Bitcoin Chill Meter

Is Bitcoin in chill mode?

Lots of noise or barely any movement? See how restless Bitcoin is right now.

…

Calm score

Loading the data…

Calm measures restlessness, not upside potential.

Is it moving a lot?

Over 7 days

…

Over 30 days

…

Typical move in a day

What range is the price in?

The last 30 completed days

High: …
Low: …
Range: …

Price change over 30 days…

How long has it lasted?

…

The score moves along a sliding 30-day window: two neighboring days share 29 of their 30 returns. A streak of calm days is therefore not a string of independent observations.

Calm, day by day

At the top, Bitcoin's price. Below, the calm score for the same day. The bluish bands mark calm days, with a score of 60 or more.

Loading the data…

Selected day

Hover or tap the chart, or use the arrow keys, to read a day.

Very chill 80 to 100
Calm 60 to 79
In between 40 to 59
Restless 20 to 39
Very restless 0 to 19
Calm days (score of 60 or more)

BTC/USD price: Bitstamp, daily candles, completed UTC days · How the score is calculated

See the data in a table

When Bitcoin gets boring

With Bitcoin, there are weeks when everyone has an opinion. The price jumps, drops again, and every candle becomes a conversation. And then there are the other weeks, when almost nothing happens.

Those moments are precious. Fewer spectacular moves to comment on, more time to think: go over your plan, check the security of your keys, prepare a decision with a clear head rather than in the middle of a big red candle.

The Chill Meter is there to spot those periods. It only looks at the price: it measures neither the noise on social media nor public interest, only how big the moves are from one day to the next.

And sometimes the calm lasts. Since 2015, the longest streak held for 286 days in a row, from April 2023 to January 2024.

Calm doesn't mean cheap

The score tells you how much the price moves, not whether it is high or low. An asset can be calm at a high price, or after a big drop. Volatility measures the size of the moves, not the value of what you buy.

Example: in January 2023, after a 2022 in which the price had been cut by almost three, the score was above 95 with Bitcoin around $16,800. In August 2025, it hit 100 with Bitcoin around $119,300, seven times more expensive. Same calm, opposite context.

And calm doesn't mean still. From mid-June to late November 2020, the score stayed at 60 or more for 165 days in a row, while the price doubled, from about $9,300 to $18,700. Small moves that often go the same way add up to little volatility, but a long way traveled.

Since 2015, on 42% of calm days the price had moved more than 10% over 30 days. That's why the tool shows the price change and range next to the score.

Is it a good time to buy?

Straight answer: the score alone can't tell you.

It doesn't measure value and it doesn't predict anything. A calm period always ends at some point, and the exit can go either way. In late 2018, after 71 calm days, the price lost 34% over the following month. In early 2023, after 94 calm days, it gained 36% over the following month.

What the score can do for you: if you prefer to look at the market and prepare your decisions when things move less, it shows you those moments. The decision itself depends on your situation, your time horizon and what you're ready to live with.

Calm and fear: two different things

Volatility is observed: it's calculated from real prices, day after day. Sentiment is estimated: it's a reading of the market's mood, built from several signals.

The two don't tell the same story. The market can be calm and worried, or restless and euphoric. So the Chill Meter uses no sentiment index in its calculation, only prices.

To see the other side, the market's mood, take a look at Bitcoin: the market's temperature: Bitcoin's price colored by the Fear & Greed Index.

How is the score calculated?

Every day, we measure how much Bitcoin moved over the last 30 days, then compare that figure with every one from the previous four years. If the last 30 days were calmer than most past periods, the score is high. If they were more restless, it's low.

The method in detail: windows, formulas, source and limits

Data

Source: Bitstamp, BTC/USD market (US dollar, not USDT). One candle per UTC day, midnight to midnight, since August 18, 2011. Only completed days are used: the current day never enters the calculation.

The site queries Bitstamp right after midnight UTC, caches the data and shows two separate dates: the closing day used, and when the site retrieved the data. If the source doesn't respond, the last valid version stays on screen, with its age.

1. Daily returns

For each day: r = ln(day's close / previous day's close). The logarithm makes a rise and a fall of the same size symmetrical.

2. Volatility

Sample standard deviation of the last 30 returns: subtract their mean, divide the sum of squares by 29 (n − 1), then take the square root. Times 100, that gives the order of magnitude of a typical daily move, in percent. The "Is it moving a lot?" card does the same over 7 returns (divided by 6).

Annualization: daily volatility × √365, because the market never closes. It's the order of magnitude of the swings over a year if the current pace lasted. It's shown for information; the score doesn't use it.

3. Comparison with the previous four years

The day's 30-day volatility is compared with the 30-day volatilities of the 1,461 days before it (four years, leap years included). Never the day itself, never a future day: each historical score only uses what existed on its date. At least 1,388 reference values (95%) are required, otherwise there's no score.

4. Percentile and score

P = 100 × (number of lower reference values + ½ × number of equal values) / number of reference values. Ties count as half.

Score = 100 − P, rounded to the nearest whole number, halves rounded up. Low volatility, high score: a score of 80 means the last 30 days were calmer than about 80% of the reference periods.

5. The five states

0 to 19: very restless. 20 to 39: restless. 40 to 59: in between. 60 to 79: calm. 80 to 100: very chill. These are presentation markers, not validated predictive thresholds: going from 59 to 60 changes nothing about the market.

6. Streak of calm days

Number of consecutive days, up to the latest one, with a score of 60 or more. A day without a score breaks the streak. Since each score covers a sliding 30 days, two neighboring days are bound to look alike: a 40-day streak is not 40 independent observations.

7. Missing data

No price is invented or interpolated. A missing, invalid or zero-trade day counts as missing, and every window touching it becomes unavailable: 31 volatilities over 30 days, 8 over 7 days. In the current history: 30 days without trades in 2011, when the market was tiny, and January 6 to 8, 2015, when Bitstamp had suspended its platform after a hack.

First score that can be calculated: December 8, 2015. Before that, there aren't four years of usable history yet.

Limits

One exchange and one closing time (midnight UTC): elsewhere or at another time, the figures would differ slightly.

Calm is relative to the last four years, not absolute. As Bitcoin's volatility has declined over time, the score is often high: since 2015, half of all days are calm or very chill.

The first scores, in late 2015 and 2016, are compared with 2011-2015, a much smaller and more restless market.

The score lags: a very restless day weighs in the window for 30 days, then drops out all at once.

Volatility says nothing about the direction of the price or its value.

Sources consulted

Bitstamp, API documentation: daily BTC/USD candles (step, start, exclude_current_candle parameters), rate limits and data terms of use. Accessed on September 19, 2026.
Coin Metrics, Data Encyclopedia, Volatility: definition of 30-day volatility: standard deviation of daily natural log returns. Accessed on September 19, 2026.
Coin Metrics, Realized Volatility, Rolling 30 Days (USD): realized volatility annualized on a 365-day basis, because crypto markets never close. Accessed on September 19, 2026.
NIST/SEMATECH, e-Handbook of Statistical Methods, 1.3.5.6 Measures of Scale: sample standard deviation, divided by N − 1. Accessed on September 19, 2026.
SciPy, scipy.stats.percentileofscore: percentile rank, kind='mean' option: average of the strict and weak counts, so ties count as half. Accessed on September 19, 2026.
CoinDesk, Bitstamp Claims $5 Million Lost in Hot Wallet Hack (January 5, 2015): Bitstamp's suspension on January 5, 2015, after a hot wallet hack. Accessed on September 19, 2026.

An information tool, not investment advice: nothing on this page tells you when to buy or sell.