ReClamm Glide Mechanics – Explained Step by Step

Glide (gradual recentering): goal, variables, centeredness, update rule, typical values, and engineering trade-offs

Overview

ReClamm (Re-Centering Liquidity AMM) introduces a mechanism called glide (also referred to as gradual / time-proportional recentering) whose goal is:

Automatically and smoothly move the virtual liquidity center of the pool toward a target price without creating large, sudden arbitrage opportunities and without requiring active LP management.

This is one of the biggest differences compared to classic Uniswap v3 concentrated liquidity (which is static until the LP manually moves ranges) and classic constant-product AMMs (which have no concept of preferred center at all).

---

Core idea in one sentence

The pool slowly “glides” its virtual balances toward a target price center using an exponential decay function — the farther the current price is from the target, the faster it wants to move back, but the movement is always gradual and time-proportional.

---

Key concepts and variables

Symbol	Meaning	Typical default / range	Controls what?
T	Target price (the “desired” center of the pool)	oracle price, TWAP, governance vote	where the pool wants to be
Vₐ, Vₓ	Current virtual balances of token A and token B	dynamic	virtual liquidity depth on each side
Rₐ, Rₓ	Real (actual deposited) balances of token A and token B	real LP deposits	real economic exposure
c	Centeredness metric (how balanced the virtual liquidity is)	0 … 1 (ideal = 1)	trigger condition for glide
m	Minimum centeredness margin	0.75 – 0.85	when glide is activated
τ	Daily shift exponent / decay rate	0.3 – 0.7	how aggressive the recentering is per day
Δt	Time elapsed since last glide (in days)	fraction of day	controls speed of movement

---

Centeredness formula (the trigger)

Most implementations use one of these two equivalent definitions:

$$c = \min\left( \frac{R_a \cdot V_b}{R_b \cdot V_a},\ \frac{R_b \cdot V_a}{R_a \cdot V_b} \right)$$

or the more symmetric version

$$c = \frac{2}{\frac{R_a V_b}{R_b V_a} + \frac{R_b V_a}{R_a V_b}}$$

• c ≈ 1 → virtual liquidity is well balanced
• c ≪ 1 → the pool is very imbalanced (one side has almost all virtual liquidity)

Glide is only activated when c < m (e.g. c < 0.80–0.85).

---

The glide update rule (core mathematics)

When glide is triggered, the virtual balances are updated using exponential decay toward the target:

$$V_{a,\ \text{next}} = T_a + (V_{a,\ \text{current}} - T_a) \cdot (1 - \tau)^{\Delta t}$$ $$V_{b,\ \text{next}} = T_b + (V_{b,\ \text{current}} - T_b) \cdot (1 - \tau)^{\Delta t}$$

Where usually Tₐ and Tₓ are chosen such that

$$\frac{T_a}{T_b} = \text{target price (in token B / token A)}$$

Most important properties of this formula

1. Asymptotic convergence → Vₐ and Vₓ approach Tₐ and Tₓ exponentially
2. The larger the deviation (|V – T|), the larger the absolute step → strong restoring force
3. The step size decays exponentially → movement becomes slower and slower → no sudden jump
4. Time-proportional → twice the time → roughly twice the distance covered (in log scale)

---

Typical values seen in production designs

Parameter	Conservative (stable pairs)	Aggressive (volatile / meme)	Very aggressive (new launch)
m (min centeredness)	0.82 – 0.87	0.75 – 0.80	0.65 – 0.75
τ (daily exponent)	0.30 – 0.45	0.50 – 0.70	0.80 – 1.00
Min glide interval	4–12 hours	1–4 hours	15 min – 1 hour
Max glide per trigger	~3–8% of virtual depth	10–25%	30–50%

---

Visual intuition – how the glide looks

Imagine the virtual price starts below the target and moves up over time. Think of the target as the top rung and the glide as climbing toward it: each step is largest when you are far away and smallest when you are almost there.

Time runs along the journey (left to right in the story, bottom to top in the diagram): the farther below target you start, the faster the absolute correction feels at first; near the target, movement asymptotically slows so there is no sudden snap.

---

Why this is powerful (ReClamm advantages)

Property	Classic Uniswap v3	Classic x*y=k	ReClamm glide
Needs active LP to re-range?	Yes	No	No — automatic
Sudden large arb opportunity?	When LP moves	Always	Very small — gradual
Capital efficiency during drift	Drops fast	Constant	Stays high — virtual recenters
Good for trending markets?	Poor	Poor	Good — follows slowly
Good for range-bound stablecoins?	Good	Ok	Excellent — stays centered
Good for high-vol meme launches?	Very poor	Poor	Very good — damps extremes

---

Most important engineering trade-offs

1. Too high τ → pool chases price too aggressively → large arb during strong trends
2. Too low τ → pool stays imbalanced for too long → bad capital efficiency
3. Too low m → glide almost always active → unnecessary state writes / computation
4. Too high m → glide almost never triggers → defeats the purpose

Sweet spot (most production designs)
→ m ≈ 0.78–0.84
→ τ ≈ 0.45–0.65 (daily)
→ glide triggered at most once per 1–4 hours

---

Further exploration

Related topics you can go deeper into:

• How the glide affects effective slippage in different market regimes
• A numerical example with concrete numbers over 3 days
• How to choose τ and m given historical volatility of the pair
• Main failure modes / attack vectors when implementing glide

See the internal ReClamm design docs for deeper detail.