Search arXiv⌕ Search

arXiv · 2611.09592

Discrete total variation under the multinomial, Part 4: variance under a general distribution

Abstract

We study the variance of the reference-centered discrete total variation $d_{\mathbf p}=\sum_{j=1}^{n-1}|X_{j+1}-X_j-N(p_{j+1}-p_j)|$ of $X\sim\mathrm{Mult}(N;\mathbf p)$ for fixed $n\ge3$ and fixed $\mathbf p$ with all $p_i>0$, as $N\to\infty$. We prove $\mathrm{Var}(d_{\mathbf p})=N\,v(\mathbf p)+O(1)$, where $v(\mathbf p)$, the variance of the sum of absolute values of the limiting Gaussian edge vector, is expressed through the exact multinomial edge variances and correlations and Nabeya's covariance formula for folded bivariate normal variables. Unlike in the uniform case, non-adjacent edge pairs generally contribute to $v(\mathbf p)$. The $O(1)$ remainder is controlled by a lattice Edgeworth expansion and Poisson summation; it oscillates with the lattice phases $\{N(p_{j+1}-p_j)\}$, and for $n=3,4$, where a pair of edges exhausts the bins, the summation runs over a sublattice coset. The formula reduces to the known uniform result (the classical Hoel--Kamat variance), gives a closed form when the reference law is a symmetric comb, and has a shape-independent large-bin limit for positive $C^1$ profiles. For the statistic $d_{\mathbf r}$ centered at a reference law $\mathbf r\ne\mathbf p$ the absolute values linearize: when every edge is mismatched the variance is that of a linear form in the counts up to an exponentially small error, and in mixed (triangular-array) regimes it is given to $o(N)$ by bivariate folded-normal covariances. The two-bin formulas are recorded from the companion preprints, and exact finite-$N$ reductions for $n\le4$ are used for validation.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Nikola Banić, Neven Elezović. 2026-11-06. Discrete total variation under the multinomial, Part 4: variance under a general distribution. https://arxiv.org/abs/2611.09592

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Line crossing problem for biased monotonic random walks in the plane

A $β$-biased monotonic random walk in the plane moves from $(x,y)$ to $(x+1,y)$ or $(x,y+1)$ with probabilities $1/(β+1)$ and $β/(β+1)$. We study $Φ_β(α,d)$, the probability that the walk started at the origin ever visits a lattice point strictly above the line $y=αx+d$, and $Ψ_β(α,d)$, the probability that it visits a lattice point \emph{on} the line after leaving the origin. For integer slopes $α=p$ we show $Φ_β(p,d)=ρ^{\lfloor d\rfloor+1}$, where $ρ$ is the smallest positive root of $y^{p+1}-(β+1)y+β=0$, give the Fuss--Catalan series for $ρ$, identify $ρ$ with the extinction probability of a Galton--Watson process, and compute the exact law and mean of the first crossing point. We prove that the return probability is $Ψ_β(p,0)=(p+1)(1-β/((β+1)ρ))$, which equals $(p+1)/(β+1)$ for $β\ge p$. For rational slopes $α=r/s$ we obtain $Φ_β$ in closed form from the $s$ large roots of the kernel $\tfracβ{β+1}z^{r+s}-z^r+\tfrac1{β+1}$, and $Ψ_β$ as a ratio of Green's functions. For arbitrary real $α$ we prove that the walk crosses every line $y=αx+d$ almost surely if and only if $β\geα$, and we establish the two-sided bound $e^{-θ(d+1)}\leΦ_β(α,d) β$.

math.PR↗

Commutation relations for two-sided radial SLE

We study the commutation relation for 2-radial SLE in the unit disc starting from two distinct boundary points. We follow the framework introduced by Dubédat~[12]. Under an additional requirement of the interchangeability of the two curves, we classify all locally commuting 2-radial SLE$_κ$ for $κ\in (0,8)$: it is either two-sided radial SLE$_κ$ with spiral of constant spiraling rate or chordal SLE$_κ$ weighted by a power of the conformal radius of its complement. Namely, for fixed $κ$ and starting points, we have exactly two one-parameter families of locally commuting 2-radial SLE curves. We also discuss the semiclassical limit of the commutation relation as $κ\to 0$. In particular, we show that the limit for the second family with an appropriately chosen power of the conformal radius is a chord that minimizes a modified chordal Loewner energy, which is unique only when the endpoints are not antipodal.

math.PR↗

Rough Heston model as the scaling limit of bivariate cumulative heavy-tailed INAR processes: Weak-error bounds and option pricing

We study nearly unstable bivariate cumulative heavy-tailed INAR($\infty$) processes and show that, under a one-factor parameterization and a suitable scaling, they converge to the rough Heston model. This yields a discrete-time microstructural route to the joint price-variance dynamics and gives explicit formulas linking the INAR asymmetry parameters to the leverage correlation and diffusion scale of the limiting volatility process. For pricing, we reduce the exact finite-$τ$ Fourier--Laplace transform recursion to a quadratic discrete Volterra equation and compare it with the rough Heston fractional Riccati equation. Under a transform-strip condition for the continuous model and bounded-Riccati regularity, every fixed bounded frequency window admits a uniform transform-error bound of order $τ^{-α}+τ^{-(1-α)}$. The finite-$τ$ transform is entire, and a uniform exponential-moment estimate then removes the Fourier cutoff and yields convergence of the exact, untruncated call prices. We also develop an FFT-accelerated CDQ simulator with $\mathcal O(τ\log^2τ)$ complexity per path and use it to price European and path-dependent options, benchmark the Markovian case $α=1$, and analyze the model-implied volatility surface.

math.PR↗