Why total return alone cannot be compared
A 50% gain over one year and a 125% gain over three years: on total return the second looks better.
But it took three times as long. To judge which moved faster, you have to remove the time dimension.
Three steps
Step 1 — for each investment compute ROI = (end − start) ÷ start × 100%.
Step 2 — record the holding period in years, using decimals for part-years (six months is 0.5).
Step 3 — convert to CAGR = (end ÷ start)^(1 ÷ years) − 1 and compare the annualised numbers.
Three cases compared (recompute them yourself)
A: 1,000 in, 1,500 after 1 year → ROI 50%, CAGR 50%.
B: 1,000 in, 2,250 after 3 years → ROI 125%, CAGR = 2.25^(1 ÷ 3) − 1 ≈ 31.04%.
C: 1,000 in, 1,200 after six months → ROI 20%, CAGR = 1.2^2 − 1 = 44%.
Ranked by total return the order is B > A > C. Ranked by CAGR it becomes A (50%) > C (44%) > B (31.04%) — the same data, the opposite conclusion.
Now hold the return fixed and vary the period: the same +50% is 50% a year when done in one year, but only 1.5^(1 ÷ 3) − 1 ≈ 14.47% when it takes three.
Notes
CAGR uses only start, end and years, so it ignores drawdowns and volatility along the way. Two investments with the same CAGR can carry very different risk.
The shorter the period, the more the annualised figure is inflated. A six-month or one-month result should not be extrapolated to a full year.
This page harmonises the time convention only. It does not judge which investment is "better" and makes no price forecast.
FAQ
Is a higher total return always the better investment?
No. In the examples B returns 125% in total against A's 50%, yet B's CAGR is 31.04% versus A's 50%, because B took two extra years.
Why does C, at only six months, show a 44% annual rate?
Annualising restates a short result on a one-year compounded basis: 1.2² = 1.44. It is a conversion, not a repeatable outcome for the next six months.
If two investments have the same CAGR, can I compare them directly?
Not fully. CAGR omits volatility and drawdown, so you also need the path and the risk. This page only fixes the period convention.