Research
Time Is Money: Incentivized Causal Transaction Ordering
On blockchains, a user who publishes a profitable transaction is visible to attackers, who can pay validators to place their own transaction first and steal the reward.

- arXiv
- 2607.11496
- Published
- 2026-07-13
- Authors
- Hongyin Chen, Xu Zheng, Jichen Li, Ittay Eyal
AI summary
Overview
Research area: Blockchain transaction ordering, mechanism design, and game theory applied to Maximal Extractable Value (MEV) and front-running defense.
Technical level: Advanced. The paper relies on sequential game theory (Subgame-Perfect Nash Equilibrium), contest theory (Tullock contests), and mechanism design, and is written for readers comfortable with equilibrium analysis.
Scope (one sentence): The paper proposes PRECEDE, a bid-based randomized ordering rule with a super-linear power weight that makes front-running economically unprofitable, and proves that a "deterrence bid" by the first-mover forms an equilibrium supporting causal ordering, anti-spam, and user profitability.
What This Paper Is About
On blockchains, a user who publishes a profitable transaction is visible to attackers, who can pay validators to place their own transaction first and steal the reward. Existing defenses — arrival-time ordering, in-block randomization, and encryption of transaction contents — all fail to enforce causality: the rule that a transaction published in response to another must appear later on chain. The authors attack the problem from the incentive side rather than the information side, designing PRECEDE so that attacking simply does not pay.
Key Contributions
-
A general model of front-running as an economic contest. The authors formalize competition for a single on-chain reward R between a regular user and a set of attackers under a generic ordering protocol, with a partial-payment rule: the winner pays its full bid, and every loser pays a fraction γ (the losing-fee rate) of its bid.
-
Entry deterrence as a new ordering principle. Rather than ordering by time, by arrival, or by hidden content, PRECEDE orders by a power-weighted randomized lottery with weight w(b) = b^k for k > 1, so that the first mover can publish a bid that makes later entry unprofitable. The authors position this as a distinct principle from time-based and content-based approaches.
-
A closed-form deterrence bid. Lemma 1 and Theorem 1 establish the exact weight threshold W_dtr = ((k−1)^(k−1) / (γ k^k)) · R^k above which any attacker bid yields non-positive expected revenue, and the corresponding smallest deterring bid b_dtr = R · (((k−1)^(k−1)) / (γ k^k))^(1/k).
-
Anti-spam and equilibrium guarantees. The paper shows that a single bid dominates multi-transaction spamming if and only if k ≥ ln 2 / ln(1+γ), and that the deterrence strategy is a Subgame-Perfect Nash Equilibrium whenever k ≥ max{2, exp(1/γ)/γ}, with a revenue-maximizing choice of k = 1 + 1/γ.
Main Findings
-
Super-linear weight squeezes counter-bids from both sides. A small attacker bid has too little weight to win, while a large bid costs more than the reward it competes for; a bid strictly below R therefore suffices to deter entry, unlike deterministic descending-bid ordering where the user must bid the entire reward.
-
The ordering rule generalizes existing mechanisms. As k → ∞ the weight concentrates on the highest bid and PRECEDE degenerates into the descending-bid ordering used by practical blockchains; at k = 0 all weights are equal and it degenerates into uniform random ordering.
-
Exact deterrence threshold. The attacker's expected revenue with a bid b against observed total weight W is u(b; W) = (b^k (R − b) − γ b W) / (b^k + W), and entry is unprofitable for all b > 0 exactly when W ≥ ((k−1)^(k−1) / (γ k^k)) · R^k.
-
Anti-spam condition. Every rational participant publishes a single bid rather than splitting the same amount across multiple transactions if and only if k ≥ ln 2 / ln(1+γ).
-
Profitability condition. When k meets that same anti-spam bound, the deterrence strategy also yields the user strictly positive expected revenue, so she remains incentivized to discover opportunities.
-
Equilibrium condition. For k ≥ max{2, exp(1/γ)/γ}, it is a Subgame-Perfect Nash Equilibrium for the user to publish the deterrence bid in the first step and for neither player to act thereafter, achieving causal ordering regardless of game length. The authors state this bound is loose, and that their numerical experiments show the anti-spam bound alone already suffices; specific numerical values from those experiments are not reported in the provided content.
-
Revenue-maximizing parameter. For a fixed γ, the user's expected revenue is maximized at k = 1 + 1/γ, where it equals γR/(1+γ).
-
Practical implementation via scores. The pairwise weighted-ordering probability w(b(tx₁)) / (w(b(tx₁)) + w(b(tx₂))) is realized by drawing r(tx) = H(tx ‖ Q) from a random oracle against a beacon Q and sorting by the score s(tx) = r(tx)^(1/b^k(tx)). Because each score depends only on the transaction's own bid and Q, the relative order of transactions competing for one reward is independent of all other transactions, so analyzing one competition in isolation is without loss of generality.
-
Sandwich attacks are also defended. Under PRECEDE, defending against a sandwich attack is no harder than defending against front-running, and the user's deterrence cost is strictly less than the loss she would suffer from a sandwich attack in practical blockchains.
-
Context on how bad the status quo is. On Flashbots, arbitrage winners pay validators a median of over 90% of their reward, which the authors cite as evidence that bid-based ordering erodes user incentives.
Methodology in Plain English
The authors treat front-running as a game rather than a cryptography problem. Two kinds of players compete for a reward: a user who spots the opportunity first, and attackers who only learn about it once the user's transaction becomes visible. The blockchain is modeled as an idealized ordering service that collects all published transactions during a window [0, T] and then orders them using a random beacon.
Into this game they insert a tunable ordering rule: each transaction gets a weight equal to its bid raised to the power k, and its chance of being placed first is proportional to that weight. The first mover then has a lever — bid high enough that any follower's expected profit is at most zero. The authors derive that threshold in closed form using calculus on the attacker's revenue function, then check three things: that following the strategy beats splitting a bid across many transactions (anti-spam), that the user still comes out ahead (profitability), and that no player has a profitable deviation at any point in the sequential game with the attacker moving last (subgame perfection). The game-theoretic proof is backed with a numerical simulation to show how conservative the equilibrium bound is, though the specific simulation parameters and results are not reported in the provided content.
Why This Matters
Research impact. The paper reframes front-running prevention as an incentive-design problem instead of an information-hiding or timestamp-ordering problem, and introduces entry deterrence as a distinct ordering principle. It also extends contest theory beyond the all-pay, single-action assumption of prior Tullock-contest work by allowing partial payment on loss and repeated actions including spam.
Real-world applications:
- Decentralized exchange arbitrage: protecting the searchers who align prices across markets, whose rewards are currently competed away (median over 90% paid to validators on Flashbots).
- Sandwich-attack mitigation: shielding ordinary swaps from the front-run/back-run pattern that raises users' execution costs.
- Validator and block-builder ordering policies: PRECEDE is described as deployable in any censorship-resistant blockchain with a simple change to its transaction ordering mechanism.
- DeFi market health: keeping rewards with the participants who supply functional transactions, so those participants stay in the market.
Industry relevance. Because the mechanism requires only on-chain-observable signals (bids), an external random beacon, and a score-then-sort step, it fits the architecture of existing aggregated-collection protocols that already resist censorship and omission by single nodes.
Future Directions
- Tightening the equilibrium bound. The proven bound k ≥ max{2, exp(1/γ)/γ} is acknowledged as loose, and the numerical simulation suggests the anti-spam bound suffices; closing this gap analytically and reporting the experimental results in detail remain open.
- Calibrating k and γ in deployment. The losing-fee rate varies across opportunities by orders of magnitude, so choosing a single conservative k that works across the full range, and validating the assumed γ per transaction via local pre-execution, is a practical open problem.
- Multi-attacker dynamics. The deterrence analysis centers on a single entering attacker observing the user's transactions; behavior with many attackers acting in sequence or simultaneously is a natural extension.
- Empirical evaluation on a live chain. Moving from the idealized collection-and-ordering abstraction and the random beacon assumption to measurement on a deployed system, and checking interaction with application-level sequencing constraints, are untested steps.
Target Audience
Researchers and protocol designers working on MEV, transaction ordering, and blockchain mechanism design; graduate students in distributed systems or algorithmic game theory looking for a concrete application of sequential contest theory; and engineers at exchanges, rollups, or block builders who can change ordering rules and want a bid-compatible alternative to arrival-time ordering or encryption-based proposals.
Authors’ abstract
Front-running is a subtle and persistent problem for blockchains. A blockchain is a stateful virtual machine executing instructions called transactions. Users earn rewards by publishing functional transactions essential to the system. Attackers observe these transactions and publish their own ahead of the users', seizing the reward and eroding users' incentive to publish functional transactions. Preventing front-running means enforcing causality: If an attacker receives…
The full abstract on arXiv