On the Convergence Rate of the Quasi- to Stationary Distribution for the Shiryaev-Roberts Diffusion
Pith reviewed 2026-05-25 02:13 UTC · model grok-4.3
The pith
The quasi-stationary cdf of the Shiryaev-Roberts diffusion converges to the stationary cdf at rate O(log(A)/A) uniformly as the absorbing boundary A tends to infinity.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
It is shown that the rate of convergence of the diffusion's quasi-stationary cumulative distribution function Q_A(x) to its stationary cdf H(x), as A→+∞, is no worse than O(log(A)/A), uniformly in x≥0. The result is established explicitly by constructing new tight lower- and upper-bounds for Q_A(x) using certain latest monotonicity properties of the modified Bessel K function involved in the exact closed-form formula for Q_A(x) recently obtained by Polunchenko (2017).
What carries the argument
The exact closed-form formula for Q_A(x) involving the modified Bessel function of the second kind, whose monotonicity properties are used to derive explicit sandwiching bounds on the difference from the stationary cdf H(x).
If this is right
- The stated rate bound holds uniformly over the entire state space x ≥ 0.
- The explicit bounds on Q_A(x) are obtained directly from the closed-form expression without additional approximation steps.
- The convergence rate tends to zero as A increases, confirming that the effect of the absorbing boundary vanishes in the limit.
Where Pith is reading between the lines
- The same monotonicity technique on the Bessel K function could be applied to quasi-stationary distributions of related one-dimensional diffusions that admit similar closed forms.
- The O(log(A)/A) rate supplies a concrete error term that can be plugged into performance analyses of change detection procedures that rely on the Shiryaev-Roberts statistic with a large threshold.
Load-bearing premise
The exact closed-form formula for Q_A(x) from Polunchenko (2017) together with the monotonicity properties of the modified Bessel K function are sufficient to construct the stated tight bounds.
What would settle it
A direct numerical evaluation of |Q_A(x) - H(x)| for successively larger A that exceeds C log(A)/A for every fixed C and some x would falsify the claimed uniform rate.
read the original abstract
For the classical Shiryaev--Roberts martingale diffusion considered on the interval $[0,A]$, where $A>0$ is a given absorbing boundary, it is shown that the rate of convergence of the diffusion's quasi-stationary cumulative distribution function (cdf), $Q_{A}(x)$, to its stationary cdf, $H(x)$, as $A\to+\infty$, is no worse than $O(\log(A)/A)$, uniformly in $x\ge0$. The result is established explicitly, by constructing new tight lower- and upper-bounds for $Q_{A}(x)$ using certain latest monotonicity properties of the modified Bessel $K$ function involved in the exact closed-form formula for $Q_{A}(x)$ recently obtained by Polunchenko (2017).
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper claims that for the Shiryaev-Roberts martingale diffusion on [0,A] with absorbing boundary at A>0, the quasi-stationary cdf Q_A(x) converges to the stationary cdf H(x) as A→+∞ at a rate no worse than O(log(A)/A), uniformly in x≥0. The result is obtained explicitly by constructing tight lower and upper bounds on |Q_A(x)−H(x)| from the closed-form expression for Q_A(x) in Polunchenko (2017) together with monotonicity properties of the modified Bessel K function.
Significance. If the bounds are valid, the explicit rate supplies a concrete, verifiable quantification of the approximation quality between quasi-stationary and stationary regimes for large thresholds. This is useful in sequential analysis. The direct analytic derivation from the 2017 closed form, rather than an appeal to general convergence theorems, is a methodological strength that makes the result self-contained and potentially reproducible.
minor comments (3)
- Abstract: the phrase 'certain latest monotonicity properties' of the modified Bessel K function should be accompanied by an explicit citation to the source of those properties.
- The manuscript should verify in the main derivation that the implicit constants in the O(log(A)/A) bound are independent of x, so that uniformity is immediate from the stated inequalities.
- All displayed equations involving the difference Q_A(x)−H(x) should be numbered consecutively and cross-referenced in the text when the bounds are applied.
Simulated Author's Rebuttal
We thank the referee for the careful reading and positive assessment of the manuscript, including the accurate summary of the main result and the recommendation for minor revision. No specific major comments were raised.
Circularity Check
Minor self-citation to co-author's 2017 closed-form; rate bound derived independently via new inequalities
specific steps
-
self citation load bearing
[Abstract and Section 1 (introduction of Q_A(x))]
"using certain latest monotonicity properties of the modified Bessel K function involved in the exact closed-form formula for Q_A(x) recently obtained by Polunchenko (2017)"
The citation supplies the closed-form expression that is then manipulated to produce the convergence-rate bound. While the manipulation is new, the load-bearing input is a result by a co-author; however the bound itself is not tautological with the 2017 formula and therefore does not meet the stricter load-bearing circularity threshold.
full rationale
The paper cites Polunchenko (2017) solely for the explicit formula of Q_A(x) involving the modified Bessel K function. From this input the authors construct fresh lower/upper bounds on |Q_A(x) - H(x)| by invoking monotonicity properties of K and direct integral estimates, yielding the O(log(A)/A) rate uniformly in x. This is a standard use of a prior mathematical result as a starting point rather than a reduction of the target claim to the citation itself. No self-definitional loop, fitted-input prediction, uniqueness theorem, or ansatz smuggling is present. The central derivation remains analytically independent of the 2017 paper beyond the shared formula.
Axiom & Free-Parameter Ledger
axioms (1)
- standard math Monotonicity properties of the modified Bessel K function
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.