First-Passage Percolation Lab Quanta 2025 Breakthrough

Network Flooding Speed & Limit Shape Simulator

How fast does a porous network flood when fluid opens up? In late 2025, a landmark paper by five mathematicians settled long-standing bounds on first-passage percolation speeds. Test stochastic passage distributions, trace expanding shortest geodesics, and observe emergent deterministic shape norms.

Fluid Wavefront & Geodesic Tree

Ready • Click lattice to add fluid source
Flooded Cluster B(t)
Wavefront Frontier ∂B(t)
Shortest-Path Geodesic Ray
Fluid Source Node
Dry Lattice Edge

Lattice Parameters

Flooded Vertices 1
Simulation Time 0.00
Frontier Speed v(t) 0.00
Max Manhattan Radius 0

Flooding Speed v(t) vs Radius dr / dt

The Mathematical Frontier: Proved Theorems vs Open Conjectures

First-passage percolation (FPP), introduced by Broadbent & Hammersley (1965), is the standard model of fluid injection into random media. While fundamental global laws are now proved, microscopic boundary fluctuations remain one of modern probability's greatest challenges.

Rigorous Theorems (Proved)

The Asymptotic Shape Theorem (Cox–Durrett 1981)
There exists a deterministic convex compact set B0 ⊂ ℝd such that for any ε > 0, almost surely (1−ε)t B0 ⊆ B(t) ⊆ (1+ε)t B0 for all sufficiently large t. Randomness washes out on macroscopic scales.
Linear Growth Speed (Kingman's Subadditive Ergodic Theorem)
The passage time T(0, n·x) / n converges almost surely to a time constant μ(x). The asymptotic flooding speed along any ray is finite and strictly positive if and only if P(τ = 0) < pc.
The 2025 Five-Mathematician Speed Resolution
In late 2025, mathematicians resolved the sharp universal lower bound on flooding speed across disordered networks, establishing that bottleneck passages cannot decelerate the global wavefront below ballistic linear scaling.

Open Conjectures (Unproven)

KPZ Fluctuation Exponent in d = 2 (χ = 1/3, ξ = 2/3)
It is widely conjectured that Var(T(0, nx)) ~ n2/3, meaning wavefront fluctuations grow as n1/3. Proved for solvable models (corner growth), but unproved for standard FPP on ℤ2.
Strict Convexity of the Limit Ball B0
Does the limit shape have flat facets? It is conjectured that for continuous distributions like Exponential(1), the boundary ∂B0 is strictly convex with no flat edges, but a general proof remains elusive.
Fluctuations in High Dimensions (d ≥ 3)
Does the KPZ scaling relation 2χ = 2ξ − 1 persist in 3 dimensions and beyond? Fluctuation exponents in ℤ3 remain entirely open with conflicting numerical predictions.
Enjoy this tool? Build your own with Super