TimeBoost: Do Ahead-of-Time Auctions Work?

Akaki Mamageishvili, Christoph Schlegel, Ko Sunghun, Jinsuk Park, Ali Taslimi · 2025

Read TimeBoost: Do Ahead-of-Time Auctions Work? by Akaki Mamageishvili, Christoph Schlegel, Ko Sunghun, Jinsuk Park, Ali Taslimi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-18328.txt.

Introduction

Akaki Mamageishvili, Christoph Schlegel, Ko Sunghun, Jinsuk Park, and Ali Taslimi test whether TimeBoost, an ahead-of-time auction for fast-lane access in blockchain trading, actually works as intended. Traders bid for a time edge, and the question is whether those bids predict the value the winner will extract. The team compares winning bids against the later markout, a measure of value gained over the auction interval, for fast-lane trades.

The link turns out weak: bids are a noisy guide to extracted value for single trades. Using the price actually paid, the second-highest bid, improves the match slightly, hinting at a common-value auction. Over long spans, though, bids and gains line up much better, so bidders read trends well even when their timing is off. Frequent bidders keep their ranking and profits throughout.

Catalog record

CREATORAkaki Mamageishvili, Christoph Schlegel, Ko Sunghun, Jinsuk Park, Ali Taslimi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-18328

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

TimeBoost: Do Ahead-of-Time Auctions Work? Akaki Mamageishvili, Christoph Schlegel, Ko Sunghun, Jinsuk Park, Ali Taslimi Abstract We study the performance of the TimeBoost auction, by comparing cumulative fixed time markout of fast lane trades over the TimeBoost interval to bids for the fast lane. Such comparison allows us to assess how well bids predict future extracted value from the time advantage. The correlation between winning bids and markouts is weak across bidders, suggesting that bids are a noisy predictor of extracted value. The correlation slightly improves when comparing paid bids (the second highest bid) instead of winning bids to markouts, which we attribute to the fact that the auction is more of a common value type. In all settings, the relative order of the most frequent bidder performance remains the same, together with their absolute profits. Bids and markouts aggregated over long time intervals exhibit much higher correlation, indicating that bidders detect trends much better than identify when the high arbitrage value is exactly available. One possible explanation for this is the fact that the correlation between previous minute markouts and current minute bids is significant, suggesting that the previous minute markouts is used to predict the next minute value when bidding. 1 Introduction Many blockchains use auctions as part of their transaction ordering mechanism. In Ethereum, for example, Proposer Builder Separation (PBS) uses a builder auction where specialized parties, so called block builders , bid for the right to determine the content of the next proposed block. In most rollups, transactions are sorted in decreasing priority fees per gas. This is commonly called Priority Ordering or Priority Gas Auction , referring to the fact that it is an auction for earlier execution. Besides significant differences --in Ethereum builders bid on the right to order the whole block whereas in rollups the auction is on the relative ordering of individual transactions by the sequencer, Ethereum has a public mempool, rollups do not-- these auction formats usually have one commonality: the auction is ’’Just-in-Time” (JIT) in the sense that bidders can bid in it just until immediately before the content of the block is determined. As an alternative, ’’Ahead-of-Time” (AOT) auctions would have a sizable time gap between the auction and the determination of the order of transactions. In the context of Ethereum L1, these auctions have been discussed under the term ’’slot auctions”: in a slot auction, the builder only bids on the right to propose the next block without committing to any specific block content at the time of bidding and determines this content only later (usually 12 seconds later) 1 1 1 See [ 5 ] . . In rollups, TimeBoost implemented by Arbitrum 2 2 2 https://docs.arbitrum.io/how-arbitrum-works/timeboost/gentle-introduction is an instance of a transaction ordering policy that has an Ahead-of-Time component as part of its design. To compare the relative performance of AOT and JIT auctions, it is important to assess how well bids match the value that bidders get from the right they acquire. In JIT auctions, this is just a question of competition. As long as enough bidders compete in the auction the paid bids should be close to the value that the winning bidders get. For AOT auctions, besides competition, there is also a problem of prediction. Bidders need to predict the future value that they will get. If they are able to predict it well (and there is sufficient competition), the auction will yield efficient outcomes. If they do not, the auction will generally be more inefficient than a JIT auction. In this paper, we provide the first empirical analysis of an AOT auction in the context of blockchain transaction ordering, using Arbitrum’s TimeBoost auction as a case study. Related Work In Ethereum, the next block proposer sorts transactions in the block unilaterally. This allows the proposer to extract the value that comes from the power of ordering transactions. Such value is usually referred as Maximum Extractable Value, or MEV for short [ 2 ] . Transaction ordering through auctions has been theoretically analyzed in [ 12 ] . An early version of TimeBoost (without an Ahead-of-Time auction) has been proposed in [ 6 ] . Evaluating time advantage of arbitrage between centralized and decentralized exchanges (CEX-DEX arbitrage) under a First-Come-First-Serve (FCFS) ordering policy has been studied [ 3 ] , comparing it to PGA auctions and pure FCFS. The latency advantage in a PBS setting has been studied in a series of papers [ 10 ] , [ 13 ] . Possible downsides of AOT auctions have been identified in [ 11 ] , in case that there is a working secondary market at which the AOT acquired right can be resold (Just-in-Time). Methodologically, we rely on the following sources: Markouts at centralized exchange mid prices have been used in many studies to estimate CEX-DEX arbitrage profits on Ethereum, e.g., in [ 14 ] . Theoretical predictions of CEX-DEX arbitrage profits have been made in [ 9 ] . Our theoretical background for auctions with common value component comes from the classical work of [ 8 ] . TimeBoost TimeBoost is a relatively new transaction ordering policy deployed on Arbitrum rollup since April 17th, 2025. In this policy, a protocol sells a time advantage in the FCFS policy to one party in a permissionless manner, via a second-price auction. Bids are denominated in the Ethereum native asset, ETH. There is a reserve price of 0.001 ETH. If no bidder bids more than the reserve price, transaction ordering policy remains pure FCFS. If only one bidder bids more than the reserve price, it pays the reserve price. If at least two bidders bid more than the reserve price, then the winner is the highest bidder and it pays the second highest bid. The time advantage equals 200ms and it lasts 1 minute for the winner. The auction for the next minute starts at second 0 of each minute and lasts until second 45. This introduces a 15 second time window for the bidders for predicting next minute’s value. The reason for having a 15 second time window is to allow enough time for the auction settlement on-chain. The rule is enforced by a designated party, the sequencer, which executes regular transactions with a 200ms delay, while transactions coming from the TimeBoost winner are executed on their arrival, without delay. In effect, the sequencer merges these two FCFS queues into each other. The advantage, sometimes referred to as fast lane , the TimeBoost auction winner is given allows the party to react to external information faster than other users and e.g. to take advantage of CEX-DEX arbitrage. TimeBoost also allows the party to react to large price changes on the chain faster than others. Empirically, most of the extracted values and most activity in the fast lane can be attributed to the CEX-DEX arbitrage [ 7 ] . In particular, Wintermute exclusively extracts CEX-DEX arbitrage, while Selini sometimes uses fast lane for other types of arbitrage. Empirical Approach For each winning bidder, we keep track of all transactions they submit through a fast lane at the on-chain exchanges. We calculate the hypothetical profit that the party would make if they would sell assets they bought

Read the complete text