pith. sign in

arxiv: 2601.18303 · v2 · pith:JJ7C4AL5new · submitted 2026-01-26 · 💻 cs.GT · cs.LO· cs.MA

Dicey Games: Shared Sources of Randomness in Distributed Systems

Pith reviewed 2026-05-16 11:19 UTC · model grok-4.3

classification 💻 cs.GT cs.LOcs.MA
keywords dicey gamesshared randomnessdistributed systemsmatching penniesgame theoryoptimal strategiesrandomness allocationcoordination
0
0 comments X

The pith

In the four-player Matching Pennies game, a team using only pairwise shared randomness sources can win with probability strictly above one quarter.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

A team of three players faces the Devil in a game where everyone chooses heads or tails simultaneously; the team wins only if all four choices match. Independent randomization gives the team a 1/8 win chance, while full shared randomness raises it to 1/2. The paper shows that restricting shared randomness to each pair of teammates still lets the team exceed a 1/4 win probability. This observation motivates the Dicey Games framework, which models how limited shared randomness sources can be allocated to improve coordination in distributed settings. The framework then characterizes when optimal strategies exist, how they can be represented, and how to allocate scarce randomness sources most effectively.

Core claim

Dicey Games formalize team-versus-adversary games in which coordination occurs solely through an allocation of shared randomness sources; for the concrete 4-player Matching Pennies instance, pairwise sharing already yields a win probability strictly greater than 1/4, and the framework supplies existence, representation, and complexity results for optimal strategies under arbitrary allocations.

What carries the argument

Dicey Games, the model in which a team’s coordination graph determines which subsets of players share independent randomness sources, and optimal mixed strategies are computed over the resulting joint distributions.

If this is right

  • Optimal strategies exist for any finite Dicey Game and any allocation of shared sources.
  • Optimal strategies admit a finite-support representation whose size is bounded by the number of players and sources.
  • Finding an optimal strategy is computationally hard in general, but polynomial for fixed numbers of players.
  • Given a budget of randomness sources, the problem of choosing the best allocation can be solved by enumeration over graphs.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same pairwise-sharing pattern may improve coordination in other simultaneous-move distributed protocols that already use limited shared randomness.
  • Extending the model to sequential games or noisy channels would test whether the >1/4 threshold generalizes.
  • Concrete small instances could be implemented on hardware with limited entanglement or shared seeds to measure empirical gains.

Load-bearing premise

The team’s only coordination mechanism is the pre-assigned shared randomness sources; all choices are made simultaneously with no further communication.

What would settle it

An explicit calculation or linear-program solution for the 4-player Matching Pennies instance that reports a maximum win probability of at most 1/4 when each pair of teammates shares one private random bit.

Figures

Figures reproduced from arXiv: 2601.18303 by K. S. Thejaswini, L\'eonard Brice, Thomas A. Henzinger.

Figure 1
Figure 1. Figure 1: A strategy to win with probability 1/6 matching-pennies case (Ada vs. Devil) and the max min value be￾comes 1/2. We now pose an illustrative question that motivates the broader class of games that we call dicey games, addressed in this work. Triangular matching pennies with the Devil What is the optimal strategy—and corresponding max min value—for achieving an all-Heads or all-Tails outcome in the matching… view at source ↗
Figure 2
Figure 2. Figure 2: An optimal collective strategy 1.2 Motivation In the context of zero-sum games, dicey games represent a natu￾ral intermediate notion between the independent randomisation studied by Cournot [19], von Neumann and Morgenstern [39], and Nash [33], and the shared randomness inherent to correlated equi￾libria [3, 5]. In correlated equilibria, all players can coordinate using one common source of randomness, whi… view at source ↗
Figure 3
Figure 3. Figure 3: The collective strategy 𝜎¯ 0 . . . • else, we have 𝜇(𝑎¯) = 𝑥𝑛+1. For example, in a matching pennies game in which all players choose between the actions Heads (written H) and Tails (written T), the fact that the team gets payoff 1 if all actions match, and 0 other￾wise, can be described by the list (H)𝑝∈Π∪{𝔇} , 1  , [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Applying Carathéodory’s theorem Geometrically, if HHH denotes the part of the cube of rolls [0, 1] 3 in which the three players choose H, then 𝑓 (𝑥) is the area of the slice HHH ∩ ({𝑥} × [0, 1] 2 ), and similarly for 𝑔. Let us note that the probability of all players choosing H is then ∫ 𝑓 , and that the probability of all choosing T is ∫ 𝑔 (without more precision, integrals are considered on the whole dom… view at source ↗
Figure 5
Figure 5. Figure 5: The collective strategy 𝜑1 (𝜎¯ 0 ) Proposition 3.6. The collective strategy 𝜎¯ ′ has a value at least as large as the collective strategy 𝜎¯. Proof. We prove this proposition by showing the inequalities: P [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The collective strategy 𝜑2 ◦ 𝜑1 (𝜎¯ 0 ) 𝑥 𝑦 H T (a) Ada’s strategy 𝑧 𝑦 T H (b) Bertrand’s strategy 𝑥 𝑧 H T (c) Claude’s strategy 𝑥 𝑦 𝑧 (d) The sets HHH and TTT [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The collective strategy 𝜎¯ ★ The second one is analogous. Then, the value of the collective strategy 𝜎¯ ′ is: min  P [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The set of Fritz John points 4.1 A system of polynomial inequalities To formally define a system of inequalities analogous to the one presented at the end of Section 3.1, we first define some notation. Throughout this section, we assume, up to renaming them, that our dice form the set Δ = {1, . . . , 𝑛}. We let 𝑘 = |𝐴𝔇|. Definition 6 (Strategy scheme, 𝑎¯¯-strategy). Let D be a dicey game. A strategy scheme… view at source ↗
Figure 9
Figure 9. Figure 9: A few more examples Here, one might expect that the optimal strategy is similar to the one that was found for triangular matching pennies: one threshold 𝜆 obtained by some intricate polynomial equation, and each player choosing Heads if all the rolls they see are above that threshold, and Tails otherwise. The reality is quite different. The max min value of this game is 1/4, which is obtained by the follow… view at source ↗
read the original abstract

Consider a 4-player version of Matching Pennies where a team of three players competes against the Devil. Each player simultaneously says "Heads" or "Tails". The team wins if all four choices match; otherwise the Devil wins. If all team players randomise independently, they win with probability 1/8; if all players share a common source of randomness, they win with probability 1/2. What happens when each pair of team players shares a source of randomness? Can the team do better than win with probability 1/4? The surprising (and nontrivial) answer is yes! We introduce Dicey Games, a formal framework motivated by the study of distributed systems with shared sources of randomness (of which the above example is a specific instance). We characterise the existence, representation and computational complexity of optimal strategies in Dicey Games, and we study the problem of allocating limited sources of randomness optimally within a team.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper introduces Dicey Games, a formal framework for distributed systems in which agents coordinate via allocated shared randomness sources. It analyzes a 4-player Matching Pennies instance in which a team of three players faces the Devil, showing that pairwise shared randomness sources allow the team to win with probability strictly greater than 1/4 (compared with 1/8 for independent randomness and 1/2 for a single shared source). The authors characterize existence, representation, and computational complexity of optimal strategies within the framework and study the problem of optimally allocating a limited number of randomness sources among team members.

Significance. If the characterizations hold, the work supplies a new, self-contained model for reasoning about partial coordination in distributed systems, with direct relevance to multi-agent game theory and distributed computing. The concrete demonstration that pairwise sharing exceeds the 1/4 threshold is nontrivial and falsifiable. The paper provides explicit characterizations of optimal strategies together with complexity results, which constitute a reusable technical contribution.

major comments (2)
  1. [§3.2] §3.2, the explicit strategy for the pairwise-sharing case: the claimed winning probability >1/4 is stated but the derivation of the exact value and the verification that it is optimal are not cross-referenced to the general representation theorem in §4.1; without this link the central claim rests on an unverified calculation.
  2. [§5] §5, the complexity result for optimal allocation: the reduction establishing NP-completeness assumes that the number of randomness sources is part of the input, yet the paper does not discuss whether the hardness persists when the number of sources is fixed (a regime relevant to the motivating example).
minor comments (2)
  1. [Abstract] The abstract asserts a 'surprising (and nontrivial) answer is yes' but does not state the concrete probability achieved by the pairwise strategy; adding this value would improve readability.
  2. [§2] Notation for the shared-randomness graph (vertices = players, edges = shared sources) is introduced without a small illustrative diagram; a figure would clarify the pairwise-sharing configuration used in the main example.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive evaluation and the precise comments. We address each major point below and will revise the manuscript accordingly.

read point-by-point responses
  1. Referee: [§3.2] §3.2, the explicit strategy for the pairwise-sharing case: the claimed winning probability >1/4 is stated but the derivation of the exact value and the verification that it is optimal are not cross-referenced to the general representation theorem in §4.1; without this link the central claim rests on an unverified calculation.

    Authors: We agree that an explicit link is needed. In the revised version we will add a direct cross-reference from §3.2 to the representation theorem in §4.1, together with a short derivation showing how the pairwise strategy yields a winning probability strictly greater than 1/4 and confirming optimality via the general result. revision: yes

  2. Referee: [§5] §5, the complexity result for optimal allocation: the reduction establishing NP-completeness assumes that the number of randomness sources is part of the input, yet the paper does not discuss whether the hardness persists when the number of sources is fixed (a regime relevant to the motivating example).

    Authors: The referee is correct that the reduction treats the number of sources as part of the input. We will add a short discussion in §5 noting that, for any fixed constant number of sources, all possible allocations can be enumerated in polynomial time (though evaluating the value of each allocation remains hard); this regime is directly relevant to the 4-player motivating example. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected in derivation chain

full rationale

The paper introduces Dicey Games as a new formal framework for modeling distributed systems with shared randomness sources, using the 4-player Matching Pennies example to illustrate that pairwise shared randomness yields win probability strictly above 1/4. It then characterizes existence, representation, and computational complexity of optimal strategies within this framework. No equations, fitted parameters, or self-citations are visible that reduce any claimed result to its own inputs by construction; the central claims rest on the newly defined model and its internal analysis rather than self-definitional loops, renamed empirical patterns, or load-bearing prior self-citations. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The framework rests on standard simultaneous-move game assumptions and introduces the new concept of Dicey Games to capture restricted shared randomness; no free parameters or invented physical entities are mentioned.

axioms (1)
  • standard math Players choose simultaneously and payoffs depend only on the tuple of choices.
    Implicit in the definition of the matching-pennies payoff structure.
invented entities (1)
  • Dicey Games no independent evidence
    purpose: Formal model of distributed games with arbitrary patterns of shared randomness sources.
    Newly defined to study the allocation and strategy problems described.

pith-pipeline@v0.9.0 · 5474 in / 1146 out tokens · 40849 ms · 2026-05-16T11:19:30.183099+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Randomise Alone, Reach as a Team

    cs.GT 2026-03 unverdicted novelty 7.0

    In concurrent graph games with distributed private randomness, memoryless strategies decide threshold reachability (NP-hard) and almost-sure reachability is NP-complete; IRATL extends ATL for probability thresholds wi...

Reference graph

Works this paper leans on

47 extracted references · 47 canonical work pages · cited by 1 Pith paper

  1. [1]

    Wooldridge

    Alessandro Abate, Julian Gutierrez, Lewis Hammond, Paul Harrenstein, Marta Kwiatkowska, Muhammad Najib, Giuseppe Perelli, Thomas Steeples, and Michael J. Wooldridge. 2021. Rational verification: game-theoretic verification of multi-agent systems.Appl. Intell.51, 9 (2021), 6569–6584. doi:10.1007/S10489- 021-02658-Y

  2. [2]

    Ricardo Alonso and Odilon Câmara. 2016. Persuading voters.American Economic Review106, 11 (2016), 3590–3605

  3. [3]

    Robert J. Aumann. 1974. Subjectivity and correlation in randomized strategies. Journal of Mathematical Economics1, 1 (1974), 67–96. doi:10.1016/0304-4068(74) 90037-8

  4. [4]

    Robert J Aumann. 1974. Subjectivity and correlation in randomized strategies. Journal of mathematical economics1, 1 (1974), 67–96

  5. [5]

    R. J. Aumann and J. H. Dreze. 1974. Cooperative games with coalition structures. Int. J. Game Theory3, 4 (Dec. 1974), 217–237. doi:10.1007/BF01766876

  6. [6]

    Saugata Basu, Richard Pollack, and Marie-Françoise Roy. 1994. On the Com- binatorial and Algebraic Complexity of Quantifier Elimination. InFOCS. IEEE Computer Society, 632–641. doi:10.1109/SFCS.1994.365728

  7. [7]

    Saugata Basu, Richard Pollack, and Marie-Françoise Roy. 1996. On the combi- natorial and algebraic complexity of quantifier elimination.Journal of the ACM (JACM)43, 6 (1996), 1002–1045

  8. [8]

    Saugata Basu, Richard Pollack, and Marie-Françoise Roy. 1997. On Computing a Set of Points Meeting Every Cell Defined by a Family of Polynomials on a Variety. Journal of Complexity13, 1 (1997), 28–37. doi:10.1006/jcom.1997.0434

  9. [9]

    2006.Algorithms in Real Algebraic Geometry(2nd ed.)

    Saugata Basu, Richard Pollack, and Marie-Françoise Roy. 2006.Algorithms in Real Algebraic Geometry(2nd ed.). Springer. doi:10.1007/978-3-540-30140-0

  10. [10]

    Bazaraa, Hanif D

    Mokhtar S. Bazaraa, Hanif D. Sherali, and C. M. Shetty. 2006.Nonlinear Pro- gramming: Theory and Algorithms(3rd ed.). John Wiley & Sons, Hoboken, New Jersey

  11. [11]

    Mihir Bellare, Wei Dai, and Phillip Rogaway. 2020. Reimagining Secret Sharing: Creating a Safer and More Versatile Primitive by Adding Authenticity, Correcting Errors, and Reducing Randomness Requirements. InProceedings on Privacy En- hancing Technologies (PoPETs), Vol. 2020. 461–490. doi:10.2478/popets-2020-0081

  12. [12]

    Dirk Bergemann and Stephen Morris. 2016. Bayes correlated equilibrium and the comparison of information structures in games.Theoretical Economics11, 2 (2016), 487–522

  13. [13]

    Dirk Bergemann and Stephen Morris. 2019. Information design: A unified per- spective.Journal of Economic Literature57, 1 (2019), 44–95

  14. [14]

    Markus Bläser, Julian Dörfler, Maciej Liśkiewicz, and Benito van der Zander. 2024. The Existential Theory of the Reals with Summation Operators. InISAAC 2024 (LIPIcs, Vol. 322). Schloss Dagstuhl – Leibniz-Zentrum für Informatik, 13:1–13:19. doi:10.4230/LIPIcs.ISAAC.2024.13

  15. [15]

    Florian Brandl, Felix Brandt, and Hans Georg Seedig. 2016. Consistent probabilis- tic social choice.Econometrica84, 5 (2016), 1839–1880. doi:10.3982/ECTA13322

  16. [16]

    Felix Brandt. 2019. Collective choice lotteries. InStudies in Economic Design. Springer, 51–73. doi:10.1007/978-3-030-18050-8_9

  17. [17]

    Eric Budish, Yeon-Koo Che, Fuhito Kojima, and Paul Milgrom. 2013. Designing random allocation mechanisms: Theory and applications.American Economic Review103, 2 (2013), 585–623. doi:10.1257/aer.103.2.585

  18. [18]

    George E. Collins. 1974. Quantifier elimination for real closed fields by cylindrical algebraic decomposition–preliminary report.SIGSAM Bull.8, 3 (Aug. 1974), 80–90. doi:10.1145/1086837.1086852

  19. [19]

    1838.Recherches sur les principes mathématiques de la théorie des richesses

    Antoine Augustin Cournot. 1838.Recherches sur les principes mathématiques de la théorie des richesses. L. Hachette, Paris. English trans: Researches into the Mathematical Principles of the Theory of Wealth (1897)

  20. [20]

    Leonardo de Moura and Nikolaj Bjørner. 2008. Z3: an efficient SMT solver. Tools and Algorithms for the Construction and Analysis of Systems4963, 337–340. doi:10.1007/978-3-540-78800-3_24

  21. [21]

    Yossi Gilad, Rotem Hemo, Silvio Micali, Georgios Vlachos, and Nickolai Zel- dovich. 2017. Algorand: Scaling Byzantine Agreements for Cryptocurrencies. InProceedings of the 26th Symposium on Operating Systems Principles (SOSP ’17). ACM, 51–68. doi:10.1145/3132747.3132757

  22. [22]

    2010.A Primer on Pseudorandom Generators

    Oded Goldreich. 2010.A Primer on Pseudorandom Generators. University Lecture Series, Vol. 55. American Mathematical Society. doi:10.1090/ulect/055

  23. [23]

    Friedrich August Hayek. 1945. The use of knowledge in society.The American economic review35, 4 (1945), 519–530

  24. [24]

    J.V Howard. 1992. A social choice rule and its implementation in perfect equi- librium.Journal of Economic Theory56, 1 (1992), 142–159. doi:10.1016/0022- 0531(92)90073-Q

  25. [25]

    Fritz John. 1948. Extremum Problems with Inequalities as Subsidiary Conditions. InStudies and Essays Presented to R. Courant on his 60th Birthday, K. O. Friedrichs, O. E. Neugebauer, and J. J. Stoker (Eds.). Interscience Publishers, New York, 187–204

  26. [26]

    Kai-Uwe Kühn. 2001. Fighting collusion–Regulation of communication between competitors.Economic Policy16, 32 (2001), 168–204

  27. [27]

    Marta Kwiatkowska, Gethin Norman, David Parker, and Gabriel Santos. 2018. Equilibria-based Probabilistic Model Checking for Concurrent Stochastic Games. Formal Methods in System Design53, 3 (2018), 261–294

  28. [28]

    I. E. Leonard and J. E. Lewis. 2016.Geometry of Convex Sets(1st ed.). Wiley- Blackwell, Hoboken, New Jersey, United States. 336 pages

  29. [29]

    Mangasarian and Stan Fromovitz

    Olvi L. Mangasarian and Stan Fromovitz. 1967. The Fritz John Necessary Op- timality Conditions in the Presence of Equality and Inequality Constraints.J. Math. Anal. Appl.17, 1 (1967), 37–47. doi:10.1016/0022-247X(67)90163-1

  30. [30]

    John Maynard Smith and George R. Price. 1973. The Logic of Animal Conflict. Nature246, 5427 (1973), 15–18. https://royalsocietypublishing.org/doi/10.1098/ rstb.2021.0509

  31. [31]

    Saptarshi Mukherjee and Hans Peters. 2022. Self-implementation of social choice correspondences in Nash equilibrium.Social Choice and Welfare59, 4 (November 2022), 1009–1028. doi:10.1007/s00355-022-01420-8

  32. [32]

    Roger B. Myerson. 1999. Nash Equilibrium and the History of Economic Theory. Journal of Economic Literature37, 3 (September 1999), 1067–1082

  33. [33]

    John F. Nash. 1950. Equilibrium points in 𝑛-person games.Proceedings of the National Academy of Sciences36, 1 (1950), 48–49. doi:10.1073/pnas.36.1.48

  34. [34]

    Marcus Schaefer. 2013. Realizability of Graphs and Linkages. 461-482 pages. doi:10.1007/978-1-4614-0110-0_24

  35. [35]

    Gustavus J. Simmons. 1984. The Prisoners’ Problem and the Subliminal Channel. InAdvances in Cryptology, Proceedings of CRYPTO ’83 (Lecture Notes in Computer Science, Vol. 196). Springer, 51–67. doi:10.1007/978-1-4684-4730-9_5

  36. [36]

    Fischer, and Bryan Ford

    Ewa Syta, Philipp Jovanovic, Eleftherios Kokoris-Kogias, Nicolas Gailly, Linus Gasser, Ismail Khoffi, Michael J. Fischer, and Bryan Ford. 2017. Scalable Bias- Resistant Distributed Randomness. In2017 IEEE Symposium on Security and Privacy (SP). IEEE, 444–460. doi:10.1109/SP.2017.34

  37. [37]

    2010.Stochastic multiplayer games: theory and algorithms

    Michael Ummels. 2010.Stochastic multiplayer games: theory and algorithms. Ph. D. Dissertation. RWTH Aachen, Germany

  38. [38]

    2010.Information and learning in markets: the impact of market microstructure

    Xavier Vives. 2010.Information and learning in markets: the impact of market microstructure. Princeton University Press

  39. [39]

    1944.Theory of Games and Economic Behavior

    John von Neumann and Oskar Morgenstern. 1944.Theory of Games and Economic Behavior. Princeton University Press, Princeton, NJ. A Useful theorems We state here some geometry theorems that we use in our proofs. Note that, in the references cited, the theorem that is proven is often more general: we state it here in a form that fits our purposes. Theorem A.1...

  40. [40]

    (1−𝑏 ℭ 11) +𝜆 1𝜆2 (1−𝜆 3) (1−𝑏 𝔄

  41. [41]

    (1−𝑏 ℭ 12) +𝜆 1 (1−𝜆 2)𝜆3 (1−𝑏 𝔄

  42. [42]

    (1−𝑏 ℭ 11) +𝜆 1 (1−𝜆 2) (1−𝜆 3) (1−𝑏 𝔄

  43. [43]

    (1−𝑏 ℭ 12) + (1−𝜆 1)𝜆2𝜆3 (1−𝑏 𝔄

  44. [44]

    (1−𝑏 ℭ 21) + (1−𝜆 1)𝜆2 (1−𝜆 3) (1−𝑏 𝔄

  45. [45]

    (1−𝑏 ℭ 22) + (1−𝜆 1) (1−𝜆 2)𝜆3 (1−𝑏 𝔄

  46. [46]

    (1−𝑏 ℭ 21) + (1−𝜆 1) (1−𝜆 2) (1−𝜆 3) (1−𝑏 𝔄

  47. [47]

    C).Let D be a dicey game

    (1−𝑏 ℭ 22) (8) ∀𝑖∈ {1,2,3},0≤𝜆 𝑖 ≤1(9) C Proof of Theorem 3.7 Theorem 3.7(App. C).Let D be a dicey game. Let 𝑘 be the number of actions available for the Devil. Then, for every collective strategy, there exists a𝑘-grid strategy whose value is at least as large. Proof. This proof follows the structure of that of Lemma 3.3 and generalises it: it first defin...