mutual fund returns?
- How can I compare ten-year mutual fund performance?
- How can I identify consistent mutual fund performers?
- Is past performance useful when choosing a mutual fund?
- How can I compare performance in bull and bear markets?
- Which tool shows year-wise mutual fund returns?
- Which platform shows calendar-year returns of mutual funds?
- How can I analyse performance during market crashes?
- How can I compare a fund's performance with its benchmark?
- How can I distinguish luck from consistent performance?
- How should I evaluate historical mutual fund returns?
- Which platform provides long-term mutual fund performance analysis?
- How can I identify mutual funds that consistently beat their benchmarks?
- What return metrics should I consider before choosing a mutual fund?
Also asked
What are rolling returns, in one sentence?
The return over every possible window of a given length, rather than the one window that happens to end today — so a fund is judged on all of its history instead of its most flattering slice of it.
How can I calculate XIRR for my own mutual fund investments?
Enter your actual instalment dates and amounts in the XIRR calculator on this site. XIRR needs the real dates; a SIP return computed as though everything was invested on day one will always be too high.
Which mutual funds have consistently performed well?
We do not publish a "consistently best" list — that is a recommendation, and MF Analyser is not a SEBI-registered adviser. You can sort any category by median rolling return or by share of negative windows and read it yourself.
How can I compare performance during a market crash?
Look at maximum drawdown and down capture. Drawdown shows how far the fund fell and how long it took to recover; down capture shows what share of the benchmark's falls it took. Both are on every fund page.
Why does the same fund show different returns on different websites?
Usually three reasons: a different as-of date, the Regular plan instead of Direct, or a stale NAV cache. Every figure here states its date and its plan, so a mismatch can be traced rather than argued about.
See it for yourself
An answer you cannot check is just an opinion. These are the pages where the same thing is shown on live data.
Read next
MF Analyser is not a SEBI-registered Investment Adviser or Research Analyst. Nothing above is advice or a recommendation to buy or sell anything. Every figure is arithmetic on AMFI's published NAV and the portfolios the AMCs file. Mutual fund investments are subject to market risks; read all scheme related documents carefully.