• BTC$77,566.00+0.88%
  • ETH$2,257.13+0.97%
  • USDT$1.00+0.05%
  • BNB$702.85+0.52%
  • ICP$2.52-0.98%
Bitcoin Data GuideBitcoin Reference

Returns & Performance

Return measured against risk

A return figure on its own says nothing about what it cost to earn. Risk-adjusted measures divide the return by the volatility taken to produce it, and this page publishes both ratios for each window alongside the assumptions behind them.

2010-2025Source: Coinbase Exchange daily candles; Bitstamp and CoinDesk historical series for 2010-2014Each window uses the annual returns from the curated yearly table. The mean return is the arithmetic average of the annual returns in the window and the volatility is their population standard deviation, both in basis points. The Sharpe ratio is the mean excess return over the risk-free assumption divided by the volatility; the Sortino ratio divides by the downside deviation, computed over the negative returns only. The risk-free assumption is a constant 2.00 per cent per year, a round long-run average for short-dated US Treasury yields over the period, and it is applied uniformly rather than year by year. Ratios are published as integers scaled by one thousand. Windows with fewer than three observations are not published, and the early windows rest on the thinly traded 2010-2014 market, whose corrected full-year returns are about 385 per cent for 2010, 1,437 per cent for 2011 and 5,463 per cent for 2013.

Windows measured

3

Distinct measurement windows

Highest Sharpe

0.78

2020-2025 (recent)

Risk-free assumption

2.00%

Annualised, applied to every window

The ratios by window

Each row reports the same four quantities for a different measurement window. The mean return is the average annual return over the window; the volatility is the annualised standard deviation of returns over the same period. The Sharpe ratio divides the excess return — the mean return less the risk-free rate — by that volatility. The Sortino ratio does the same but divides by downside deviation only, so it does not penalise upside surprises. The observations column records how many return periods the window contains, which is the first thing to check before trusting any of the other figures.

Bitcoin risk-adjusted return measures by window, with mean return, volatility, risk-free rate, Sharpe ratio, Sortino ratio and observation count.
WindowMean returnVolatilityRisk-freeSharpeSortinoObservations
2010-2025 (full series)+612.9%1382.7%2.00%0.4420.8515
2015-2025 (post-2014)+192.0%384.9%2.00%0.496.4411
2020-2025 (recent)+94.8%118.7%2.00%0.783.526

2010-2025Source: Coinbase Exchange daily candles; Bitstamp and CoinDesk historical series for 2010-2014

The risk-free assumption. Every ratio in this table subtracts the same annualised risk-free rate, 2.00%, from the mean return before dividing by volatility. That rate is a stated assumption, not a market observation, and it is applied uniformly across windows of different lengths. A reader who prefers a different assumption will get different ratios from the same return and volatility figures; the raw components are published alongside the ratios so the arithmetic can be redone.

What the ratios assume, and where they strain

Both ratios come from the same intellectual tradition: they treat volatility as the definition of risk and reward the investor for bearing it. That framing works well for a diversified portfolio of assets whose returns are roughly symmetric. It works less well for an asset like Bitcoin, whose return distribution is not symmetric and whose largest moves have historically been upward. A measure that penalises all deviation from the mean treats a sudden doubling as a risk of the same kind as a sudden halving, which is not how most holders experience it.

The Sortino ratio exists precisely because of that objection. By dividing only by downside deviation, it stops charging the asset for its positive surprises. The two ratios in the table therefore disagree in a systematic way: Sortino is higher than Sharpe wherever the return distribution is skewed upward, and the size of the gap is a rough measure of how much of the volatility was upside. For Bitcoin the gap is wide, and that width is itself informative.

There is a second and more serious limitation. Both ratios are computed from historical returns and assume that the past distribution is a reasonable guide to the future one. For an asset whose market structure has changed as much as Bitcoin's — from unregulated venues to regulated ones, from retail-only to institutional participation — that assumption is weak. A Sharpe ratio measured over the full record blends periods that had very little in common, and a ratio measured over a short window rests on too few observations to be stable.

The practical guidance is to read the ratios as a way of comparing windows rather than as an absolute verdict on the asset. A window with a higher Sharpe than another earned more per unit of volatility over that period, which is a fact about the period. It is not a prediction, and it is not a substitute for looking at the drawdown record, which describes the losses a holder would actually have had to endure. The volatility explainer covers how the denominator of both ratios is measured.

Related reading