Methodology

A protocol collects fees: a lending protocol earns interest, an exchange earns commissions. We compare its market value with the part of those fees it keeps: its revenue.

Fair value = annual revenue × usual multiple ÷ circulating tokens

1. Smoothed annual revenue. We take the median of three measures: the last 12 months, the last 6 months × 2 and the last 3 months × 4. Exceptional days (more than 5 times normal) are capped, so a single day never sets the value.
2. Usual multiple. The protocol’s median price-to-revenue ratio over its last 3 years (extremes removed). With a short history, it is completed with its sector’s median, all the more so as the history is short. It stays between 3 and 150.
3. Range. The dotted lines on the chart use the low (25th percentile) and high (75th percentile) multiples of its history: a fair value is a zone, not a figure to the cent.

The gap compares the price with this fair value. It gives the temperature:

• Freezing: more than 40% below the estimated fair value
• Cold: 15% to 40% below
• Mild: within 15% of the fair value
• Warm: 15% to 40% above
• Hot: more than 40% above

No fair value is shown when the multiple has changed regime for more than 6 months (a new business model, for example): its past no longer serves as a reference.

Each estimate carries a confidence grade, which says how solid the data behind it is:

• History: how many years of revenue and prices we know.
• Size: revenue of a few thousand dollars a month moves too much to serve as a basis.
• Regularity: revenue concentrated on a few days makes the estimate fragile.
• Liquidity: a thinly traded token has a less representative price.

A: solid estimate. B: reasonable estimate, worth cross-checking. C: fragile estimate, shown in grey.

Each protocol or blockchain page compares its indicators with the median of its sector: price-to-revenue, revenue growth over one year, net yield for holders (revenue paid out minus dilution), dilution, volatility. The rank shows its place among comparable cryptos (1st = the best on that criterion).

Bitcoin, blockchains, memecoins and projects whose revenue is too small or too recent have no usable revenue. For them, there is no revenue-based fair value but a trend value, drawn from price history. That way, every crypto has a gap in numbers that reads the same way.

Trend value = average price over the past year

• 365-day geometric mean: it counts a 100% rise and a 50% fall the same way, so an isolated spike does not inflate the value.
• Bitcoin, Litecoin, Bitcoin Cash: we take the average of the last 200 weeks (close to four years, a full cycle), corrected by the usual gap between the price and that average over the last six years.
• Range: the zone where the price spent half the time around its average (25th to 75th percentiles).
• Confidence: B at best. A price trend is not a measure of intrinsic value; a large gap says the price has moved away from its trend, not that it must return to it.

Temperatures follow the same thresholds as for fair value. For blockchains, the market cap-to-fees ratio is still shown, to compare networks with one another.

A low price is not enough: a crypto can be “freezing” simply because the hype has faded. The quality grade describes how solid the project is, regardless of its price, out of 100 points:

• Real revenue (35 points): revenue (or a blockchain’s fees) over the last 12 months, on a logarithmic scale. For Bitcoin and stores of value, adoption replaces revenue.
• Supply and dilution (25 points): share of tokens already in circulation; fewer points if many tokens remain to be issued or if supply grows by more than 10% a year.
• Maturity (15 points): length of the price history; fewer points after a fall of more than 90% from the peak.
• Liquidity and size (15 points): volume traded relative to market cap, and rank. A volume above the market cap signals very speculative trading.
• Price stability (10 points): volatility, worst drawdown and liquidity compared with other cryptos.

A: 75 and above · B: 60–74 · C: 45–59 · D: 30–44 · E: below 30. Two fixed rules: memecoins, political tokens and fan tokens (CoinGecko categories) are capped at E and hidden by default in the ranking and the map; a crypto without measurable revenue cannot rank above B.

On the chart, the scenario (shown over 4 years by default, 1 to 5 years to choose from) extends every line except the price: fair value and its range, trend value, 50- and 200-day moving averages. It is illustrative, not a forecast.

• Underlying trend: a regression of the price (in logarithm) over 8 years at most, once the cycle is removed, with a slope capped between −25% and +25% a year that fades over time.
• 4-year cycle: Bitcoin’s average shape between two halvings (2016 and 2020 cycles): a surge 12 to 18 months after the halving, a fall of 70 to 80% afterwards, a low about two and a half years later. The amplitude is measured on the current cycle; for the next one, it sits between the current cycle’s and the usual reduction seen from one cycle to the next.
• Sensitivity of each crypto: measured on its own past cycles when it has at least 3 years of history (a crypto that went × 20 then −90% amplifies the cycle), otherwise on its moves over the last two years compared with Bitcoin; capped between 0.5 and 2.
• Current gap: half of the gap between the price and “trend + cycle” fades in about a year and a half; the other half is treated as lasting.
• Fair value: revenue that follows its own underlying trend by default (a regression of monthly revenue over the last 4 years, one cycle, once the cycle is removed, capped between −25% and +25% a year and fading), or the pace you choose (stable, last 3 months, one year); it also follows the cycle (volumes rise with prices, at 80% of the price’s sensitivity); usual multiple, observed dilution.
• Cryptos without revenue (Bitcoin, most blockchains, tokens without revenue): the scenario rests only on the price’s underlying trend and the cycle; there is no revenue to project.

The simulator calculates what €10,000 would have become between a point on the price and a point on another line, excluding fees and currency effects. With a point in the projection, the result is a hypothetical simulation.

No estimate for stablecoins (their price is pegged at $1), assets backed by something external (gold, money-market funds, tokenized dollars) and cryptos whose price history is shorter than 18 months. Making up a number would be easier, but it would be wrong.

A revenue-based fair value requires at least $1M of revenue over a year, measured for more than 6 months, and a market cap of $20M. Below that, the trend value applies.

Prices, market caps and token supply: CoinGecko, refreshed every 30 minutes for the top 1,000 cryptos. Revenue, fees and holder share: DefiLlama, refreshed daily. Long price history: DefiLlama.

• Past revenue does not guarantee future revenue.
• A crypto can stay “cold” for a long time, or get even colder.
• The market can be right to pay more: a fair value based on the past cannot see strong growth ahead.
• History grows every day: for recent projects, the usual multiple relies mostly on their sector.
• DefiLlama may correct its data; our grades follow.

Dibagger is an independent project created by Cédric. The idea comes from the tools stock investors have used for a long time: compare a company’s price with what it earns, on a chart you can read in a few seconds. Dibagger applies that reading to cryptos that earn real revenue, and places the others within their price trend.

The site lives on its Pro subscriptions: no advertising on the site, no paid link to an exchange. A question, a mistake spotted: cedric@dibagger.com.