Dynamics of the Liquidity Profile in Concentrated Liquidity Provision: An SPIDE Approach

summary

The gist

The paper develops a stochastic partial integro-differential equation (SPIDE) framework to model the dynamics of liquidity profiles in concentrated liquidity automated market makers like Uniswap V3…

In short

The paper uses a stochastic partial integro-differential equation (SPIDE) to mathematically model how liquidity profiles change in concentrated liquidity automated market makers like Uniswap V3. It shows that this profile evolves through mean-reverting concentration and nonlocal redistribution, transforming the complex dynamics into a tractable system of linear stochastic differential equations for analysis.

Key concepts

Liquidity Profile
This is defined as the total aggregation of all liquidity positions across different price ticks in an automated market maker. It represents how capital is distributed across the entire price spectrum and changes based on market conditions and provider actions.
SPIDE Framework
A stochastic partial integro-differential equation is a mathematical tool used to describe continuous evolution. Here, it captures two key features of decentralized markets: nonlocal interactions (redistribution) and continuous change, allowing the model to simulate how liquidity flows dynamically.
Gegenbauer Polynomials
These are special mathematical functions used in the analysis. The paper uses them to express the eigenmodes of the system, which helps transform an infinite-dimensional problem into a finite, coupled system of linear stochastic differential equations that can be solved practically.

Terminology used across episodes

This episode discusses

The paper

Dynamics of the Liquidity Profile in Concentrated Liquidity Provision: An SPIDE Approach · Read on arXiv

CHEUK YIN LEE, JIMMY RISK, SHEN-NING TUNG, TAI-HO WANG

National Cheng Kung University · Cal Poly Pomona · National Tsing Hua University · City University of New York

Transcript

Introduction to the show: ident: Quantitative Finance. The day in quantitative finance: models of markets, risk and prices.

Ruth: I'm Ruth.

Warren: And I'm Warren, here with Iris.

Ruth: We're looking at the paper: "Dynamics of the Liquidity Profile in Concentrated Liquidity Provision".

Warren: Here's the gist of the paper. The paper develops a stochastic partial integro-differential equation (SPIDE) framework to model the dynamics of liquidity profiles in concentrated liquidity automated market makers like Uniswap V3, which is important because it provides a mathematical tool to understand how capital is distributed and how this distribution evolves under market conditions Page 1</ref:2610.30539).

Ruth: Modeling the Liquidity Profile Dynamics The core of the paper is modeling the liquidity profile, defined as "the aggregation of all overlapping LP positions across tick space" Page 1</ref:2610.30539).

Warren: This profile evolves under a degenerate Sturm–Liouville operator that encodes mean-reverting concentration around the market price, coupled with a Hilbert–Schmidt integral operator that captures nonlocal redistribution driven by provider repositioning Page 1</ref:2610.30539).

Ruth: The SPIDE framework is uniquely suited to capture two critical features of decentralized markets: "Nonlocal Interactions" and "Continuous Evolution" Page 2</ref:2610.30539).

Title and authors: Ruth: So, we're looking at a paper called "Dynamics of the Liquidity Profile in Concentrated Liquidity Provision: An SPIDE Approach". The title itself tells you immediately that they're not just looking at the price movement, but how the actual distribution of liquidity across those price ticks changes.

Warren: Yeah, it sounds like they’re trying to get past just tracking the asset price and actually figure out where all the capital is sitting in those concentrated pools.

Ruth: Exactly. They introduce this stochastic partial integro-differential equation, or SPIDE framework, as a tool to model that evolution of the liquidity profile itself.

Iris: It seems like they are modeling something continuous, which is important because real market data isn't just discrete points; it’s a fluid density that shifts.

Ruth: Right. They say the profile is defined as the aggregation of all overlapping liquidity positions across tick space, which means it's this big picture view of where everyone has placed their capital across the whole price range.

Warren: And they use this SPIDE to capture two main things: nonlocal interactions and continuous evolution, which are definitely key features when you think about how a decentralized market actually works.

Iris: That sounds like a very ambitious way to model it, moving beyond simpler local diffusion models that just look at immediate neighbors.

Ruth: They do that by coupling a degenerate Sturm–Liouville operator, which handles the mean-reverting concentration around the price, with a Hilbert–Schmidt integral operator for those nonlocal redistributions driven by providers repositioning.

Warren: That Hilbert–Schmidt part is interesting because it’s explicitly capturing how liquidity moves around based on provider decisions, not just random noise.

Iris: And then they use orthonormal eigenfunction expansions to take this infinite-dimensional SPIDE and turn it into a much more manageable system of linear stochastic differential equations.

Ruth: That transformation is crucial because it takes something mathematically huge and makes it solvable through standard SDE techniques, which is where the tractability comes from.

Warren: So what does that mean practically? It means they’re not just getting a theoretical equation; they’re creating a system you can actually run simulations on for things like predicting pool reserves.

Iris: But I wonder how robust this is when we look at the math itself, especially with that infinite-dimensional setup.

The paper's summary: Ruth: Let's break down what the paper says about how this SPIDE framework actually works in practice. They set up the evolution of a random function ut(x) governed by this SPIDE, and they establish that there’s a proof for the existence and uniqueness of the mild solution in a weighted L two space built on the Gegenbauer eigensystem <ref:2610.30539#pg1,existence and uniqueness of the mild solution in a weighted L 2>.

Warren: So, they’ve handled the theory part—they know this system actually has a unique mild solution under certain conditions, which is a big step for any complex stochastic process.

Iris: I'm looking at how they reduce the SPIDE to a coupled system of linear SDEs, and they say that the low-order modes of these equations carry the same level, slope, and curvature interpretations used in their empirical Legendre diagnostics.

Ruth: That’s the link between theory and observation; it means the mathematical modes directly map onto what we see empirically as clustering around a price point.

Warren: And they specifically mention that the Gegenbauer order alpha can vary in both the theoretical framework and when they perform calibration, which gives them flexibility.

Iris: That variation is important because it suggests that different values of alpha might correspond to different ways liquidity clusters or redistributes nonlocally, so it’s a parameter we have to figure out.

Ruth: They also mention that for p=two the mild solution is unique among all solutions satisfying some specific condition, which helps narrow down the possibilities when you're trying to identify the right model parameters <ref:2610.30539#pg1>.

Warren: So, they are essentially building a mathematical structure that links price dynamics, liquidity concentration, and provider repositioning into one coherent picture.

Iris: It’s like they’ve taken a messy real-world observation and built a rigorous mathematical machine to predict how that observation will evolve over time.

The paper's improvements: Ruth: The authors point out several things they think the framework can do better or what future work should look like. They suggest revisiting the noise specification to explore additional noise modes near the center, which could add more detail to how liquidity behaves right at the price point.

Warren: And they also mention exploring the BS/rawliquidity branch in section four point two.two which seems like a different way to look at how providers might be acting—maybe focusing on raw liquidity instead of just the final positions <ref:2610.30539#pg1>.

Iris: They also suggest addressing structural drift through price-aware coordinates and time-varying parameters, which is really important because markets aren't static; they evolve over time and context.

Ruth: That structural drift idea is what I think is most critical for real-world application; if we don't account for how the underlying market structure itself shifts, our models will always be slightly off.

Warren: Price-aware coordinates and time-varying parameters sound like the most direct candidates to fix that drift issue they've identified, which would give us a way to track that deterioration you see in real data.

Iris: And one limitation they explicitly flag is that weighted-space spectral estimators often rely on spatio-temporal infill to pre-estimate something, which means their current estimation methods have some assumptions baked in.

Ruth: So the improvement isn't just about making the equation cleaner; it’s about acknowledging where the mathematical model itself stops working and pointing toward what needs to be added next.

Conclusion: Ruth: To wrap up, this paper on "Dynamics of the Liquidity Profile in Concentrated Liquidity Provision: An SPIDE Approach" shows that you can use a single calibration to jointly estimate the concentration rate alpha, the concentration order alpha, and that relocation block VK.

Warren: That’s a big win because it gives you three key pieces of information from just one calibration run, rather than trying to pin down each parameter separately.

Iris: I think the implication is that this SPIDE framework provides a unified way to look at these dynamics, connecting the theoretical structure directly to the empirical diagnostics.

Ruth: And for me, it means we have a disciplined calibration method for understanding how concentrated liquidity works in these AMMs. It’s modular because you can choose different noise specifications and coordinate systems conditionally.

Warren: But I think the authors are honest when they say that their framework is a disciplined calibration of one branch, and the broader framework still has to address those missing pieces like price-aware coordinates and time-varying parameters.

Iris: That makes sense; it’s a strong model for what it can do, but the limitations point directly at where you need to go next to capture the full complexity of these markets.

Ruth: Exactly. This SPIDE approach gives us a powerful tool to study liquidity dynamics in concentrated liquidity AMMs, and we have this framework that allows practitioners to choose different noise types and coordinate systems conditionally.

More episodes

← Home