Benford Behavior in Stick Fragmentation Problems
Bruce Fang, Ava Irons, Ella Lippelman, Steven J. Miller · 2025
Read Benford Behavior in Stick Fragmentation Problems by Bruce Fang, Ava Irons, Ella Lippelman, Steven J. Miller (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-17360.txt.
Introduction
Bruce Fang, Ava Irons, Ella Lippelman, and Steven J Miller study Benford behaviour in stick breaking tasks. Their paper is from two thousand twenty-five. Benford's law says that in many real data sets, small first digits show up more often than large ones, at a set rate. They ask when the lengths of stick pieces, made by breaking a stick at random points, follow that law.
They build on past work on the one-way case. With sum identities for multinomial counts, they shrink the high-way case down to the one-way case. They give a test that is both needed and enough for piece lengths to settle into strong Benford behaviour as breaks grow.
Catalog record
| CREATOR | Bruce Fang, Ava Irons, Ella Lippelman, Steven J. Miller |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-17360
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Benford Behavior in Stick Fragmentation Problems Bruce Fang, Ava Irons, Ella Lippelman, Steven J. Miller Abstract Benford’s law is the statement that in many real-world data set, the probability of having digit d in base B , where 1\leq d\leq B , as the first digit is \log_{B}\left((d+1)/d\right) . We sometimes refer to this as weak Benford behavior, and we say that a data set exhibits strong Benford behavior in base B if the probability of having significand at most s , where s\in[1,B) , is \log_{B}(s) . We examine Benford behaviors in stick fragmentation model. Building on the work on the 1-dimensional stick fragmentation model, we employ combinatorial identities on multinomial coefficients to reduce the high dimensional stick fragmentation model to the 1-dimensional model and provide a necessary and sufficient condition for the lengths of the stick fragments to converge to strong Benford behavior. 1. Introduction In the late nineteenth century, astronomer Simon Newcomb observed that in the logarithmithic books at his workplace, certain pages were “more worn than others” [ New ] . In particular, there was more wear and tear in the earlier pages than the later pages. He deduced that there is a “bias” towards smaller leading digits, with the digit 1 showing up roughly 30\% of the time, the digit 2 showing up roughly 18\% of the time, and so on. Newcomb’s findings were practically ignored until about fifty years later when physicist Frank Benford published his own research on the distribution of leading digits in Reader’s Digest [ Ben ] . Benford displayed a table of roughly 20,000 observations from twenty different sets of data, shown in Figure 1 . Figure 1. From [ New ] : Benford’s 20,000 observations. Figure 2. The distribution of Benford’s law. Benford called this distribution the “law of Anomalous Numbers”, but due to the popularity of his publication, the phenomena eventually became known as “Benford’s law”; see Figure 2 for probabilities. Benford’s law is a powerful phenomena that occurs in a variety of data, including accounting, elections, finance, geosciences, physics, population data, street addresses, and more. The law’s prevalence makes it useful for ensuring data integrity, including using it as a method of fraud detection for tax returns, insurance claims, and expense reports [ Ni , PTTV ] . For more information on the history of Benford’s law, see [ BeHi , Hi1 , Hi2 , Mil , Rai ] . In 1986, Lemons [ Lem ] used Benford’s law to analyze the partitioning of a conserved quantity. Since then, mathematicians and physicists have examined Benfordness of various fragmentation processes. Among these processes is the stick fragmentation process , the subject of this paper. In the 1-dimensional multi-proportion stick fragmentation model, one starts with a stick of length L and fixes an integer m\geq 2 . At Stage 1, the stick is split into m substicks of lengths p^{(1,1)}_{1}L,p^{(1,1)}_{2}L,\dots,p^{(1,1)}_{m}L according to proportions p^{(1,1)}_{1},p^{(1,1)}_{2},\dots,p^{(1,1)}_{m} drawn from a probability distribution on (0,1) , such that p^{(1,1)}_{1}+p^{(1,1)}_{2}+\cdots+p^{(1,1)}_{m}=1 . At Stage 2, each of the m substicks created from Stage 1 is split into m smaller substicks according to proportions p_{1}^{(2,k)},p_{2}^{(2,k)},\dots,p_{m}^{(2,k)} drawn from the same probability distribution such that p_{1}^{(2,k)}+p_{2}^{(2,k)}+\cdots+p_{m}^{(2,k)}=1 for some 1\leq k\leq m . In general, at Stage N , every substick created from Stage N-1 is further split into m smaller sticks according to proportions p^{(N,k)}_{1},p^{(N,k)}_{2},\dots,p^{(N,k)}_{m} drawn from the same probability distribution on (0,1) such that p^{(N,k)}_{1}+p^{(N,k)}_{2}+\cdots+p^{(N,k)}_{m}=1 for some 1\leq k\leq m^{N-1} . By the end of N stages, there are m^{N} sticks in total, and a question of particular interest to us is whether or not this fragmentation process results in stick lengths that are distributed according to Benford’s law. In this paper, we specialize our investigation to a particular case of the multi-proportion stick fragmentation model, by fixing beforehand the m proportions p_{1},p_{2},\dots,p_{m} with which we split every substick. We provide a necessary and sufficient condition for such a fixed multi-proportion stick fragmentation model to result in sticks with lengths converging to a generalized Benford’s law. Theorem 1.1 . For any integer m>2 , choose p_{1},p_{2},\dots,p_{m-1}\in(0,1) such that p_{1}+p_{2}+\dots+p_{m-1}<1 . Set p_{m}:=1-(p_{1}+p_{2}+\dots+p_{m-1}) . At each stage, we cut a given stick according to proportions p_{1},p_{2},\dots,p_{m-1} to create m pieces. After N iterations, we have m^{N} sticks in total, of lengths \displaystyle A_{k_{1},k_{2},\dots,k_{m}}\ :=\ Lp_{1}^{k_{1}}p_{2}^{k_{2}}\cdots p_{m}^{k_{m}}, (1.1) for 0\leq k_{1},k_{2},\dots,k_{m}\leq N such that k_{1}+k_{2}+\cdots+k_{m}=N . Let y_{i}=\log_{10}(p_{i}/p_{i+1}) for 1\leq i\leq m-1 . Then the stick fragmentation model results in stick lengths that converge to strong Benford’s law if and only if y_{i}\not\in\mathbb{Q} for some 1\leq i\leq m-1 . 1.1. Definitions and Theory of Benford’s Law We begin with a definition of Benford’s law (see for example [ Dia , Mil ] ). Definition 1.2 (Benford’s Law for the Leading Digit) . We say a data set exhibits Benford’s law for the leading digit if the frequency of leading digit d is \log_{10}(\frac{d+1}{d}) . There are many methods of proving that a data set follows Benford’s law. A common one is to use the Uniform Characterization Theorem. An important definition for this characterization is the notion of the significand of a real number. Definition 1.3 (The Significand) . For any positive x , we can express it as x\ =\ S_{10}(x)\cdot 10^{k_{10}(x)} (1.2) for unique S_{10}(x)\in[1,10) and integer k_{10}(x) . We call S_{10}(x) the significand of x . There is a more generalized version of Benford’s law for the entire significand, not just for the first digit. Definition 1.4 . We say that a sequence of random variables \{X^{(N)}\}_{N=1}^{\infty} , converges to strong Benford’s law if \lim_{N\rightarrow\infty}\mathbb{P}(S_{10}(X^{(N)})\leq s)\ =\ \log_{10}(s) (1.3) for all s\in[1,10] . Note that the increasing indices usually correspond to the growing size of the data set, so convergence to strong Benford’s law is really an asymptotic statement. Definition 1.5 (Uniform Distribution Modulo 1) . A sequence of random variables \{X^{(N)}\}_{N=1}^{\infty} converges to being equidistributed mod 1 if \lim_{N\rightarrow\infty}\mathbb{P}(X^{(N)}\textup{ mod }1\leq s)\ =\ s (1.4) for all s\in[0,1] . Now, we are ready to state the Uniform Characterization Theorem [ Mil ] . Theorem 1.6 (Uniform Characterization Theorem) . A sequence of random variables converges to strong Benford’s law base 10 if and only if the sequence of the base 10 logarithms of the random variables converges to being equidistributed mod 1 . Thus, convergence to strong Benford’s law and convergence to uniform distribution modulo 1 are in fact equivalent conditions. Thus, to prove (or disprove) convergence to strong Benford’s law, it suffices to prove (or disprove) convergence to uniform distribution modulo 1. 1.2. Fixed Proportion Stick Fragmentation Model and the Multinomial Distribution Now that we have a proper foundation for Benford’s law, it is time to describe in detail the model for this paper. The model can be thought of as an extension of a model from Becker et. al called the fixed single proportion stick fragmentation model [ Be ] . Suppose we start with a stick of length L . We are going to split the stick at one fixed proportion 0<p<1 . After the first break, we have two sticks, namely of lengths Lp and L(1-p) . Now we split the two sticks again at the same fixed proportion p , resulting