The Setting
A pending coarse vertex comes with a finite graph in and an event that the origin reaches its inner cube. Revealing its coarse block raises the question of how to describe the remaining random variables and the probabilities attached to the other pending vertices.
Let be the bond or site model on the nearest-neighbor lattice . Write for the positive unit vector in coordinate direction . A lattice bond has endpoints and ; incidence with two vertices means that they are these endpoints in either order. The coordinate set is the set of lattice bonds for the bond model and for the site model.✦ A configuration assigns closed or open states to these coordinates.✦ Two lattice vertices are nearest neighbors when their difference is for some .✦ Open adjacency requires their joining bond to be open in the bond model, and both vertices to be open in the site model.✦ A vertex is available unconditionally for bonds and precisely when it is open for sites.✦
A finite lattice subgraph consists of finite sets of lattice vertices and bonds, with both endpoints of every bond in .✦ The underlying simple graph of a finite lattice subgraph has vertex set and adjacency given by endpoint incidence with a bond of .✦ Write for connection by a finite open path in .✦ A bond path may have length zero at any vertex. In the site model, an open path requires all its vertices to be open, including the coincident endpoints of a path of length zero. The finite target connection event for a vertex and a target set is the set of configurations specified by .✦ The graph variables are for bonds and for sites.✦
A lattice box has coordinatewise ordered corners.✦ The vertices of a lattice box are the integer triples with in every coordinate, and the bonds of have both endpoints among those vertices. Write for this finite lattice graph✦ and for its variables.✦ A graph contains when it contains all these vertices and bonds.✦ For a lattice center and a coordinate radius , the cube is the box with corners and .✦
Finite Histories
For a coordinate set and a parameter vector , let be the independent Bernoulli product measure, with opening probability at coordinate .✦ The notation means the constant parameter vector .
A finite pattern is the event prescribing a configuration on a finite set .✦ The pinned parameter vector fixes these coordinates to their prescribed states, as in (1).✦ (1) The partial replacement is the configuration equal to on and to outside .✦
A finite observation history consists of a finite observed set and a configuration specifying its values.✦ For a finite set , the extended history records the values of on , overwriting any earlier values there.✦ Its definition and the history parameter vector are given in (2). (2)
Coarse Blocks and Packets
A coarse point belongs to .✦ The four directions have outward steps , in that order.✦ For a nonnegative integer scale , the embedding and the surrounding and inner cubes are defined by (3). (3) For coarse points , the interface is the set of lattice bonds with one endpoint in and the other in , in either order. For coarse points , the interbox variable set is in the bond model and the empty set in the site model.✦
The block consists of the variables in and its interface with its parent .✦ The root variables and the exposed variables for a finite processed set are defined alongside it in (4). (4)
An exploration packet consists of a coarse vertex, a parent vertex, and a finite lattice graph containing and the lattice origin.✦ The root connection event of an exploration packet is .✦ For a graph containing and a coarse vertex , the extension of toward adds the vertices and bonds of and the interface bonds .✦ For an exploration packet and a coarse vertex , the child packet is the packet with vertex , parent , and graph obtained from by adding the vertices and bonds of and the interface bonds . The root packet has vertex and parent and graph .✦
The Exploration Rule
An exploration state consists of a finite history, a finite processed set, a finite ordered queue of packets, and finite sets of good and bad coarse vertices.✦ The visited set is .✦ For a packet , the candidate list consists of the children in direction order, retaining only those whose vertices are outside the visited set.✦
Write for a nonempty queue with first packet and remaining queue . The revelation set consists of the unobserved variables of the first queued block, and is the history after their revelation, as specified in (5). (5) For and , acceptance of a packet in state is the condition in (6), where denotes the values of the extended history. (6)✦
The single exploration step uses the history . For a nonempty queue , it adds to and removes from the queue. On acceptance, it appends to and adds to ; on rejection, it leaves as the queue and adds to . The other classification set is unchanged in either case. An empty queue leaves the state unchanged. The finite run applies this step times with the same configuration , starting from at .✦
A valid exploration state at is a state satisfying the conditions in (7). The last condition is the strict pending connection promise. (7)✦
Scale Parameters and Initialization
A quarter-face index is a pair consisting of a coordinate and a sign vector .✦ The quarter-face of the cube corresponding to the index is the set in (8); the 24 signed-coordinate quarter-faces may overlap on their boundaries. (8) The confined seed event consists of configurations having an available vertex and a vertex joined by a finite path of open lattice adjacencies restricted to .✦ Restricted adjacency requires both vertices of each step to belong to the containing cube.✦
Finite scale parameters form a tuple satisfying , , and . The radii and trial count are nonnegative integers with , , , and . Write for the seed variable count✦ and for the contour error tolerance.✦ The error and size budgets are (9). (9) The quarter-face requirement on these parameters is (10), for every center , every quarter-face index , and each . (10)
The initial state declares the root variables open, processes the coarse origin, and queues its four child packets as in (11), where is the configuration with every coordinate open. (11)✦ Write for the initial pinned product measure.
The Claim
Theorem 1. Let , let be finite, and let be a configuration. For every measurable event , the finite-history factorization is (12). (12)✦ For any finite history , parameter , finite observation set , and configuration , the parameter update is (13). (13)✦
Fix either model , a nonnegative integer , , and . Let be valid at , with queue , and let . For every configuration , the later packet probability satisfies (14). (14)✦ For any exploration state valid at and every configuration , a single exploration step preserves validity, as in (15). (15)✦ For any exploration state valid at , every configuration , and every nonnegative integer , the finite-run statement is (16). (16)✦ The initial state associated with finite scale parameters is valid, as in (17). (17)✦
In the paper, the result reads: conditional on any positive-probability finite revealed history , all unrevealed variables are independent Bernoulli, and for every pending packet at every step.
What Is Checked
The factorization concerns measurable events and arbitrary finite patterns; the exploration statements concern finite observation histories and packet graphs in either percolation model.