Bitcoin Cycle
Bitcoin dashboard

Scroll on the chart to zoom, drag to pan, hover for the mood and the halving clock on any day. Green bands are the 18 months after each halving; the red band is where past bear markets bottomed, timed from this cycle's peak.

Loading sixteen years of prices…

Bitcoin, USDdaily close, UTC

The four-year cycle

Every cycle rebased to 1× on its halving day. Hover to read each cycle at the same point on the clock.

Peak: highest close within two years of the halving. Trough: lowest close within two years of that peak.

Market mood

    Each bar is today's reading ranked against the last three years (100 = the most risk-on reading in that time). The mood is their average.

    What bitcoin did next, by mood

    Median price change over the next 1, 3, 6 and 12 months, and how often it was higher. Spells counts separate visits to each band, the real sample size. Years higher averages the 12-month result within each calendar year first, so one long boom or bust can't dominate.

    The same bands, split by where the halving clock was median 12-month return

    Does any of it trade?

    Growth of $1 from the first halving, log scale. Signal at the close, trade the next day, no costs, cash earns nothing. Click a rule to hide it.

    The cycle held, four times out of fourPeaks came 371–546 days after the halving and every completed bear market fell 77–85%. Sitting out from month 18 to the next run-up beat buy-and-hold while holding coins less than half the time.
    Fear is a buy, greed is not a sellExtreme fear was followed by a higher price a year later in every year it showed up. Extreme greed came before the biggest gains: in bitcoin, greed is momentum. Selling it was the worst rule here.
    Four cycles is not a sampleThe multiples shrink every cycle (92×, 30×, 8×, 2×) and the theory is public, so it can be front-run. Timings set from the first two cycles missed the next two peaks by about three months.

    Trend

    How the mood is built, and how far to trust it

    Five inputs, each taken at the daily close and using only data known that day: price against its 200-day average; MVRV, the market cap over the realised cap (every coin valued at the price it last moved); active addresses, 30-day average against the one-year average; distance below the 52-week high; and 30-day realised volatility, inverted. Each is ranked against its own last three years, and the five ranks are averaged into a 0–100 score.

    Same build as the STI dashboard, with two swaps: the STI's breadth and new-highs inputs need 30 member stocks, which bitcoin doesn't have, so on-chain valuation and network activity take their places. Every window is in calendar days because bitcoin never closes.

    Limits. Four halvings, five spells per extreme band, and an asset that went up 900,000× over the sample, so almost any long rule looks good. Treat the tables as base rates, not a model. Not investment advice. Full methodology and tests.

    Key events

    Data: Coin Metrics community API (price, MVRV, active addresses), · Events curated from public reporting · Not investment advice.