Paper: arXiv 2505.05113
Authors: Alex Nezlobin, Martin Tassy
Abstract
Although modern blockchains almost universally produce blocks at fixed intervals, existing models still lack an analytical formula for the loss-versus-rebalancing (LVR) incurred by Automated Market Makers (AMMs) liquidity providers in this setting. Leveraging tools from random walk theory, we derive the following closed-form approximation for the per block per unit of liquidity expected LVR under constant block time: [" \overline{"\mathrm{ARB"}}= \frac{",σ_b^{2"}} {",2+\sqrt{2π"},γ/(|ζ(1/2)|,σ_b),}+O!\bigl(e^{"-\mathrm{const"}\tfracγ{“σ_b”}}\bigr);\approx; \frac{“σ_b^{2”}}{",2 + 1.7164,γ/σ_b"}, “] where \(σ_b\) is the intra-block asset volatility, \(γ\) the AMM spread and \(ζ\) the Riemann Zeta function. Our large Monte Carlo simulations show that this formula is in fact quasi-exact across practical parameter ranges. Extending our analysis to arbitrary block-time distributions as well, we demonstrate both that–under every admissible inter-block law–the probability that a block carries an arbitrage trade converges to a universal limit, and that only constant block spacing attains the asymptotically minimal LVR. This shows that constant block intervals provide the best possible protection against arbitrage for liquidity providers.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper presents advanced theoretical derivations using random walk theory and special functions (Riemann Zeta), yielding complex closed-form formulas, but also validates them with extensive Monte Carlo simulations showing high accuracy, indicating strong empirical backing.
Research Flowchart
flowchart TD A["Research Goal<br>Derive analytical LVR for AMMs<br>with deterministic/variable block times"] --> B["Key Methodology<br>Random Walk Theory<br>Monte Carlo Simulations"] B --> C["Data/Inputs<br>Constant vs. Variable Block Time Laws<br>Intra-block Volatility σb<br>AMM Spread γ"] C --> D["Computational Process<br>Stochastic calculus &<br>Closed-form approximation derivation"] D --> E["Key Finding 1: Closed-Form LVR<br>`ARB ≈ σ_b^2 / (2 + 1.7164γ/σ_b)`<br>(Quasi-exact across parameters)"] D --> F["Key Finding 2: Block Time Optimization<br>Constant intervals attain asymptotically minimal LVR<br>(Universal limit proved)"] E & F --> G["Outcome<br>Constant block spacing offers<br>best protection for Liquidity Providers"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a discrete‑time, continuous‑state Markov chain model of an AMM that accommodates arbitrary block‑time distributions and derives closed‑form approximations for loss‑versus‑rebalancing (LVR), especially a tractable formula for constant (deterministic) block intervals.
Source sentence
We address this gap by developing a discrete-time, continuous-state Markov chain model of an AMM that supports arbitrary block-time distributions, with a focus on the deterministic block-time case. Evolving block-by-block, the model reflects AMM state dynamics and provides tractable expressions for LP losses.
Assumes 6
- Block times are i.i.d. draws from a base distribution with finite first and second moments, normalized to mean 1 and scaled to the average block time t. model
Source sentence
Let µ be a base distribution on [0, ∞) with finite first and second moments (we assume E[µ] = 1 for normalization). For a given average block-time t > 0, denote by µt the distribution of tX when X ∼µ (so that µt has mean t).
- A special case of constant block time is modeled as a Dirac delta at t (deterministic block spacing). model
Source sentence
One particular case of interest is the constant block-time model, where µt is a Dirac delta at t (i.e., each block has exactly length t).
- The asset price follows a driftless geometric Brownian motion with volatility σ, so log‑price changes over interval Δ are N(0, σ²Δ). market
Source sentence
We assume the asset price St follows a geometric Brownian motion with zero drift and volatility σ. In other words, over an interval of length ∆, the log-price change ln(St+∆/St) is N(0, σ2∆).
- The AMM spread γ is constant over time. model
Source sentence
LOSS-VERSUS-REBALANCING UNDER DETERMINISTIC AND GENERALIZED BLOCK-TIMES 5 The AMM is assumed to have a constant internal spread γ, and the liquidity provider supplies liquidity with constant density ℓ(USD per percentage point of price change).
Show 2 more
- Liquidity is supplied with constant density ℓ (USD per percentage‑point of price change). model
Source sentence
LOSS-VERSUS-REBALANCING UNDER DETERMINISTIC AND GENERALIZED BLOCK-TIMES 5 The AMM is assumed to have a constant internal spread γ, and the liquidity provider supplies liquidity with constant density ℓ(USD per percentage point of price change).
- The analysis restricts to the zero‑drift case; non‑zero drift is left for future work. model
Source sentence
(We restrict attention to the driftless case; incorporating a non-zero drift for St is left for future work.)
Data
- Large Monte‑Carlo simulation of AMM price dynamics synthetic unknown
Source sentence
Our large Monte Carlo simulations show that this formula is in fact quasi-exact across practical parameter ranges. Monte-Carlo experiments confirm that the resulting approximation is remarkably accurate, thereby extending the theoretical foundation of LVR to the contemporary blockchain systems.
Costs
Not applicable
Finds 5
- A closed‑form approximation for per‑block per‑unit‑liquidity expected LVR under constant block time is ̅ARB ≈ σ_b² / (2 + 1.7164 γ/σ_b), with an exponentially small error term. theoretical
Source sentence
Leveraging tools from random walk theory, we derive the following closed-form approximation for the per block per unit of liquidity expected LVR under constant block time: \[ \overline{\mathrm{ARB}}= \frac{\,σ_b^{2}} {\,2+\sqrt{2π}\,γ/(|ζ(1/2)|\,σ_b)\,}+O\!\bigl(e^{-\mathrm{const}\tfracγ{σ_b}}\bigr)\;\approx\; \frac{σ_b^{2}}{\,2 + 1.7164\,γ/σ_b}, \] where $σ_b$ is the intra-block asset volatility,
- Monte‑Carlo experiments confirm that the above formula is quasi‑exact across practical parameter ranges (error < 0.01%). statistical
Source sentence
Our large Monte Carlo simulations show that this formula is in fact quasi-exact across practical parameter ranges. This formula is practically exact for typical market parameters (yielding error < 0.01% in simulations).
- Constant (deterministic) block intervals uniquely minimize asymptotic LVR among all block‑time distributions with a given mean. theoretical
Source sentence
Extending our analysis to arbitrary block-time distributions as well, we demonstrate both that--under every admissible inter-block law--the probability that a block carries an arbitrage trade converges to a universal limit, and that only constant block spacing attains the asymptotically minimal LVR. As an important implication, we show that among all block-time distributions with a given mean, the constant
- The asymptotic decomposition ARB = Ptrade × LVR holds for any block‑time law, extending a result previously known for Poisson blocks. theoretical
Source sentence
Combining these two limits, we obtain the asymptotic decomposition ARB = Ptrade × LVR, (2.1) which was first observed by Milionis et al.[8] in the special case of Poisson-distributed block times.
- Moving from Poisson‑distributed to constant block times reduces per‑block LVR by up to 17.4% in the fast‑block regime. performance
Source sentence
Moreover, our results indicate that, all else equal, moving from a Poisson block-time to a constant block-time reduces the per-block LVR by up to 17.4% in the fast-block regime (see Figure 1a and Figure 1b). When ρb decreases, the percentage gain increases to-wards an asymptotic limit of ≈17.4%.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 19 related papers). Coverage is partial, so this is neither replication nor refutation.
5 papers in the archive : Priority Gas Auction Cadence and Searcher Competition: … ; Partially Active Automated Market Makers
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.