pith. sign in

arxiv: 2605.21424 · v1 · pith:4FAL2HX3new · submitted 2026-05-20 · 🧮 math.PR

A dice game, a multinomial walk, and the inverted Dirichlet distribution

Gunther Leobacher , Alexander Steinicke This is my paper

Pith reviewed 2026-05-21 03:02 UTC · model grok-4.3

classification 🧮 math.PR
keywords dice gamemultinomial walkinverted Dirichlet distributionregularized beta functionwinning probabilityasymptotic analysisnegative multinomial
0
0 comments X

The pith

A dice game modeled as a multinomial walk has winning probabilities that vary monotonically with its parameters.

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

The paper presents a dice game in which players take turns rolling until a target sequence appears and shows that this process is equivalent to a multinomial walk. Using conjugacy between the negative multinomial and inverted Dirichlet distributions, it obtains exact expressions for each player's winning and losing probabilities. A separate monotonicity result for the regularized beta function then implies that these winning probabilities are monotone in the game parameters. The work further determines the limiting behavior when one or more parameters become large and gives an expression for the probability that a given player finishes last.

Core claim

For the generalized dice game the probability that a designated player wins equals a ratio of inverted Dirichlet integrals; this probability is monotone in each parameter because the regularized beta function is monotone in its first two arguments when the third argument is fixed.

What carries the argument

The conjugacy relation between the negative multinomial distribution and the inverted Dirichlet distribution, which converts the probability that the multinomial walk never hits a losing barrier into a ratio of beta integrals.

If this is right

  • The probability that any given player wins changes monotonically when the success probabilities on the die faces are altered.
  • When one or more game parameters tend to infinity the winning probabilities converge to explicit limits that can be read off the beta function.
  • The probability that a player is last admits an analogous integral representation and inherits the same monotonicity.
  • Exact closed-form probabilities replace the need for Monte-Carlo simulation of the game.

Where Pith is reading between the lines

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

  • The same conjugacy may let researchers compute absorption probabilities for other barrier problems driven by multinomial or negative-multinomial steps.
  • The monotonicity of the regularized beta function could be used directly in statistical comparisons that involve incomplete beta ratios.
  • The asymptotic limits suggest how large-sample versions of the game behave when the die becomes biased or the number of faces grows.

Load-bearing premise

The successive dice rolls are independent and follow the fixed probability vector that defines the multinomial step.

What would settle it

An explicit numerical evaluation of the winning probability for two players and three faces that violates the claimed monotonicity in one of the probability parameters.

Figures

Figures reproduced from arXiv: 2605.21424 by Alexander Steinicke, Gunther Leobacher.

Figure 1
Figure 1. Figure 1: The game board The question is: Which of the players is/are most likely to win? For this, let Nℓ denote the number of rounds played until player ℓ reaches their goal. Then Nℓ − nℓ follows a negative binomial distribution with parameters nℓ and nℓ 36 , since it is the number of failures before the nℓ-th success with the probability of success equal 1 arXiv:2605.21424v1 [math.PR] 20 May 2026 [PITH_FULL_IMAG… view at source ↗
Figure 2
Figure 2. Figure 2: Any walk reaching (n1, y) before (x, n2) crosses the line x + y = n1 + n2 − 1 in a point x ′ , y′ where y ′ < n2. the probability of player 1 winning is π1(n1, n2) := nX2−1 k=0  n1 + n2 − 1 k  p k 2 p n1+n2−1−k 1 , the probability that a random variable X, having binomial distribution with parameters n1 + n2 − 1 and p2, is smaller than n2. It is known that the median of X is unique if p2 is rational, [26… view at source ↗
Figure 3
Figure 3. Figure 3: Note that the winning probability for player 1 is the hitting probability of one face of an m-dimensional cuboid. For m = 3 there can be paths exiting from different faces of the cuboid but still hitting the plane x + y + z = n1 + n2 + n3 − 1 in the same point. stronger result, since it actually shows that π1(n1, n2) is strictly increasing in n2 (and then the claim follows by symmetry, π1(n1, n2) = π2(n1, … view at source ↗
read the original abstract

We consider a simple dice game, which leads to an intriguing study of multinomial walks, with surprising and seemingly paradoxical properties. The winning and losing probabilities of a general version of the game are investigated via conjugacy relations between Gamma and Poisson distributions, as well as between negative multinomial and inverted Dirichlet distributions. We show a monotonicity property of the regularized beta function, which implies a monotonicity property of the winning probability. Furthermore, the asymptotic behavior of the game for one or several parameters of the game tending to infinity is analyzed, as well as the probability of being last in the game.

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

0 major / 3 minor

Summary. The paper models a dice game as a multinomial walk stopped when any coordinate reaches a threshold, derives winning probabilities via conjugacy between the negative multinomial distribution and the inverted Dirichlet (equivalently, via Gamma-Poisson relations), proves a monotonicity property of the regularized incomplete beta function that transfers to monotonicity of the winning probability in the game parameters, and analyzes the asymptotic regime when one or more parameters tend to infinity together with the probability that a given player is last.

Significance. The central contribution is the explicit transfer of a monotonicity result for the regularized beta function to the winning probability of the game, together with the asymptotic analysis. The conjugacy approach is standard in the field but is applied here to yield concrete, computable expressions and a clean monotonicity statement. The work also supplies explicit formulas for the probability of being last. These elements are of interest to researchers working on multivariate discrete processes and their continuous limits.

minor comments (3)
  1. [§2.2] §2.2, after Eq. (8): the mapping from the stopping time of the multinomial walk (first time any coordinate hits its threshold) to the shape parameters of the inverted Dirichlet is stated but the boundary case when two coordinates hit simultaneously is not explicitly ruled out or handled; a short remark confirming that this event has probability zero under the continuous embedding would remove any ambiguity.
  2. [Figure 1] Figure 1: the caption does not indicate whether the plotted curves are exact or Monte-Carlo; adding this information would improve reproducibility.
  3. [§4] §4, the statement of the asymptotic result for the probability of being last: the limit is given but the rate of convergence is not discussed; a brief remark on the order of the error term would strengthen the claim.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our manuscript and the recommendation for minor revision. We appreciate the recognition that the central contribution lies in transferring the monotonicity property of the regularized incomplete beta function to the winning probabilities of the dice game, together with the asymptotic analysis and explicit formulas for the probability of being last.

Circularity Check

0 steps flagged

No significant circularity; derivations use external conjugacy and independent monotonicity proof

full rationale

The paper models the dice game as a multinomial walk and invokes standard conjugacy relations between negative multinomial and inverted Dirichlet (and Gamma-Poisson) distributions to obtain winning probabilities. It then proves a monotonicity property of the regularized beta function directly and transfers it to the game probabilities. No equation or claim reduces by construction to a fitted parameter, self-definition, or load-bearing self-citation chain; the conjugacy is treated as an established external fact and the monotonicity result is a new, self-contained argument. The derivation chain is therefore independent of its own outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

The analysis rests on standard probability-theory conjugacy relations and the definition of the regularized beta function; no free parameters or invented entities are introduced.

axioms (3)
  • standard math Conjugacy between Gamma and Poisson distributions
    Invoked to obtain winning and losing probabilities for the game.
  • standard math Conjugacy between negative multinomial and inverted Dirichlet distributions
    Used to express the winning probabilities of the general game.
  • standard math Monotonicity of the regularized beta function in its parameters
    Proved in the paper and applied to the winning probability.

pith-pipeline@v0.9.0 · 5621 in / 1262 out tokens · 40190 ms · 2026-05-21T03:02:04.663813+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.

Reference graph

Works this paper leans on

44 extracted references · 44 canonical work pages

  1. [1]

    An infinite mixture of inverted D irichlet distributions

    Bdiri, T., and Bouguila, N. An infinite mixture of inverted D irichlet distributions. In Neural Information Processing\/ (Berlin, Heidelberg, 2011), B.-L. Lu, L. Zhang, and J. Kwok, Eds., Springer Berlin Heidelberg, pp. 71--78

    work page 2011

  2. [2]

    Positive vectors clustering using inverted D irichlet finite mixture models

    Bdiri, T., and Bouguila, N. Positive vectors clustering using inverted D irichlet finite mixture models. Expert Systems with Applications 39 , 2 (2012), 1869--1882

    work page 2012

  3. [3]

    Bayesian learning of inverted D irichlet mixtures for SVM kernels generation

    Bdiri, T., and Bouguila, N. Bayesian learning of inverted D irichlet mixtures for SVM kernels generation. Neural Computing and Applications 23 , 5 (Oct 2013), 1443--1458

    work page 2013

  4. [4]

    Berman, A., and Plemmons, R. J. Nonnegative matrices in the mathematical sciences . Classics in applied mathematics 9. Society for Industrial and Applied Mathematics, 1987

    work page 1987

  5. [5]

    M., and Smith, A

    Bernardo, J. M., and Smith, A. F. M. Bayesian Theory . John Wiley & Sons Ltd., New York, 2000

    work page 2000

  6. [6]

    Probability and Measure

    Billingsley, P. Probability and Measure . Wiley Series in Probability and Statistics. John Wiley & Sons Ltd., New York, 1986

    work page 1986

  7. [7]

    P., Evnine, J., and Gibbs, S

    Boyle, P. P., Evnine, J., and Gibbs, S. Numerical evaluation of multivariate contingent claims. The Review of Financial Studies 2 , 2 (1989), 241--250

    work page 1989

  8. [8]

    Chen, R.-R., Chung, S.-L., and Yang, T. T. Option pricing in a multi-asset, complete market economy. Journal of Financial and Quantitative Analysis 37 , 4 (December 2002), 649--666

    work page 2002

  9. [9]

    Fang, K.-T., Kotz, S., and Ng, K. W. Symmetric Multivariate and Related Distributions . Chapman and Hall/CRC, New York, 1990

    work page 1990

  10. [10]

    On the inverted D irichlet distribution

    Ghorbel, M. On the inverted D irichlet distribution. Communications in Statistics - Theory and Methods 39 , 1 (2009), 21--37

    work page 2009

  11. [11]

    C., and Maher, M

    Gittins, J. C., and Maher, M. J. A generalised incomplete B eta function and its application to multi-line stock control. Journal of Applied Probability 10 , 4 (1973), 748--760

    work page 1973

  12. [12]

    Bounds for median and 50 percetage point of binomial and negative binomial distribution

    G \"o b, R. Bounds for median and 50 percetage point of binomial and negative binomial distribution. Metrika 41 , 1 (Dec 1994), 43--54

    work page 1994

  13. [13]

    Convergence from discrete- to continuous-time contingent claims prices

    He, H. Convergence from discrete- to continuous-time contingent claims prices. The Review of Financial Studies 3 , 4 (04 1990), 523--546

    work page 1990

  14. [14]

    Two monotonicity results for B eta distribution functions

    Hornik, K. Two monotonicity results for B eta distribution functions. Entropy 26 , 11 (2024)

    work page 2024

  15. [15]

    L., Kemp, A

    Johnson, N. L., Kemp, A. W., and Kotz, S. Univariate Discrete Distributions . John Wiley & Sons, Inc., New York, 1992

    work page 1992

  16. [16]

    L., Kotz, S., and Balakrishnan, N

    Johnson, N. L., Kotz, S., and Balakrishnan, N. Continuous Univariate Distributions , 2nd ed., vol. 1. John Wiley & Sons Ltd., New York, 1994

    work page 1994

  17. [17]

    L., Kotz, S., and Balakrishnan, N

    Johnson, N. L., Kotz, S., and Balakrishnan, N. Discrete Multivariate Distributions . John Wiley & Sons Ltd., New York, 1997

    work page 1997

  18. [18]

    Mean, median and mode in binomial distributions

    Kaas, R., and Buhrman, J. Mean, median and mode in binomial distributions. Statistica Neerlandica 34 , 1 (1980), 13--18

    work page 1980

  19. [19]

    Variational bayesian inference for finite inverted D irichlet mixture model and its application to object detection

    Lai, Y., Ping, Y., He, W., Wang, B., Wang, J., and Zhang, X. Variational bayesian inference for finite inverted D irichlet mixture model and its application to object detection. Chinese Journal of Electronics 27 , 3 (2018), 603--610

    work page 2018

  20. [20]

    Against All Odds? -- a surprisingly not so easy dice game

    Lerchenberger, E., and Oberhuber, C. Against All Odds? -- a surprisingly not so easy dice game . Manuscript submitted for publication, 2026

    work page 2026

  21. [21]

    H., Balakrishnan, N., and Bae, S

    Ling, M. H., Balakrishnan, N., and Bae, S. J. On the application of inverted D irichlet distribution for reliability inference of completely censored components with dependence structure. Computers & Industrial Engineering 196\/ (2024), 110452

    work page 2024

  22. [22]

    On the generalised inverted Dirichlet distribution

    Lingmnwah, G. On the generalised inverted Dirichlet distribution . Demonstratio Mathematica 9 , 3 (1976), 119--130

    work page 1976

  23. [23]

    Morris, K. W. A note on direct and inverse binomial sampling. Biometrika 50 , 3-4 (12 1963), 544--545

    work page 1963

  24. [24]

    D irichlet and related distributions: Theory, methods and applications

    Ng, K., Tian, G., and Tang, M. D irichlet and related distributions: Theory, methods and applications . Wiley, 2011

    work page 2011

  25. [25]

    Norris, J. R. Markov Chains . Cambridge University Press, Cambridge, UK, 1997

    work page 1997

  26. [26]

    Uniqueness of a median of a binomial distribution with rational probability

    Nowakowski, S. Uniqueness of a median of a binomial distribution with rational probability. Advances in Mathematics: Scientific Journal 10 , 4 (2021), 1951--1958

    work page 2021

  27. [27]

    Multivariate beta distributions and independence properties of the W ishart distribution

    Olkin, I., and Rubin, H. Multivariate beta distributions and independence properties of the W ishart distribution. The Annals of Mathematical Statistics 35 , 1 (1964), 261--269

    work page 1964

  28. [28]

    F., Ferreira, J

    Otto, A. F., Ferreira, J. T., Bekker, A., Punzo, A., and Tomarchio, S. D. A refreshing take on the inverted D irichlet via a mode parameterization with some statistical illustrations. Journal of the Korean Statistical Society 54 , 1 (Mar 2025), 314--341

    work page 2025

  29. [29]

    Tables of the incomplete -function

    Pearson (ed.), K. Tables of the incomplete -function. C omputed by the S taff of the D epartment of A pplied S tatistics, university of L ondon, U niversity C ollege

    work page

  30. [30]

    Some asymptotic properties of the negative binomial distribution

    Pessin, V. Some asymptotic properties of the negative binomial distribution. Ann. Math. Statist. 32 , 3 (1961), 922--923

    work page 1961

  31. [31]

    Phillips, P. C. The characteristic function of the D irichlet and multivariate F distributions. Cowles Foundation Discussion Papers 865, Cowles Foundation for Research in Economics, Yale University, Jan 1988

    work page 1988

  32. [32]

    M -matrix characterizations

    Plemmons, R. M -matrix characterizations. I ---nonsingular M -matrices. Linear Algebra and its Applications 18 , 2 (1977), 175--188

    work page 1977

  33. [33]

    \"U ber eine A ufgabe der W ahrscheinlichkeitsrechnung betreffend die I rrfahrt im S tra ennetz

    P \'o lya, G. \"U ber eine A ufgabe der W ahrscheinlichkeitsrechnung betreffend die I rrfahrt im S tra ennetz. Mathematische Annalen 84 , 1 (1921), 149--160

    work page 1921

  34. [34]

    Die U ngleichungen von V ietoris

    Raab, W. Die U ngleichungen von V ietoris. Monatshefte f \"u r Mathematik 98 , 4 (Dec 1984), 311--322

    work page 1984

  35. [35]

    Die multivariate inkomplette B eta- F unktion

    Raab, W. Die multivariate inkomplette B eta- F unktion. Monatshefte f \"u r Mathematik 123 , 1 (Mar 1997), 71--84

    work page 1997

  36. [36]

    Applied Statistical Decision Theory

    Raiffa, H., and Schleifer, R. Applied Statistical Decision Theory . The M. I. T. Press, Cambridge, 1961

    work page 1961

  37. [37]

    On Global Inversion of Homogeneous Maps

    Ruzhansky, M., and Sugimoto, M. On Global Inversion of Homogeneous Maps . Bulletin of Mathematical Sciences 5 , 1 (2015), 13--18

    work page 2015

  38. [38]

    Negative multinomial distribution

    Sibuya, M., Yoshimura, I., and Shimizu, R. Negative multinomial distribution. Annals of the Institute of Statistical Mathematics 16 , 1 (Dec 1964), 409--426

    work page 1964

  39. [39]

    Pricing derivatives on multiple assets

    Sierag, D. Pricing derivatives on multiple assets. Master's thesis, Universiteit Leiden, Leiden, March 2013

    work page 2013

  40. [40]

    G., and Cuttman, I

    Tiao, G. G., and Cuttman, I. The inverted D irichlet distribution with applications. Journal of the American Statistical Association 60 , 311 (1965), 793--805

    work page 1965

  41. [41]

    Variational learning of finite inverted D irichlet mixture models and applications

    Tirdad, P., Bouguila, N., and Ziou, D. Variational learning of finite inverted D irichlet mixture models and applications. In Artificial Intelligence Applications in Information and Communication Technologies , Y. Laalaoui and N. Bouguila, Eds. Springer International Publishing, Cham, 2015, pp. 119--145

    work page 2015

  42. [42]

    On certain inequalities concerning the incomplete beta function

    Vietoris, L. On certain inequalities concerning the incomplete beta function. Sitzungsber., Abt. II, \"O sterr. Akad. Wiss., Math.-Naturwiss. Kl. 191\/ (1982), 85--92

    work page 1982

  43. [43]

    Inverted D irichlet distribution and multivariate logistic distribution

    Yassaee, H. Inverted D irichlet distribution and multivariate logistic distribution. The Canadian Journal of Statistics / La Revue Canadienne de Statistique 2 , 1 (1974), 99--105

    work page 1974

  44. [44]

    Probability integral of inverted D irichlet distribution and its applications

    Yassaee, H. Probability integral of inverted D irichlet distribution and its applications. Comp. Statist. 76\/ (1976), 64--71

    work page 1976