What is Puell Multiple? definition, context and sources

01What is Puell Multiple?

Puell Multiple is the ratio of the daily value of newly issued bitcoins in USD to its 365-day moving average. We use the issuance variant without transaction fees; a variant based on total miner revenue must be labeled and calculated separately. The numerator and denominator require the same methodology and a positive reference average.

02What is compared with the yearly average

Glassnode defines Puell Multiple as the dollar value of daily bitcoin issuance divided by its 365-day moving average and credits David Puell with the metric. Our use compares the same quantity in the numerator and denominator. The result is a dimensionless multiple, not a bitcoin price or an investor’s percentage return. [1]

03Issuance and revenue including fees are different variants

Coin Metrics distinguishes a Puell Multiple based on issuance from a variant based on miner revenue; revenue includes both new coins and fees. Our main explanation uses only the value of new issuance. A fee transfers existing coins and must not be presented as new issuance in this variant. When comparing charts, first check the metric identifier and what both parts of the ratio contain. [2] [3] [4]

04The average requires a complete, time-aligned window

Glassnode’s definition uses daily issuance and its moving average. Our reproducible example uses a simple average of completed daily values including the current model day. For an actual provider, check the day boundary, price source, inclusion of the current day and treatment of missing values. A missing day is not automatically zero issuance; we do not interpret this ratio with a non-positive denominator. [1]

05Dollar value combines quantity and price

Coin Metrics’ dollar-denominated issuance series values new units. In our simplified model, daily value is the product of the amount of new BTC and a single price for that day. A change in the ratio can therefore arise from changes in quantity, valuation or the reference average. Without further information, it cannot be attributed solely to price movement or miners’ decisions to sell. [2]

06Halving changes the issuance base even at the same price

Bitcoin Core’s GetBlockSubsidy function reduces the block subsidy according to the number of elapsed halving intervals. Our interpretation is that, with the same block count and price, a lower subsidy can reduce the value of new issuance while the moving average still includes the previous regime. The actual block count in a daily window must be measured; a calendar day does not contain a fixed number of blocks. A low multiple after such a change does not by itself prove a price bottom. [5]

07Aggregate revenue is not an individual miner’s net profit

Coin Metrics defines revenue as the value of new units and fees, not a result after deducting costs. Our interpretation therefore does not derive an operation’s margin, electricity price, debts, hardware efficiency or share of discovered blocks from Puell Multiple. Neither the issuance nor the revenue variant alone proves forced selling, immediate cash receipts or machines being switched off. [4]

08A historical test needs the data available at the time

Glassnode Point-in-Time distinguishes values known at a given moment from later updates. Our reproduction requires saving the provider, version, price inputs and availability time; an unfinished day cannot be treated as a final total known in advance. Point-in-Time availability must be checked for the specific series, and choosing such data alone does not fix timing errors elsewhere in the test. [6]

09An extreme value is neither an instruction nor a protocol threshold

Coin Metrics cautions that historical results of on-chain indicators do not guarantee future developments. Our interpretation therefore does not treat colored chart bands as consensus rules or certain moments to buy or sell. The multiple describes a relation to its own historical baseline; by itself, it determines neither fair value, cycle length nor the probability of a reversal. [7]

ExampleEXAMPLE · PUELL

The same history, different issuance or price on the final day

Our hypothetical model compares alternatives A, B and C for the same final day, not three consecutive days. In every alternative, each of the previous 364 days issued 1 BTC at a model price of 100 USD/BTC, giving daily issuance value of 100 USD. For the final day, new issuance quantities are [1, 0.5, 1] BTC and prices [100, 100, 200] USD/BTC. Daily issuance values are therefore [100, 50, 200] USD. We exclude fees and consider all daily observations complete; these are neither actual network data nor a model of a particular block subsidy. For each alternative, we add the previous daily values to the final day’s value and divide by 365 days. The sums are [36500, 36450, 36600] USD and the average daily values, rounded to six decimal places, [100, 99.863014, 100.273973] USD. Puell Multiple calculated from the unrounded averages is [1, 0.500686, 1.994536] after the same rounding. The chart shows only these dimensionless ratios for A, B and C; the horizontal line at 1 denotes equality with the respective average, not a profit threshold. In B, price is the same as in A but issuance is lower; in C, quantity is the same as in A but valuation is higher. Because the final day also enters the average, the results are not exactly 0.5 and 2. The example includes neither operating costs, a completed sale nor probabilities of future prices.

Check whether the series measures new issuance or revenue including fees.
Use consistent daily windows, price inputs and a documented data version.
A lower subsidy can reduce the multiple even without a price decline.
The ratio alone shows neither net profit, an actual sale nor the next price reversal.

Read in context

For the clearest picture, read this entry together with Block Subsidy, On-chain Analysis, Moving average, Hashprice, Halving.

?Practical questions

01Does a low Puell Multiple mean mining is loss-making?

No. The ratio compares the aggregate dollar value of issuance with its own history. It does not deduct a particular miner’s costs or establish their operating margin. A lower block subsidy can reduce the value even if bitcoin’s price is unchanged.

02Can the issuance variant be substituted for the total revenue variant?

Not without changing the definition. The revenue variant also includes fees and must use the corresponding revenue average. Mixing an issuance numerator with a revenue denominator creates a different quantity; a similar chart name alone is not enough.