Rolling returns and drawdown answer different questions about historical mutual-fund data. Rolling returns show the result of repeatedly measuring the same holding period across many start dates. Drawdown shows the fall from a previous high to a later low in a chosen series. Read together, they can describe historical range and stress; neither predicts the next return, recovery time or a household’s ability to bear loss.
Direct answer (53 words): Before reading either metric, identify the exact scheme, plan, option, benchmark or return series, observation frequency, period and source document. Rolling returns describe many overlapping holding-period outcomes. Drawdown describes a peak-to-trough decline. Record the method and dates, then treat missing definitions as unknowns rather than filling them with assumptions.
Reviewed by GFS Research Desk.
The reader problem: one chart, several hidden definitions
A factsheet or data page may place a rolling-return table beside a maximum-drawdown figure. The layout makes the figures look comparable, but they may be built from different inputs. One could use daily NAV observations and another monthly values. One could use a growth option or total-return series while another uses a different plan. The return window may be one year, while the drawdown history covers five years.
The current Scheme Information Document (SID) is the starting point for the scheme’s objective, investment policy, benchmark and risk disclosures. SEBI’s mutual-fund master circular is the official regulatory source to check alongside current scheme material. AMFI’s official data pages are useful for locating published mutual-fund data, but a table still needs its own as-of date and methodology.
This draft was checked on 24 July 2026. Data pages, files, disclosure formats and scheme documents can change. A live review should use the current official document, not this article as a substitute for it.
What rolling returns actually measure
A one-year rolling return asks: what was the result of a one-year window ending on each selected observation date? Instead of choosing only 1 January and 31 December, the calculation moves the window through the available history. A three-year rolling series similarly repeats a three-year window, subject to the source’s stated convention and enough history.
A simple hypothetical illustration makes the structure clearer. Suppose a data set has monthly observations and a stated one-year window. The first observation might cover January to December, the next February to January, then March to February. The windows overlap. These are not twelve independent investments, and they should not be added together or treated as a sequence of future forecasts.
Before interpreting the output, capture:
- the holding-period length;
- the first and last observation dates;
- whether returns are point-to-point or annualised;
- the plan and option;
- the NAV or total-return series used;
- treatment of distributions, if relevant; and
- the source’s frequency, rounding and missing-data rules.
A median, minimum or percentage of positive rolling periods is meaningful only inside those definitions. Two rolling-return summaries with different windows or end dates are not a fair comparison merely because both are labelled “rolling.”
What drawdown actually measures
Drawdown is a decline from a prior high to a later value. A generic point-in-time drawdown can be expressed as:
`(current value ÷ previous running peak) − 1`
If a hypothetical series rises from 100 to 120 and later reaches 90 before making another high, the fall from 120 to 90 is 25%. That arithmetic describes the selected series and dates. It does not say that a person entered at 120, exited at 90, or experienced exactly that result after cash flows, taxes, charges or transaction timing.
“Maximum drawdown” is the largest peak-to-trough fall found in the chosen history under the source’s method. Ask whether the series is NAV, an adjusted or total-return series, or another index; whether observations are daily, weekly or monthly; whether the window has a fixed start date; and how a new peak and recovery are defined. A month-end series can miss an intra-month low that a daily series would capture.
Drawdown also has a recovery dimension. A peak-to-trough percentage alone does not tell you the date of recovery, whether recovery occurred within the sample, or what happened between observations. If recovery information is absent, write “not disclosed” rather than inferring it.
Why the two metrics should not be collapsed
Rolling returns organise outcomes by holding period. Drawdown organises declines by path. A series may have mostly positive five-year rolling periods while still experiencing a substantial interim drawdown. Conversely, a shallow maximum drawdown over a short sample says little about a longer holding-period distribution.
The measures also respond differently to cash flows. A published NAV series usually has no individual investor’s contribution schedule. A household’s SIP or withdrawal path can produce an experience that differs from a lump-sum chart. Do not use a rolling-return percentage as a promise of what a recurring cash-flow pattern will earn, and do not use a maximum drawdown as a precise personal loss estimate.
The useful question is not “which metric is better?” It is “what historical question does each metric answer, and are the inputs aligned?”
The seven-field alignment card
Make one card before copying any figures into notes.
- Identity: exact scheme, plan, option and structure.
- Series: NAV, total return, benchmark or another stated series.
- Window: start date, end date and rolling holding period.
- Frequency: daily, weekly, monthly or another observation rule.
- Return convention: point-to-point, annualised, reinvested distributions or other stated method.
- Drawdown method: running peak, trough, sample boundary and recovery treatment.
- Document trail: URL, file name, publication date, as-of date and section.
If any line is missing, the correct result is a limitation. Do not copy a number into a comparison table while silently assuming that the missing field matches another source.
A safe reading workflow
Start with the definition layer
Read the SID and current KIM for the scheme’s objective, permitted investments, benchmark and risk language. A category name or objective is not a performance series. A permitted allocation range is not proof of a current holding. Use the document stack to identify what is being measured.
Then save the dated data
Download the factsheet or official data file and record its publication and as-of dates. AMFI’s official data repositories can help locate industry information, while the AMC’s current document remains important for scheme-specific definitions. Preserve the original file. A later file at the same URL may replace an earlier snapshot.
Reproduce only simple, matched arithmetic
If the source gives two matched returns for the same period, you may label their subtraction as an arithmetic return gap. For rolling returns, do not recalculate a published statistic unless the observation dates, series and method are clear. For drawdown, identify the peak and trough actually used. Your own calculation should be labelled “illustration” or “reader calculation,” never presented as an official scheme statistic.
Separate observation from interpretation
Write three columns: “source says,” “arithmetic shown,” and “cannot establish.” For example, a source may say that a one-year rolling series had a stated minimum. Your arithmetic may show how a 25% hypothetical fall is calculated. The limitation is that neither establishes future loss, suitability, liquidity or recovery time.
Common traps
Trap: comparing different plans. Direct and regular plan data can have different expense structures. Align identity before comparing a metric.
Trap: using a low NAV as evidence of lower risk. NAV level is not a risk ranking. Units, exposure, costs and distributions need context.
Trap: treating positive rolling periods as a success probability. The historical sample and method determine the statistic. It is not a forecast probability.
Trap: calling every decline a maximum drawdown. A drawdown requires a prior peak and a defined observation series. A single negative month is not automatically the maximum peak-to-trough fall.
Trap: explaining the whole gap with the expense ratio. Expense is one possible contributor to realised differences, but cash, trading, valuation, taxes and other operations can matter. AMFI’s official expense data should be checked for the same plan and period, not used as a mechanical forecast.
Trap: mixing a benchmark chart with a scheme chart. A benchmark may have different construction and return conventions. Keep the series labels visible.
A reader worksheet
For each figure, fill: “metric,” “exact definition,” “series,” “window,” “frequency,” “as-of date,” “source section,” and “limitation.” Mark each field as verified, not shown or not applicable. Keep the original document and note the date you accessed it.
For a rolling-return table, ask whether the displayed period is annualised and whether the endpoints are inclusive under the stated method. For a drawdown chart, ask whether the peak and trough are visible, whether the chart is daily or periodic, and whether recovery is shown. If the answer is not in the source, stop at “not disclosed.”
This small discipline makes the output auditable. It also prevents a polished chart from carrying more certainty than its documentation supports.
FAQs
Are rolling returns the same as annual returns?
No. Annual returns usually describe selected annual periods. Rolling returns use overlapping windows that move through the history. The source must state its window and convention.
Does a maximum drawdown mean an investor lost that exact percentage?
No. It describes a selected series from a prior high to a later low. Entry date, cash flows, distributions, taxes and transaction effects can differ.
Does a positive rolling-return history prove future performance?
No. It is a historical summary under a specified sample and method. Past performance is not a forecast.
Can I compare a scheme’s drawdown with its benchmark?
Only after aligning the series, dates, frequency, return treatment and definitions. Even then, the comparison describes history and not personal suitability.
Is a longer rolling window always safer?
Not necessarily. A longer window changes the question and can hide interim declines. It does not remove market, liquidity or sequence-of-returns risks.
Where should I look when a factsheet omits the method?
Start with the current AMC factsheet, SID, KIM, portfolio and data disclosures, then check the relevant official AMFI or SEBI source. If the method remains absent, record it as an unknown.
Honest limitations
This guide cannot calculate an individual’s return, taxable result, liquidity need, risk capacity or suitable investment. It does not rank schemes or predict returns. A historical chart may contain survivorship, data-quality, plan, option, valuation and sample-period limitations. Verify current scheme documents and the exact source method before acting, and obtain qualified professional help for personal tax or legal questions.
> Mutual fund investments are subject to market risks. Read all scheme-related documents carefully.
> This content is educational and is not investment advice or a recommendation. Verify independently before acting.
> Past performance is not indicative of future returns.