A measurement of Ethereum block position: swaps at the front pay 8–17x the tip

Victoria Shinder17 October 2026, 09:152 min read

An ethresear.ch post, "Measuring priority fees, positions and transaction classification of Ethereum blocks, July 2025 to October 2026" (13 October 2026, corrected 15 October after a base-fee decoding fix), tries to answer a concrete question behind block-building and MEV: what does the front of an Ethereum block actually pay, and who is there? This is a measurement study by one researcher, run on public data with an open-source tool; the figures are theirs and labelled as such, not an Ethereum Foundation result.

The tool and the sample. The author wrote TEMBRI (Transaction, Ethereum, Mempool, Block and Relay Intelligence), an open-source Rust tool (Apache-2.0), and ran it on public Xatu canonical execution data from ethpandaops. The sample is execution data only — about 115,000 blocks across sixteen sampled days: 1 July 2025, 1 July 2026, the week of 1–7 August 2026 and 3–9 October 2026. "Position" is defined as gas used up to and including a transaction divided by total block gas, so a transaction's place in the block is a 0–100% figure.

The headline finding. Swaps in the first 10% of block gas pay, on average, 8 to 17 times the priority fee of swaps between 30% and 60% of the block — on every day sampled. Most of that drop happens before the 20% mark. In other words, the premium for being near the top of the block is large and concentrated right at the front.

What held and what changed. Between 2025 and 2026, fee levels fell by about three-quarters, but the front-to-middle ratio did not fall — the premium for position persisted even as absolute tips collapsed. The composition at the front also shifted: about half the swaps in the first 5% of gas paid no priority fee in early August, falling to about one in eight in early October; the author attributes much of that fall in zero-tip front transactions to the builder Titan, at a nearly unchanged block share. Swaps that do pay a tip are about twice as frequent at the front as in the block overall, in both months, and builder labels show persistent differences in how priority fee correlates with position.

Why it matters, and its limits. Pricing position within the block is central to both out-of-protocol MEV (through MEV-Boost) and in-protocol fees (EIP-1559); a design needs to know what the front pays and who sits there, which is what this measures. The author is explicit about scope: it is execution data only, mempool/relay/direct-payment data are still in progress, and the fee numbers were recomputed after a decoding fix — so it is a reproducible snapshot of block-position economics, not a complete account of MEV payments. The tool and the October run's configuration are published for anyone to reproduce.

Source: ethresear.ch, "Measuring priority fees, positions and transaction classification of Ethereum blocks, July 2025 to October 2026" — https://ethresear.ch/t/measuring-priority-fees-positions-and-transaction-classification-of-ethereum-blocks-july-2025-to-october-2026/26156