Complete Supermartingale Certificates for ω-Regular Properties
Pith reviewed 2026-05-21 01:40 UTC · model grok-4.3
The pith
Reactivity properties on Markov chains decompose into almost-sure termination obligations to absorbing regions.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Every reactivity property admits decomposition into multiple obligations of almost-sure termination into absorbing regions, and appropriate absorbing regions always exist on general state spaces. This enables the extension of every complete proof rule for almost-sure termination into a proof rule for reactivity that is complete in the almost-sure case, and complete up to an arbitrarily small ε-approximation in the quantitative case. Applied to supermartingales, the approach produces the first sound and complete supermartingale certificates for almost-sure ω-regular properties and the first sound and ε-complete supermartingale certificates for quantitative ω-regular properties on time-homog
What carries the argument
the decomposition of reactivity properties into almost-sure termination obligations into absorbing regions
If this is right
- Every complete proof rule for almost-sure termination extends to a complete proof rule for reactivity.
- Supermartingale certificates for almost-sure termination become sound and complete certificates for almost-sure ω-regular properties on countably infinite state spaces.
- Quantitative supermartingales become sound and ε-complete certificates for quantitative ω-regular properties.
- Standard quantitative safety results extend to quantitative reactivity via the same construction.
Where Pith is reading between the lines
- The same decomposition technique could be tested on non-countable or uncountable state spaces to check whether the existence of absorbing regions remains guaranteed.
- One could explore whether the number of required absorbing regions can be bounded for common reactivity patterns such as fairness or persistence.
- The approach suggests a route to lift other termination certificates, such as ranking functions, to full ω-regular verification without new machinery.
Load-bearing premise
For any reactivity property on a time-homogeneous Markov chain with general state space, appropriate absorbing regions exist that allow decomposition into almost-sure termination obligations.
What would settle it
A concrete reactivity property together with a time-homogeneous Markov chain on which no absorbing regions exist that permit the required decomposition into almost-sure termination obligations.
read the original abstract
We introduce a general methodology for the construction of sound and complete proof rules for the almost-sure and quantitative acceptance of reactivity properties on time-homogeneous Markov chains with general state spaces. Reactivity captures $\omega$-regular properties and subsumes linear temporal logic. Our core technical result establishes that every reactivity property admits decomposition into multiple obligations of almost-sure termination into absorbing regions, and that appropriate absorbing regions always exist on general state spaces. This enables the extension of every complete proof rule for almost-sure termination into a proof rule for reactivity that is complete in the almost-sure case, and complete up to an arbitrarily small $\varepsilon$-approximation in the quantitative case. We apply our new methodology to recent results on sound and complete supermartingale certificates for almost-sure termination in the special case of countably infinite state spaces, alongside standard results on quantitative safety. As a result, we obtain the first sound and complete supermartingale certificates for almost-sure $\omega$-regular properties and the first sound and $\varepsilon$-complete supermartingale certificates for quantitative $\omega$-regular properties on time-homogeneous Markov chains with countably infinite state spaces.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces a general methodology for sound and complete supermartingale certificates for almost-sure and quantitative acceptance of reactivity (ω-regular) properties on time-homogeneous Markov chains with general state spaces. The core technical result asserts that every reactivity property decomposes into finitely many almost-sure termination obligations into absorbing regions, which are shown to always exist; this allows every complete termination certificate to be lifted to a complete (or ε-complete) certificate for reactivity. The approach is instantiated on recent supermartingale results for countable state spaces, yielding the first such certificates for almost-sure and quantitative ω-regular properties.
Significance. If the central decomposition and existence claims hold, the work provides a reusable bridge from termination analysis to full ω-regular verification, with explicit completeness guarantees. The decomposition into absorbing-region termination obligations is a clean technical contribution that preserves soundness and completeness (up to ε in the quantitative case). The application to countable spaces produces novel certificates that were previously unavailable.
major comments (1)
- [Core technical result / existence theorem] Core technical result (abstract and the existence theorem): the claim that 'appropriate absorbing regions always exist on general state spaces' is load-bearing for completeness, yet the construction does not explicitly verify that the chosen regions are measurable with respect to the underlying σ-algebra. Without this, the induced subprocesses may fail to remain Markovian and the almost-sure termination obligations cannot be expressed via supermartingales in the required sense. A concrete counter-example construction or an additional regularity assumption (e.g., standard Borel) is needed to close the gap.
minor comments (2)
- [Introduction] The distinction between the fully general state-space theorem and the countable-space instantiation is not always sharp in the introduction; a short paragraph clarifying which results require only countability would improve readability.
- [Preliminaries] Notation for reactivity properties and absorbing regions is introduced gradually; a single consolidated table or definition block early in the paper would help readers track the decomposition.
Simulated Author's Rebuttal
We thank the referee for the careful reading and for highlighting the importance of measurability in the core construction. We address the single major comment below and will incorporate a clarification to strengthen the presentation.
read point-by-point responses
-
Referee: Core technical result (abstract and the existence theorem): the claim that 'appropriate absorbing regions always exist on general state spaces' is load-bearing for completeness, yet the construction does not explicitly verify that the chosen regions are measurable with respect to the underlying σ-algebra. Without this, the induced subprocesses may fail to remain Markovian and the almost-sure termination obligations cannot be expressed via supermartingales in the required sense. A concrete counter-example construction or an additional regularity assumption (e.g., standard Borel) is needed to close the gap.
Authors: We appreciate the referee drawing attention to this point. The absorbing regions in our decomposition are constructed as the sets of states from which there exists a scheduler ensuring that the ω-regular reactivity condition is satisfied by reaching designated absorbing components. Because the reactivity property is expressed via a measurable acceptance set on path space and the transition kernel is measurable, these regions arise as the greatest fixed points of measurable set-valued operators and are therefore measurable with respect to the given σ-algebra. Consequently the restricted subprocesses remain time-homogeneous Markov chains. We will add an explicit lemma (and its proof) in the revised manuscript that verifies this measurability directly from the construction, thereby confirming that the almost-sure termination obligations are well-defined and that the supermartingale certificates apply without change. No additional regularity assumption such as standard Borel spaces is required, and we do not believe a counter-example exists under the stated hypotheses. revision: yes
Circularity Check
No circularity: core existence result and decomposition are independent of the termination certificates being extended
full rationale
The paper's central technical result asserts that every reactivity property decomposes into almost-sure termination obligations into absorbing regions that always exist on general state spaces. This is presented as a new contribution that then enables extension of prior complete termination proof rules (applied to recent supermartingale results for countably infinite spaces). No equation or definition in the provided abstract reduces the existence claim or completeness to a self-referential fit, renaming, or load-bearing self-citation chain; the termination results are treated as independent inputs whose completeness is presupposed rather than redefined. The derivation chain therefore remains self-contained against external benchmarks for termination.
Axiom & Free-Parameter Ledger
axioms (1)
- domain assumption Time-homogeneous Markov chains admit absorbing regions for any reactivity property such that the property decomposes into almost-sure termination obligations.
Reference graph
Works this paper leans on
-
[1]
An introduction to probability theory and its applications, Volume 1 , author=. 1968 , publisher=
work page 1968
-
[2]
Rick Durrett , title =
- [3]
-
[4]
Meyn, Sean and Tweedie, Richard L. , year=. Markov Chains and Stochastic Stability , publisher=
-
[5]
Sean P. Meyn and R. L. Tweedie , journal =. Stability of Markovian Processes I: Criteria for Discrete-Time Chains , urldate =
- [6]
-
[7]
Lecture Notes on Limit Theorems for Markov Chain Transition Probability Functions , author=. 1968 , publisher=
work page 1968
-
[8]
A User's Guide to Measure Theoretic Probability , publisher=
Pollard, David , year=. A User's Guide to Measure Theoretic Probability , publisher=
-
[9]
Amir Pnueli , title =
- [10]
-
[11]
Principles of model checking , publisher =
Christel Baier and Joost. Principles of model checking , publisher =
-
[12]
From Spot 2.0 to Spot 2.10: What's New? , booktitle =
Alexandre Duret. From Spot 2.0 to Spot 2.10: What's New? , booktitle =
-
[13]
Automata, Logics, and Infinite Games:
-
[14]
Rupak Majumdar and V. R. Sathiyanarayana , title =. Proc
-
[15]
A new proof rule for almost-sure termination , journal =
Annabelle McIver and Carroll Morgan and Benjamin Lucien Kaminski and Joost. A new proof rule for almost-sure termination , journal =
-
[16]
Alessandro Abate and Mirco Giacobbe and Diptarko Roy , title =
-
[17]
Henzinger and Kaushik Mallik and Pouya Sadeghi and
Thomas A. Henzinger and Kaushik Mallik and Pouya Sadeghi and. Supermartingale Certificates for Quantitative Omega-Regular Verification and Control , booktitle =
-
[18]
Shmuel Safra , title =
-
[19]
Zohar Manna and Amir Pnueli , title =
-
[20]
Doblin, W. , title =. Annales scientifiques de l'\'Ecole Normale Sup\'erieure , pages =. 1940 , publisher =. doi:10.24033/asens.883 , zbl =
-
[21]
Jean-François Mertens and Ester Samuel-Cahn and Shmuel Zamir , journal =. Necessary and Sufficient Conditions for Recurrence and Transience of Markov Chains, in Terms of Inequalities , urldate =
- [22]
-
[23]
A. G. Pakes , journal =. Some Conditions for Ergodicity and Recurrence of Markov Chains , urldate =
-
[24]
Stefan Kiefer and Richard Mayr and Mahsa Shirmohammadi and Patrick Totzke , title =
-
[25]
Hannah Mertens and Joost. Accurately Computing Expected Visiting Times and Stationary Distributions in Markov Chains , booktitle =
-
[26]
Static Posterior Inference of Bayesian Probabilistic Programming via Polynomial Solving , journal =
Peixin Wang and Tengshun Yang and Hongfei Fu and Guanyan Li and C. Static Posterior Inference of Bayesian Probabilistic Programming via Polynomial Solving , journal =
-
[27]
Henzinger and Mathias Lechner and
Matin Ansaripour and Krishnendu Chatterjee and Thomas A. Henzinger and Mathias Lechner and. Learning Provably Stabilizing Neural Controllers for Discrete-Time Stochastic Systems , booktitle =
-
[28]
F. G. Foster , title =. The Annals of Mathematical Statistics , number =. 1953 , doi =
work page 1953
-
[29]
Kevin Batz and Benjamin Lucien Kaminski and Joost. Relatively complete verification of probabilistic programs: an expressive language for expectation-based reasoning , journal =
-
[30]
Deductive Proofs of Almost Sure Persistence and Recurrence Properties , booktitle =
Aleksandar Chakarov and Yuen. Deductive Proofs of Almost Sure Persistence and Recurrence Properties , booktitle =
-
[31]
Toru Takisaka and Yuichiro Oyabu and Natsuki Urabe and Ichiro Hasuo , title =
-
[32]
Krishnendu Chatterjee and Hongfei Fu and Amir Kafshdar Goharshady , title =
-
[33]
Aleksandar Chakarov and Sriram Sankaranarayanan , title =
- [34]
-
[35]
Stochastic invariants for probabilistic termination , booktitle =
Krishnendu Chatterjee and Petr Novotn. Stochastic invariants for probabilistic termination , booktitle =
-
[36]
Learning Control Policies for Stochastic Systems with Reach-Avoid Guarantees , booktitle =
- [37]
-
[38]
This is the moment for probabilistic loops , journal =
Marcel Moosbrugger and Miroslav Stankovic and Ezio Bartocci and Laura Kov. This is the moment for probabilistic loops , journal =
- [39]
-
[40]
The Probabilistic Termination Tool Amber
Moosbrugger, Marcel and Bartocci, Ezio and Katoen, Joost-Pieter and Kov \'a cs, Laura. The Probabilistic Termination Tool Amber. Formal Methods. 2021
work page 2021
-
[41]
Automated Termination Analysis of Polynomial Probabilistic Programs , booktitle =
Marcel Moosbrugger and Ezio Bartocci and Joost. Automated Termination Analysis of Polynomial Probabilistic Programs , booktitle =
-
[42]
ThanhVu Nguyen and Deepak Kapur and Westley Weimer and Stephanie Forrest , title =
-
[43]
Sipma and Zohar Manna , title =
Sriram Sankaranarayanan and Henny B. Sipma and Zohar Manna , title =
-
[44]
Michael D. Ernst and Jeff H. Perkins and Philip J. Guo and Stephen McCamant and Carlos Pacheco and Matthew S. Tschantz and Chen Xiao , title =. Sci. Comput. Program. , volume =
-
[45]
Practical Methods for Proving Program Termination , booktitle =
Michael Col. Practical Methods for Proving Program Termination , booktitle =
-
[46]
Linear Invariant Generation Using Non-linear Constraint Solving , booktitle =
Michael Col. Linear Invariant Generation Using Non-linear Constraint Solving , booktitle =
-
[47]
Stephen Prajna and Antonis Papachristodoulou and Pablo A. Parrilo , title =
-
[48]
Antonis Papachristodoulou and Stephen Prajna , title =
- [49]
-
[50]
Krishnendu Chatterjee and Hongfei Fu and Petr Novotn. Algorithmic Analysis of Qualitative and Quantitative Termination Problems for Affine Probabilistic Programs , journal =
-
[51]
Sheshansh Agrawal and Krishnendu Chatterjee and Petr Novotn. Lexicographic ranking supermartingales: an efficient approach to termination of probabilistic programs , journal =
-
[52]
Synthesis of Linear Ranking Functions , booktitle =
Michael Col. Synthesis of Linear Ranking Functions , booktitle =
-
[53]
Motwani and Maximilian Seeliger and
Krishnendu Chatterjee and Ehsan Kafshdar Goharshady and Mehrdad Karrabi and Harshit J. Motwani and Maximilian Seeliger and. Quantified Linear and Polynomial Arithmetic Satisfiability via Template-based Skolemization , booktitle =
-
[54]
Ali Asadi and Krishnendu Chatterjee and Hongfei Fu and Amir Kafshdar Goharshady and Mohammad Mahdavi , title =
-
[55]
Krishnendu Chatterjee and Amir Kafshdar Goharshady and Tobias Meggendorfer and Dorde Zikelic , title =
-
[56]
Sound and Complete Witnesses for Template-Based Verification of
Krishnendu Chatterjee and Amir Kafshdar Goharshady and Ehsan Kafshdar Goharshady and Mehrdad Karrabi and. Sound and Complete Witnesses for Template-Based Verification of
- [57]
-
[58]
She, Zhikun and Li, Haoyang and Xue, Bai and Zheng, Zhiming and Xia, Bican , year =
-
[59]
Representing polynomials by positive linear functions on compact convex polyhedra , volume =
Handelman, David , year =. Representing polynomials by positive linear functions on compact convex polyhedra , volume =
-
[60]
Equivalence and Similarity Refutation for Probabilistic Programs , journal =
Krishnendu Chatterjee and Ehsan Kafshdar Goharshady and Petr Novotn. Equivalence and Similarity Refutation for Probabilistic Programs , journal =
-
[61]
Differential cost analysis with simultaneous potentials and anti-potentials , booktitle =
-
[62]
Probabilistic Program Verification via Inductive Synthesis of Inductive Invariants , booktitle =
Kevin Batz and Mingshuai Chen and Sebastian Junges and Benjamin Lucien Kaminski and Joost. Probabilistic Program Verification via Inductive Synthesis of Inductive Invariants , booktitle =
-
[63]
Mathias Lechner and. Stability Verification in Stochastic Control Systems via Neural Network Supermartingales , booktitle =
-
[64]
Grigory Neustroev and Mirco Giacobbe and Anna Lukina , title =. 2025 , booktitle =
work page 2025
-
[65]
Henzinger and Mathias Lechner and
Krishnendu Chatterjee and Thomas A. Henzinger and Mathias Lechner and. A Learner-Verifier Framework for Neural Network Controllers and Certificates of Stochastic Systems , booktitle =
-
[66]
Compositional Policy Learning in Stochastic Control Systems with Formal Guarantees , booktitle =
-
[67]
Daniele Ahmed and Andrea Peruffo and Alessandro Abate , title =
-
[68]
Mirco Giacobbe and Daniel Kroening and Abhinandan Pal and Michael Tautschnig , title =. NeurIPS , year =
-
[69]
Alessandro Abate and Alec Edwards and Mirco Giacobbe and Hashan Punchihewa and Diptarko Roy , title =. Log. Methods Comput. Sci. , volume =
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.