Blockchain/Level II · Mechanics/2.2 Transactions/Lesson 3
Fees and fee rates (sat/vB)
The auction for block space and the metric that prices it
MaterialsAccess to mempool.space or a similar block explorerA calculator capable of division
Wonder
- Why does a transaction with a higher total fee sometimes get left behind by one with a lower total?
- How does a single byte of signature data cost less than a byte of address data in the same block?
- What determines whether a payment confirms in ten minutes or two weeks?
Explore
- Miners select transactions for their blocks by fee rate, not total fee. A block has finite capacity; maximizing revenue requires fitting the most value into that space. This is an auction where each transaction bids on its own density. The metric compares price per unit of volume, ensuring that a small, high-value bid outranks a large, low-density one when space is tight.
- Since SegWit activated in August 2017, block capacity is measured in weight units rather than bytes. The limit is 4,000,000 weight units per block. Non-witness data counts as four weight units per byte; witness data (signatures) counts as one. Witness data never enters the UTXO set that every node must keep, and can be pruned after validation, so it is charged less.
- Virtual bytes (vB) normalize this weighting: vsize equals total weight divided by four. Fee rates are therefore quoted in satoshis per virtual byte (sat/vB). A typical one-input, two-output native SegWit transaction is roughly 141 vB. The same payment as a fully legacy (P2PKH) transaction would be approximately 226 bytes, about 60% more block space for the same payment.
- Fee rates fluctuate with demand for block space. In December 2017 and again in spring 2021, congestion drove rates to historic highs. Around the April 2024 halving, the launch of the Runes token protocol caused sharp spikes as users rushed to mint new tokens. During quiet periods, rates fall toward the minimum relay threshold, allowing even small transactions to confirm quickly.
- Wallets estimate appropriate fee rates by analyzing mempool conditions and recent blocks. They project how long a transaction will wait at various price points. This estimation is probabilistic; there are no guarantees. Users may choose higher fees for faster confirmation or lower fees to save cost, accepting the risk of delay. The trade-off is between certainty and expense.
- Open mempool.space to observe current fee rates. Note the recommended rate for the next block versus one hour out. Select a recent transaction from the list and examine its virtual size (vsize) and total fee. Divide the fee by the vsize to verify the displayed sat/vB rate, confirming how the market prices that specific unit of data.
Key facts
- Miners evaluate transactions by dividing total fee by transaction weight to maximize revenue density.
- The block size limit is 4,000,000 weight units; non-witness bytes count as 4 units, witness bytes as 1.
- Virtual bytes (vB) are calculated as weight divided by 4, serving as the standard unit for fee rates.
- A typical native SegWit transaction is approximately 141 vB, while a legacy equivalent is roughly 226 bytes.
- Fee rates spike during periods of high demand, such as December 2017 and the April 2024 Runes launch.
- Wallets estimate fees based on mempool congestion and recent block inclusion times.
Exercise nine, six, three, one
9CollectRecord 9 recent Bitcoin transactions from a block explorer. For each, note the virtual size (vB), total fee in satoshis, and the resulting fee rate in sat/vB.
6ClusterGroup these 9 transactions into 6 categories based on their fee rate relative to the network median at that time: significantly higher, slightly higher, near median, slightly lower, significantly lower, or unconfirmed after 24 hours.
3DistillReduce these groups to 3 distinct user behaviors driving those fee choices (e.g., urgency, cost-optimization, error).
1CrystallizeWrite one sentence defining what a sat/vB rate actually purchases from the mining network.
Fact check
Dates, figures and claims checked against the published sources listed below.
- BIP141: Segregated Witness (consensus layer)
- A. M. Antonopoulos & D. A. Harding, Mastering Bitcoin, 3rd ed. (2023), ch. 9
- mempool.space documentation on fee estimation