Calculator
Why an average return is not what you get
Up 50, then down 50
−25%
Lowest drag measured
0.60%
Highest drag measured
16.45%
Steady 8% for 20 years
$466k
Same average, that variability
$17k
Instruments measured
8
The cost of variability
Set the average, the spread and the horizonThe simple average, which is how returns are usually advertised.
How far a typical year lands above or below that average.
Measured from real price history
Annualised variability of daily moves over the past year. Our calculation on published closing prices.
Advertised average
8.00%
What actually compounded
6.13%
Lost to variability
1.87%
The same average return, with and without the variability
Up 50% then down 50% is an average of zero and leaves you down 25%, because the fall applies to a larger balance than the rise did. That gap is the whole effect, and it grows with the square of the variability, which is why a very volatile holding needs a much higher average return to finish level with a steady one. It is also why an advertised average return is not what you would have received. Returns here alternate above and below the average by a fixed amount rather than being drawn at random, so the result is a property of the arithmetic and the same inputs always give the same answer.
What real instruments would cost
An identical 8% average return, held 20 years| Instrument | Variability | Compounds at | Lost to variability |
|---|---|---|---|
| FTSE 100 | 11.3% | 7.40% | −0.60% |
| S&P 500 | 12.9% | 7.23% | −0.77% |
| Gold | 28.4% | 4.19% | −3.81% |
| Nikkei 225 | 28.6% | 4.14% | −3.86% |
| Bitcoin | 35.7% | 1.92% | −6.08% |
| Ethereum | 53.3% | -6.08% | −14.08% |
| Crude oil WTI | 54.0% | -6.46% | −14.46% |
| Solana | 57.3% | -8.45% | −16.45% |
Every row is given the same 8% average annual return and differs only in how much it varies around it. The final column is what each instrument actually did over the past year, included so the variability figures are not mistaken for returns. Variability is our calculation from published closing prices; the drag columns are arithmetic on the stated assumption, not a measurement of that instrument.
Where the gap comes from
A fall applies to a larger balance than the rise that follows it. Lose 50% of $100 and you have $50; gain 50% of $50 and you have $75. The percentages cancel, the money does not.
Repeat that across many years and the shortfall compounds. The average return is unchanged throughout — only the path differs.
Why it is worth knowing
Marketing quotes averages. Averages are the higher of the two numbers, always, unless returns never vary at all.
The figure you would have received is the compounded one, and the gap between them widens quickly: it is negligible for a bond fund and can be several percentage points a year for a cryptoasset.
What this does not claim
It does not say a volatile holding will underperform. A high enough average return overcomes the drag easily, and several instruments listed here have done exactly that.
It says that for any given average, more variability leaves you with less — so the two have to be judged together, and an average quoted alone tells you very little.
How it is calculated
Annual returns alternate a fixed amount above and below the average you set, which holds the arithmetic average exactly at that figure and isolates the effect of the spread.
Nothing is random, so the same inputs always give the same answer. Variability is the annualised standard deviation of daily moves over the past year.