Oscillator Risk, explained

History used

All

every day the measure has ever had, nothing dropped

Low score

≤ 25%

blue guide line, blue background once the history is long enough

High score

≥ 90%

turns red only after two closes in a row

Measures

8

ratios and dominance charts, chosen by us

A rank, not a formula in the usual sense. For every bar on the chart, it takes every earlier value plus the current one and counts what fraction of them sits below the current value. Matching values count as half. That fraction, scaled to 0–100, is the score.

If 86% of the measure’s history closed below today’s value, the reading is 86%. There is no smoothing, no extra weight on recent readings, and no separate treatment of up days and down days the way RSI has. It is a straight count.

All of it. The history expands as the measure grows — nothing rolls off the back. This is not a rolling window of the last so-many bars: today is ranked against every reading the chosen measure has ever had, at whichever timeframe you are viewing.

Two rules decide when that history is long enough to trust. A reading counts as reliable once there are at least 30 bars covering at least one calendar year. Before that, the line is drawn faint and no coloured background is shaded — early readings are shown, but marked as not yet comparable to later ones.

For each bar, every value seen so far is lined up in order. Count how many sit below the current value, add half of any that match it exactly, divide by the total number of values (the current one included), and multiply by 100.

Two things fall out of that which are easy to miss. First, because the current bar is always in the total but never counts as "below itself," a brand-new all-time high scores (n − ½) / n — just under 100, never 100 exactly. Second, the bar still in progress updates live, but the two-in-a-row confirmation that turns 90+ red only counts finished bars, so a spike part-way through a day cannot confirm a high reading on its own.

The chart draws two guide lines. A reading at or below 25% puts the measure in the low band — historically low compared with itself — and, once the history is long enough, shades the background blue. A reading at or above 90% is high, but the line and background only turn red after two finished bars in a row at or above 90%. A single bar above 90% stays orange.

Both bands are entirely relative to the measure’s own history. Nothing here compares one measure to another, and nothing here refers to an absolute level. A dominance percentage and a market-cap ratio are scored on exactly the same 0–100 scale, because that scale only ever measures a value against its own past.

Oscillator Risk doesn’t try to call tops or bottoms the way the Pi Cycle Top does, so there is no hit rate to quote. It is a measure of position, and its limits come from what a rank can and cannot tell you.

Eight measures of market structure. Three of them are ratios and totals: Altcoins vs Bitcoin (total crypto market cap excluding BTC, ETH, stablecoins and tokenized RWAs, divided by Bitcoin’s market cap), Ethereum vs Bitcoin (market cap against market cap, not the ETH/BTC price pair) and Others Market Cap (coins ranked 11 through 125). And five dominance measures: Bitcoin, Ethereum, Bitcoin excluding stablecoins, USDT + USDC, and BTC + ETH + stablecoins combined. Individual coins are not offered — this is about how money moves between parts of the market, not about one coin’s price.

The dominance measures, Ethereum vs Bitcoin and Altcoins vs Bitcoin are all built from one unbroken CoinGecko market-cap history, so the line never switches source part-way along. Altcoins vs Bitcoin is calculated and stored every day from the definition above. Others Market Cap comes from a maintained daily history instead, refreshed by hand, because which coins sat in the top 125 on any past day cannot be worked out from today’s list. The chart opens on Altcoins vs Bitcoin, weekly.

It never asks whether a measure is high in absolute terms — only whether it’s high compared with its own entire past.

Watch the band, not the exact number. The gap between, say, 82% and 88% matters far less than the fact that neither has crossed 90 — and a single bar above 90 matters less than two finished ones. The score is a rank, and ranks squash the detail together near the extremes.

Change the timeframe deliberately. Switching from weekly to daily doesn’t just zoom the chart — it changes which closing values are being ranked, and can move the score even though the underlying data hasn’t changed. Faint early readings mean there isn’t yet enough history at that timeframe.

Oscillator Risk is a position indicator, not a timing one. It says where a market measure sits inside its own history — nothing about direction, momentum, or what happens next. That flexibility is the point: the same 0–100 scale, the same two levels, applied to a ratio, a total, or a dominance chart without tuning it for each one.

It is not one of the weighted inputs behind the Bitcoin or Ethereum Heatmeter — it is its own chart. If you want a single number that blends timing indicators like Pi Cycle Top with valuation and on-chain measures, the heatmeters do that. Oscillator Risk is for asking one narrower question about one measure at a time: compared with its own entire past, how stretched is this right now?

Others in price & cycle models — What price is doing relative to its own history, trend and long-run models.

• BDZOur own measure of how far Bitcoin’s price has stretched from its recent average, in either direction.
• RSI Cross SignalA popular momentum gauge crossing its own average — a sign the direction is turning, not just that price has run hot.
• RSI HeatmapMomentum across many coins and short time windows at once, plus a 0-100 risk reading for whichever coin you pick.
• Fear & Greed IndexThe widely published 0–100 sentiment gauge, tracked alongside price so you can see when emotion led or lagged.
• Everything vs BTCAltcoin risk read from Bitcoin’s share of the market — when money moves out of Bitcoin and into everything else.

Every indicator we track — the full glossary.