The Crypto Fear and Greed Index: What It Measures and Its Real Limits

What the Fear and Greed Index is

The Crypto Fear and Greed Index, published by Alternative.me, is a single number from 0 to 100 intended to summarise market sentiment. Zero is labelled extreme fear and 100 extreme greed, with fear, neutral and greed bands in between.

Two facts about its scope are frequently missed and both matter.

It covers Bitcoin only. Alternative.me states on the index page that the current index is for Bitcoin only, with separate indices for large altcoins offered as a future addition. Treating the number as a reading on the whole crypto market, or on any particular altcoin, is reading something into it that is not there.

One of its six components is switched off. The published methodology lists surveys at a 15% weight and marks that component as currently paused. So the number you see is produced by five factors, not six, and the remaining weights carry proportionally more of the result.

The index is also published through a public API at api.alternative.me/fng/, which returns each reading as a value, a classification string, and a timestamp. That is worth knowing if you want to check the history yourself rather than rely on a screenshot.

How the index is computed

Alternative.me publishes the factor weights directly.

The circularity problem

This is the single most important thing to understand about the index, and the arithmetic makes it concrete.

Working it through

Start from the published weights and remove the paused component.

So roughly 71% of the active index is a transformation of Bitcoin's own volatility, volume and dominance, and roughly 29% is attention data.

Why this produces the behaviour people observe. Traders frequently note that the index tracks price closely, and interpret this as evidence that sentiment and price are tightly linked. The simpler explanation is that most of the index is price data. Volatility rises in declines, volume rises in declines, and Bitcoin dominance typically rises when risk appetite falls. All three push the index toward fear during a sell-off before any human being feels anything. The correlation is largely definitional.

What this actually tells you. When the index prints extreme fear during a decline, you have not received two independent pieces of evidence. You have received the price move, and then a partly rescaled version of the same price move with a sentiment label attached. Counting it as confirmation is the double-counting error described in how to read crypto market signals, where the fix is to count independent dimensions rather than indicators.

The actionable version. If you want a sentiment reading that is genuinely independent of price, you need the components that are not derived from market data: social volume and search interest. Those are available separately and they are the 29%. Using the composite instead means accepting that most of what you are looking at, you already knew from the chart.

What extreme readings actually correlate with

Extreme fear readings coincide with periods of high volatility, elevated volume, rising Bitcoin dominance, and falling attention or panicked attention. That is a genuine and reliable relationship, and it is not a discovery. It is what the formula computes.

Extreme greed readings coincide with the reverse: compressed volatility relative to recent history, elevated volume in an advance, falling dominance as capital rotates into altcoins, and rising search interest.

The tempting next step is to say that extreme fear readings have historically been followed by rallies. Two things need to be said about that claim.

First, it is unfalsifiable as usually stated. Over a sufficiently long horizon in an asset that has trended upward, almost any entry has been followed by a rally. A claim that extreme fear precedes gains needs a stated holding period, a stated benchmark, and a comparison against entering at random on the same schedule. Without those three, it is not a testable statement.

Second, we are not quoting a historical hit rate here, deliberately. Published backtests of index-based rules exist, but they depend on the index's own revision history, on which of the six components were active during the period tested, and on the threshold chosen for "extreme." The surveys component being paused means recent readings are not produced by the same formula as older ones. Quoting a percentage derived across that boundary would be false precision, so this article omits it rather than repeating a figure it cannot stand behind.

The contrarian reading and where it breaks

The contrarian idea is simple and not stupid: buy when the index shows extreme fear, reduce when it shows extreme greed. It fails on a specific, identifiable mechanism rather than on general scepticism.

Extreme readings persist. The index does not print extreme fear for one day and revert. Because every factor is scored against a 30-day and 90-day baseline, a sustained decline keeps volatility elevated, volume elevated, and dominance rising for as long as the decline lasts. Readings can sit in extreme fear for weeks.

That single fact dismantles the naive rule. A trader following "buy at extreme fear" buys on the first day of the extreme reading, and then again, and again, throughout a decline that may still have a long way to run. The rule has an entry condition and no exit condition, so its results are determined almost entirely by how long the downtrend lasts and how much capital the trader had left by the end of it. The index is not what determines the outcome.

The rule is a schedule, not a signal. Once you notice that it fires repeatedly through a decline, you can see what it actually is: a form of averaging in, triggered by market conditions rather than by the calendar. Whether that is a good idea is a separate question, examined in dollar-cost averaging in crypto, and it should be evaluated on those terms rather than credited to the index.

Reflexivity works against it. Any sentiment measure widely followed enough to be tradable is followed widely enough to change what it measures. When a large number of participants buy an extreme fear print, the buying itself alters volume, volatility and price, and therefore the next reading.

What the index cannot tell you

• It cannot tell you about altcoins. It is published for Bitcoin only. An altcoin's own sentiment can diverge sharply, particularly since the dominance component moves against altcoins by construction.
• It cannot lead price. Roughly 71% of its active weight is computed from market data that has already printed.
• It cannot tell you how long a state will last. Persistence is exactly what a contrarian entry needs to know, and it is exactly what the index does not contain.
• It cannot be compared across long periods without care. Component availability changes, and a paused component changes the effective formula.
• It cannot tell you about positioning in the derivatives market. For that you need funding, open interest and liquidation data, which are separate measurements with their own failure modes.

How to use it as context

For the underlying emotional cycle the index is attempting to proxy, crypto market psychology covers the behavioural pattern in detail, and why most crypto traders lose money covers the practical consequences of acting on it.

Conclusion

The Crypto Fear and Greed Index is a competently constructed composite that is widely misdescribed. It is not a measure of what the market feels. With the surveys component paused, roughly 71% of its active weight comes from Bitcoin's own volatility, volume and dominance, which means most of what it reports is a rescaled version of price and trading data the reader already has. The remaining 29%, social volume and search interest, is the part that carries information the chart does not.

That composition explains the behaviour everyone notices. The index tracks price closely because it is largely built from price. Treating a fear reading during a sell-off as independent confirmation of the sell-off is counting the same evidence twice, and it produces exactly the false confidence that composite indicators are supposed to prevent.

The contrarian rule fails on a specific mechanism rather than on general grounds. Because every component is scored against a 30-day and 90-day baseline, extreme readings persist for as long as the conditions that produced them persist. A rule that buys extreme fear therefore fires repeatedly through a decline, which makes it an averaging-in schedule with no exit condition rather than a timing signal. Its results are governed by the length of the downtrend, not by the index.

Two further limits deserve to be remembered. The index is published for Bitcoin only, so applying it to altcoins imports a dominance component that moves against them by construction. And it says nothing about leverage, which is where the most dangerous crowding in crypto actually sits. Used as a crowding gauge that informs how large to be, the index earns its place. Used as a trigger for when to buy, it is asking a coincident measurement to answer a question about the future.

Frequently asked questions

Sources and further reading

Primary sources:

• Alternative.me — Crypto Fear and Greed Index and methodology

Related CoinBeaver articles:

• How to read crypto market signals
• What Bitcoin dominance measures
• Crypto market psychology
• Dollar-cost averaging in crypto
• How funding rates work
• Open interest in crypto
• Why most crypto traders lose money

This article is educational and is not financial advice. No historical hit rate for extreme index readings is quoted here: any such figure depends on the holding period, the threshold used for "extreme," the benchmark, and which of the six components were active over the period tested, and the currently paused surveys component means recent readings are not produced by the same formula as older ones. Verify the live methodology on the publisher's own page before relying on any reading, since component weights and availability can change.