Why Two Correct Numbers Can Still Disagree: Reconciling Different Return Calculations

Two trustworthy numbers that still don’t match

Every number quoted about a fund comes from a method, and different, equally valid methods can produce different answers for the same underlying NAV history. This is not a data error or a sign one source is wrong. It is the expected result of asking slightly different questions: “what did this fund return from exactly this date to exactly that date” and “what has this fund typically returned over windows of this length” are two different questions, and they can have two different, both-correct answers.

A real example

For SBI Focused, a single point-to-point 3-year calculation, from the NAV nearest 2023-10-31 to the NAV nearest 2026-10-23, gives an annualised return of 13.99%. The same fund’s official 3-year rolling CAGR, the median across every 3-year window in its history, is 16.6418%. The gap is 2.65 percentage points.

Both numbers are computed correctly, from the same real NAV series. The point-to-point figure is the actual outcome of one specific 3-year holding period, starting exactly on 2023-10-31. The rolling median describes the typical outcome across every 3-year window the fund has had, which is a different, broader question.

Where this kind of gap usually comes from

A handful of genuine methodology differences account for most real-world disagreements between two sources reporting on the same fund:

• Single window vs the full rolling distribution. As above: one specific start and end date versus the median, minimum or maximum across many windows.
• Different as-of dates. NAV updates daily, so a figure computed yesterday and one computed today are answering a question about two different end dates, even if everything else about the method is identical.
• Direct vs Regular plan. The same scheme’s two plans have different expense ratios and therefore different NAVs and returns, as covered in Direct vs Regular Plan. A return quoted without saying which plan it refers to is incomplete.
• Growth vs IDCW option. IDCW payouts reduce the NAV each time they are paid, so comparing a Growth NAV return with an IDCW NAV return for the “same” scheme compares two different things, as covered in IDCW vs Growth.
• TRI vs price-return benchmark. A benchmark index quoted as a Total Return Index, which assumes dividends reinvested, will show a different number than the same index quoted as a price index.

How to reconcile two conflicting numbers yourself

Before assuming one source is wrong, check: the exact start and end dates used, whether it is a single window or a rolling statistic, which plan and option, and whether the benchmark (if one is quoted) is a TRI or a price index. In most real cases, once these are matched, the numbers converge or the remaining gap becomes small and explainable, rather than a genuine data error.

Frequently asked questions

If two sources disagree, which one is right?

Often both are, because they are measuring different things. Check the methodology notes - window length, single point vs rolling statistic, plan type, as-of date - before concluding either is wrong.

Why does Punji show a rolling median instead of a single point-to-point return?

Because a single point-to-point figure depends heavily on the exact dates chosen, as covered in Rolling Returns vs Point-to-Point Returns. A rolling median is less sensitive to that choice and describes the fund’s typical behaviour more fairly.

Can genuine data errors still happen?

Yes. A stale NAV, a wrong plan or option mapped to the wrong scheme, or an unadjusted corporate action can all cause a real discrepancy, separate from the methodology differences above. If a gap is far larger than these explanations account for, it is worth flagging as a possible data issue rather than assuming it is methodology.