Work in the prime end closure of the marked domain, so that its two conformal boundary points are distinct even if their physical boundary representatives coincide. A chordal filling is a closed connected set joining those points, meeting the conformal boundary nowhere else, and full relative to the two boundary arcs between them: complementary pockets that do not reach either boundary arc have been added. In the normalized complex upper half-plane, its two complementary sides are simply connected and abut the negative and positive real axes. For a proper boundary-to-boundary curve, chordal filling means adjoining all such trapped pockets to its range. Independent curves are combined by taking their union and then applying this same chordal filling operation.
An admissible marked subdomain contains neighborhoods of the two marked boundary points and has a simply connected component joining them. The restriction property says that, conditional on the random chordal filling being contained in , its law is the chordal filling law for . Equivalently, mapping the conditional chordal filling conformally back to recovers the original law. The contained-set events are understood away from the shared marked endpoints.
By transporting the family under conformal maps, it suffices to work in . Let be an admissible compact H-hull whose closure avoids zero, and normalize
In terms of the single normalized measure , restriction is precisely
The Schwarz reflection principle extends analytically near zero, with , so the stated conditioning events have positive probability for the two laws considered below. This is the chordal restriction property for filled sets. An unbounded marked subdomain can be approximated by such compact-hull removals; conformal maps transport the formulation to general domains.
We need a determination principle for filled sets, since a probability of one avoidance event alone would not establish equality of laws. Choose countably many thin polygonal compact H-hulls with rational data, attached to real boundary intervals away from zero. For each chordal filling, every point of its complement has a path within one of the two complementary sides to the appropriate real boundary interval. A sufficiently thin tube along such a path avoids the chordal filling and contains a neighborhood of that point. Hence the union of interiors of all avoided test tubes is exactly the complement.
Their joint avoidance probabilities are also determined by single-hull probabilities. For a finite union of test compact H-hulls, either the union separates zero from infinity, in which case a connected chordal filling cannot avoid it, or its unbounded complementary component contains both marked endpoints. In the latter case fill the union's bounded pockets to get an admissible compact H-hull . A connected set joining zero to infinity and avoiding the union cannot enter one of these separated pockets, so it avoids as well. Thus joint avoidance is either impossible or is the avoidance event for . The inclusion-exclusion principle gives every finite joint probability distribution of the tube-avoidance indicators, and these indicators determine the random complement and chordal filling. This proves avoidance probabilities determine a chordal filling law.
In the given avoidance formulas, the argument of the conformal map must be the retained open domain, rather than the random closed filling. We express that domain as to keep the two objects distinct. The formulas then imply for the two filled processes
The chordal filling operation does not change avoidance of an admissible compact H-hull: every piece of that compact H-hull is attached to the real boundary away from the endpoints, and so cannot be inside a trapped pocket while its generating curve avoids the compact H-hull. These laws are conformally invariant as unparameterized chordal filling laws. For SLE this follows from Question 1; for the Brownian half-plane excursion, Brownian scaling of its independent one- and three-dimensional coordinates shows that its unparameterized law is unchanged by positive dilations, and mapping the Brownian half-plane excursion to another marked domain gives the conformally natural law. Conformal invariance of planar Brownian motion describes the corresponding change of time, which does not alter the chordal filling.
Here is the conditional-law calculation for either exponent . For another admissible compact H-hull in the image complex upper half-plane, set
The complement of is mapped to by this composition, so is again admissible; both maps fix zero and have derivative one at infinity. The chain rule gives . Therefore
These are exactly the original chordal filling's avoidance probabilities for every . The determination principle proves equality of the conditional image law with the original law. Hence both generated chordal fillings have restriction, with exponents and respectively. This is a complete implication from the two allowed avoidance formulas; the restriction conclusion was not assumed.
For independent chordal fillings, avoidance of by their filled union is equivalent to avoidance by every component. Independence multiplies the probabilities, and thus adds their restriction exponents of a chordal filling. Consequently
The determination principle applies once more, giving
The equality concerns the filled, unparameterized closed sets. The numerical reason is .
Each marginal is a Bessel process of dimension , but finite lifetime can also be proved directly. Apply the Itô formula before reaches zero:
With clock , the Itô integral is a Brownian motion . The inverse clock satisfies
The clock runs through all up to the maximal lifetime: a finite terminal clock would leave at a finite positive limit, allowing continuation, or, at an infinite physical lifetime, would make its clock grow at a positive rate, a contradiction. There is no finite-time explosion to infinity; for example and localization bounds the probability of reaching arbitrarily large levels on a fixed time interval. Since almost surely and , the last exponential integral has finite limit, while . Therefore
The same argument applies to , whose noise is also Brownian motion, so its lifetime is finite too. No independence between the two lifetimes is being claimed.
Set and before . Brownian motion terms cancel in the sum:
Thus simultaneous extinction is impossible. At the finite first lifetime, precisely one coordinate is zero and the other remains positive. Because has finite variation, differentiating the ratio gives
After clock change , its generator is
A scale function of a one-dimensional diffusion solves , so
Both endpoint singularities are integrable since . Normalize to obtain
The Itô formula makes a bounded martingale, first by localization away from the endpoints and then by bounded convergence theorem. Its terminal value is one on , when , and zero on the opposite event, when . optional stopping theorem therefore proves the oppositely driven Bessel exit probability
Here is the beta function and is the Gamma function.
The integral printed in the question is incorrect for general . Its first exponent is , whereas the two coupled equations give . At , which lies in the allowed interval, its integral from zero diverges for every positive upper limit; no finite normalizing constant gives a probability. Moreover exchanging and while changing to shows that at the probability must be . The corrected symmetric integral has this value for every , whereas the printed asymmetric normalized integral generally does not. For , even when the printed integral converges, its derivative density satisfies
so it is harmonic for the required exit problem only at . The two exponents agree exactly at that value.
To relate that value to percolation, write the two SLE boundary gaps as and . In Brownian-time units , they have the displayed coupled equations with , after reversing the Brownian motion sign. Hence corresponds to . The event is that the positive marked point is swallowed before the negative marked point .
For a quadrilateral with boundary points in the order , its conformal cross-ratio coordinate is . In the critical percolation exploration, wire the arcs and blue and the other two arcs yellow. Exploration from zero, stopped when one of the two marked sides is swallowed, detects a blue crossing between those two blue arcs exactly in the event just computed. The continuum exploration has the SLE6 law; its Locality property of SLE lets one stop the ordinary chordal exploration before the distant boundary colors affect it. Thus the crossing probability is
the Cardy boundary crossing formula. This is the rigorous scaling-limit formula for critical site percolation on the triangular lattice; applying it to other critical planar models requires the corresponding universality assumption. At this percolation value of , the printed integral is already the correct one.
Let and . We use these precise Bessel process facts, which the question permits: for dimension a process started at exists for all time, never reaches zero, and has a strictly positive all-time minimum. More quantitatively,
No assertion about SLE simplicity or transience is being assumed.
For a real point before its Loewner lifetime, set . Its equation is
Thus is a Bessel process of dimension . For each rational it stays positive on every finite interval, almost surely. These countably many events hold together. They cover every real as well: choose rational and observe that the noise cancels in the difference,
Consequently cannot reach zero while remains positive. The negative side has the identical argument with the reflected Brownian motion.
This excludes Loewner trace contact with every real point other than zero, rather than just with each fixed rational point. On any finite interval a positive gap at a real point stays bounded away from zero. Continuous dependence of the Chordal Loewner equation then solves it on a complex neighborhood of that point, and the analytic flow extends the map conformally across a real interval there. Its real derivative is , and its image stays away from the driving point. The Loewner trace tip therefore cannot be there. We have proved .
Now use Brownian motion independent increments to restart at every deterministic rational time . The centered future maps have Loewner driver and therefore the original SLE law, independently of the past. Its Loewner trace stays in by the preceding argument. The inverse map extends continuously to the real boundary for a continuous trace-generated compact H-hull: the finite Loewner trace gives a locally connected boundary, and the Caratheodory boundary extension theorem applies. Its value at is . Thus the future in the original plane can meet only at . All these restart conclusions hold simultaneously for rational .
Suppose with . The Loewner trace cannot be constant throughout , since its compact H-hull capacity increases strictly. By continuity there is a rational with . But lies in , contradicting the restart conclusion at . This also excludes any return to the starting point. Hence the Loewner trace is simple and lies in the open complex upper half-plane at every positive time.
For transience, first strengthen the boundary conclusion to avoidance of the closure at each fixed real point. Fix and . If the now-simple curve first enters the closed half-disc of radius about at time , join its tip to by the segment inside that disc. Before the segment misses the Loewner trace. Together with the Loewner trace and the real interval from zero to , it bounds a pocket whose boundary includes the positive side of the Loewner trace and . A Brownian motion path from high on the imaginary axis reaching that boundary arc must enter the pocket through the segment. The maximum principle for harmonic functions bounds its harmonic measure by the chance of hitting the entire half-disc.
The boundary arc maps under to . Its harmonic measure from , as , is
The half-disc maps out by ; its semicircle becomes an interval of length . Its harmonic measure has asymptotic . Comparing and letting yields
But by the all-time minimum of a transient Bessel process fact. The curve cannot enter any such disc with . Therefore it stays a strictly positive distance from for all time. Reflection proves the same for fixed . This is the boundary approach forces a small SLE Bessel gap argument; the crosscut comparison is also described in the boundary-closure lemma in the primary SLE paper.
Map out the simple initial slit at time one. The starting point zero has two distinct real prime end images under , lying on opposite sides of . Conditionally on the initial slit, the image future is a fresh independent SLE translated by . The just-proved fixed-point closure avoidance applies to each of and . Therefore neither nor lies in the closure of the mapped future. If a sequence of future Loewner trace points approached zero in the original domain, their images would approach one of these two prime ends. This is impossible. Thus
Finally Scaling invariance of SLE gives . For every fixed ,
The events of returning to the closed radius- disc after time decrease as increases. Continuity of probability on decreasing events implies that the probability of returns at arbitrarily late times is zero. Taking the countable intersection over positive integer proves
This proves actual eventual escape, not merely unboundedness of the growing compact H-hulls.
A compact H-hull is a bounded relatively closed subset of with simply connected complement, considered together with its compact closure in . Its unique mapping-out function has hydrodynamic normalization at infinity
The nonnegative coefficient is its half-plane capacity. Thus the half-plane-capacity parameterization here gives . Strict increase means whenever ; the strictly increasing capacity already excludes equality.
The Loewner local growth property means that the new growth becomes small after removing the old compact H-hull. More precisely set , taking the closure from the complex upper half-plane. On each bounded time interval, the diameters of these image compact H-hulls tend uniformly to zero as . Equivalently the new growth can be separated from infinity by crosscuts of vanishing diameter in the mapped domain. It is this single-point local growth, together with capacity continuity, that gives a continuous real Loewner driver.
The Loewner transform is the limiting point at which this mapped growth occurs:
The limit is independent of the chosen new-growth points. The Loewner correspondence theorem says that the resulting mapping-out functions satisfy
Conversely, given a continuous real , solve this ordinary differential equation until the maximal lifetime of each . The surviving set is mapped conformally onto by ; its complement, with the appropriate boundary closure, is . The Laurent coefficient is , and the composition law reconstructs the local growth. A continuous Loewner driver always determines the compact H-hulls; generation by a continuous Loewner trace is an additional property, not a conclusion for every continuous Loewner driver.
A chordal SLE Loewner trace in the complex upper half-plane with parameter is the continuous curve associated with the Loewner driver , where is standard real Brownian motion started at zero. It starts at zero, its compact H-hull at time is the curve's past with all pockets cut off from infinity filled, and
The capacity clock has . For this gives the vertical slit .
To prove Scaling invariance of SLE, fix and put
The Laurent series of has coefficient , and direct differentiation gives
Brownian scaling says has the same law as . The Loewner drivers therefore agree in law, and uniqueness of the Chordal Loewner equation gives equality in law of the maps, compact H-hulls, and continuous Loewner traces:
This identifies the time change as well as the spatial scaling.
For a simply connected Jordan domain with distinct marked boundary points, choose a conformal bijection taking zero to the initial point and infinity to the target, and use as an unparameterized curve. The Caratheodory boundary extension theorem supplies the boundary values of . Any second such map is for , because the automorphisms of fixing zero and infinity are positive dilations. The Scaling invariance of SLE just proved makes the two image laws identical, up to the corresponding deterministic change of capacity clock. Thus this defines SLE consistently in marked Jordan domains. The assumed and the continuous boundary extension give , the specified target. The law is conformally natural and does not impose a distinguished half-plane-capacity parameterization on an arbitrary domain.
Let and put , counting each edge once. If only the spin at changes from to , the change in is , so
The given opposite transition probabilities have the same ratio. More explicitly,
For pairs differing at more than one site both sides vanish, and for equal configurations balance is automatic. This proves detailed balance, hence the chain is reversible with stationary law . All single-spin changes have positive probability, so the finite chain is irreducible; it also has positive holding probability and is aperiodic.
Its random-map version is a heat-bath Markov chain. Choose a vertex uniformly and a uniform , independently, and set its spin to when
and to otherwise, leaving all other spins unchanged. This also specifies the holding probabilities omitted from the off-diagonal formulas. The conditional Ising law at is exactly this Bernoulli choice.
Order spin configurations coordinatewise with . Since both the neighbour sum and are increasing, every map defined by a fixed preserves this order. For monotone coupling from the past, generate one two-sided independent sequence of such maps , , and, for a horizon , apply to the all-minus and all-plus states at time . If their time-zero images agree, every initial state has that same image by monotonicity. Return it. Otherwise enlarge , for example by doubling it, retaining all maps already generated at existing negative times. Any earlier starting configuration is also forced to the same output once a horizon coalesces, so the result is consistent under further backward extension.
To prove termination, let and
List the vertices . A block of updates choosing them in that order and having every uniform less than sets every vertex to , regardless of the starting configuration. Each choice has probability , and every uniform condition has probability . Thus a block is a universal reset with probability
Independent disjoint blocks in the past give
Once such a block occurs, later maps preserve agreement. Therefore the backward search terminates in finite time almost surely, including the doubling implementation.
Finally we prove that the common output has law . For a deterministic horizon , independently start another chain at time with initial law and use the same maps. Its time-zero state has law by stationarity. On the event that this deterministic horizon coalesces, , so for every event ,
Letting deterministic proves
This argument justifies exact sampling without treating a random, data-dependent horizon as a deterministic stationary run.
For the graphical representation of the contact process, put independent death marks on each vertical time-line at rate , and independent infection arrows at rate for every ordered nearest-neighbour pair. A time-directed path moves vertically upward, cannot pass a death mark, and may follow an infection arrow. From an initial set , the infected set consists of sites reachable at time from . Thus infected sites recover at rate , and a healthy site with infected neighbours becomes infected at rate . Rates here are per ordered edge, fixing the normalization of .
With , write
Couple two infection rates by adding independent arrows of rate to the rate- construction, retaining the same deaths. Every path available at the lower rate remains available at the higher rate. Hence is nondecreasing.
For the lower bound, dominate the infection population by a linear birth-death branching process , started from one particle, with per-particle birth rate and death rate . Assign one branching particle to every infection; allow an offspring at every attempted transmission, including attempts into an already infected site, and let the extra particles continue independently. Matched deaths remove their infection. This constructs . The linear birth rate is nonexplosive, so in particular the contact process from a finite initial set is well-defined with finite population at finite times. Its branching mean solves
Consequently
If , let to obtain . This proves the branching bound for contact-process survival
For the upper bound use the dimension comparison for the contact process. Define the height . Construct an auxiliary infected set together with an infected representative of height for every . Initially and . Use the original graphical marks as follows.
  • A death at removes from as well as removing that representative from . Other deaths change only .
  • From there are exactly arrows into height and exactly arrows into height . When one of these arrows occurs and the target height is absent from , infect in the auxiliary process and use the arrow's target as its representative. That target belongs to after the arrow, whether it was newly infected or already infected.
  • Arrows from other sites and arrows into an already occupied auxiliary height produce no auxiliary change.
This rule maintains a live representative for every auxiliary infection. Conditional on the full past, each active representative has death rate and outgoing-arrow rate in each of the two height directions. Representatives at distinct heights are different sites, using distinct death clocks and distinct outgoing ordered-edge clocks. Independence and memorylessness of the underlying Poisson processes give exactly the one-dimensional contact-process generator
Thus has the law of the one-dimensional contact process with infection rate , even though the representatives themselves depend on the higher-dimensional history. Since implies at every time,
Every for which is therefore supercritical whenever the one-dimensional survival set is nonempty; if that critical value were infinite the inequality would be trivial. Taking infima gives
On the coordinatewise ordered configuration space, stochastically dominates when for every increasing real function . Equivalently, for every increasing event. The Holley condition is the sufficient cross-lattice inequality
Here join and meet are coordinatewise maximum and minimum. This condition implies by the Holley inequality; it is not asserted to be necessary. It can be checked on unnormalized weights, since their normalization constants cancel.
Use the weights
For the upper bound, take and . Since open-edge count is modular, the Bernoulli factors cancel in the Holley ratio, leaving
The last inequality uses and the fact that removing edges cannot reduce the number of components. Thus .
For the lower bound, let . Its odds satisfy
Set , the number of edges added to in the union. For and , the corresponding ratio is
Adding one edge either merges two components or leaves the count unchanged. Therefore adding edges reduces by at most , proving the exponent nonnegative. This also proves the hinted monotonicity of rather than assuming it. The Holley condition now proves both Bernoulli bounds for the random-cluster model:
The same argument holds with wired boundary conditions: identify the wired boundary vertices first and count components on that quotient graph. Taking finite-volume wired limits preserves every increasing cylinder-event inequality. Approximating by the decreasing events consequently gives
Let . For , the upper bound makes the wired percolation probability zero, so . If and , the lower bound makes it positive. Solving this strict inequality,
Taking infima proves
For , the lower bound and the trivial upper bound give the same conclusion. No assumption about percolation at the critical parameter is required.
Let be the number of open edges, the number of connected components in the open subgraph, counting isolated vertices, and . Summing the joint weight over spins requires a constant spin on every open component. There are exactly choices. Thus the edge marginal is
This is the random-cluster measure, and the joint normalizing constant is .
Summing instead over edge configurations factors independently over edges. For a fixed spin configuration the resulting weight is
since . The factor is independent of the spins. Hence the spin marginal is exactly the ferromagnetic Potts measure in the convention
Its partition function satisfies .
Conditionally on , edge states are independent: an edge with different endpoint spins is closed with probability one, while one with equal endpoint spins is open with probability . Conditionally on , color its open components independently and uniformly from . Equivalently, every compatible spin configuration has conditional probability , and all incompatible configurations have probability zero. These are the two conditional laws of the Edwards-Sokal coupling.
If belong to the same open component, their colors agree surely; otherwise the two independent component colors agree with probability . Therefore
Rearranging proves the spin-connectivity identity for the Potts model:
Phase transitions refer to infinite-volume limits, not a single finite graph. The identity shows that connectivity is exactly the excess spin agreement above its independent-color baseline. More directly, use a wired random-cluster boundary condition and fix the corresponding boundary spin to color . Its component has color , while every finite component has a uniform color. In the compatible infinite-volume coupling,
Consequently normalized spontaneous magnetization is
The onset of a wired infinite cluster and the onset of spontaneous magnetization in the Potts model are the same transition under , giving . This also transfers any jump or continuity of that order parameter between the two descriptions; it makes no assumption about behavior exactly at the critical point.
Order configurations coordinatewise: means that every open edge of is open in . An increasing event satisfies . For , let be the configuration with exactly the edges of open. For increasing events,
The disjoint sets are open witnesses: fixing those edges open forces the respective events whatever happens elsewhere. This is disjoint occurrence of increasing events. The BK inequality states that under the independent density- product measure,
Disjoint occurrence means disjoint witnesses, not disjoint events.
Now take the lattice rectangle with width and height . Make the dual bonds open exactly when the primal bonds they cross are closed. A primal open left-right crossing and a dual open top-bottom crossing cannot coexist: the two curves must intersect, requiring a bond and its crossing dual bond both to be open. Conversely, if no primal left-right crossing exists, the boundary separating vertices reachable from the left side from the right side contains a closed-bond dual top-bottom path. One obtains it by tracing the interface of the left-reachable set; it cannot terminate in the interior, and it separates the two vertical boundary sides. Thus exactly one of the two crossing events occurs.
For the precise dual dimensions, its vertices have coordinates
The top and bottom rows are the terminals of dual edges crossing the primal boundary horizontal bonds. Initially no horizontal bonds along these two terminal rows are needed. Adding independent density- horizontal bonds there does not change a top-bottom crossing event: an initial run along the top row can be removed by starting at its last vertex, and a final run along the bottom row can be removed similarly. The completed dual rectangle has width and height . Rotate it through a right angle and translate it: its top-bottom crossing has exactly the law of the original left-right event at density . Therefore
Write and . This event is determined by finitely many bonds, since a path can be stopped on first hitting that boundary. On , erase loops from a crossing path and choose a vertex where it meets the middle column. Its two portions run to the left and right boundary respectively, each reaching distance at least from . Stop each portion at its first radius- boundary hit. The two portions have no common edges, giving disjoint witnesses for twice. Hence
There are exactly relevant middle-column vertices. The additional hinted vertex at is outside this rectangle and need not be included. Translation invariance, the union bound and the BK inequality yield
The cluster radius is its maximum distance in the norm, taking value for an unbounded cluster; thus . Taking square roots proves the square-lattice one-arm lower bound from disjoint occurrence
Write for an edge conductance and . Take , as implicit in a unit source-to-sink electrical network. An oriented current flow is antisymmetric, . Kirchhoff's first law is conservation of current: at every vertex other than the source and sink, with the divergences there equal to the prescribed total flow and its negative. Kirchhoff's second law says that the sum of oriented voltage drops around any closed cycle is zero. Ohm's law is
Thus the resistance of an edge is , and the cycle law reads for .
Let . Its expectation is finite. Indeed, from each state there is a path to of length at most with positive probability; finiteness gives a common lower bound on reaching within that many steps. The Markov property then gives , proving the assertion. Define the expected occupation count before absorption
Conditional expectation at every pre-stopping visit to gives for the number of directed traversals. Hence the expected signed count is
This proves Ohm's law with voltage , grounded at . Summing its voltage drops along a cycle telescopes, proving Kirchhoff's second law.
For Kirchhoff's first law, count entrances and exits on each stopped path. Their net difference at is the initial indicator minus the final indicator. Taking expectations,
Thus the source emits exactly one unit, the sink receives one unit, and all other vertices conserve current.
We also establish the uniqueness required for the conclusion. The cycle law implies that edge voltage drops are differences of vertex potentials: define a potential by summing drops along a path to , and cycle sums make this independent of the path. If two unit flows obey both laws, their difference has the form with zero divergence at every vertex. Therefore
Strict positivity of the conductances and connectedness force to be constant, so . We conclude the electrical current from stopped random-walk traversals identity
The exponent printed in the original PDF is too strong under the stated increment bound. Take , let equal or with equal probabilities, and set for . This is a martingale with and for . For and ,
Thus the printed inequality is false. The intended two-sided Azuma-Hoeffding inequality is
Here the are deterministic increment bounds. We prove this corrected assertion from conditional exponential moments.
Let . The martingale property gives . For , convexity of the exponential function gives the secant bound on ,
Taking conditional expectation removes the terms containing , so
The last inequality follows term by term from in the power series for and . If the increment is zero and the same bound holds directly. This is the bounded-increment case of the Hoeffding lemma.
Repeated conditioning, without any independence assumption, now gives
Write . If , the Markov inequality gives for ,
Minimizing at gives . Apply the same argument to the martingale for the lower tail, and use the union bound to obtain the boxed result. If , all increments vanish almost surely, so the deviation probability is zero for every . This handles the degenerate case without dividing by zero and proves the correct bound in full.
Let be the set of three-vertex subsets and let indicate that forms a triangle. Then
In particular . We will use the Stein-Chen method with edge-sharing dependency neighborhoods, rather than wrongly treating triangle indicators as mutually independent.
Here is the needed Poisson approximation with dependency neighborhoods, together with its proof. For indicators with means , let contain and suppose is independent of the entire family of indicators outside . With and , define
The total variation distance convention is . We claim
For an indicator test function on the nonnegative integers, solve the Poisson Stein equation
with . This elementary Stein factor can be proved as follows. Take the immigration--death process with immigration rate and per-particle death rate one. Starting at , its time- distribution is
with independent summands. Its stationary law is that of . Coupling the initial particles with a stationary Poisson initial population makes the centered expectations decay at least as a constant times , so
is finite. Its generator equation is , obtained by integrating the semigroup derivative. Set for ; choose , which does not affect the equation. To bound , couple the processes starting at by adding two independent initial particles. Both survive to time with probability , and otherwise their second difference cancels in expectation. Since ,
This proves the Stein factor for , and the chosen handles .
Put . Independence gives . Inserting this identity into the Stein equation yields
The first sum in absolute value is at most , since the two arguments differ by . In the second sum, on the arguments differ by , giving bound . Taking the supremum over event indicators proves the claimed total variation distance bound.
For triangle , choose to consist of itself and all triangles sharing an edge with . There are exactly such triangles. Every outside triangle uses none of the three edges of , even if it shares one vertex, so is independent of their entire indicator family. For distinct neighboring triangles, their union has five edges and . Therefore
The Stein-Chen method gives . Coupling Poisson distributions of two means by adding an independent Poisson variable gives
This proves the Poisson limit for triangles in a binomial random graph:
Applying the same bound to the event gives
For finite this is an approximation, not an exact formula: the error from is at most , and the error from is at most .

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact