Benford behavior resulting from stick and box fragmentation processes

Bruce Fang, Steven J. Miller · 2025

Read Benford behavior resulting from stick and box fragmentation processes by Bruce Fang, 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-2508-12915.txt.

Introduction

Bruce Fang and Steven Miller study Benford's law in breaking processes in this 2025 math paper. Benford's law says first digits in many datasets follow a fixed pattern, with one far more common than nine. The authors ask when fragments of broken objects obey that law.

They study two models: sticks split into pieces and boxes split in many dimensions. For sticks, identities on multinomial coefficients reduce the multi-split case to a simpler one. They give an exact condition for strong Benford behavior and measure the gap from it. For boxes, Fourier tools prove that face volumes tend to strong Benford behavior under mild conditions.

Catalog record

CREATORBruce Fang, Steven J. Miller
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE89 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-12915

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Benford behavior resulting from stick and box fragmentation processes Bruce Fang, 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 two different probabilistic model: stick and box fragmentation. Building on the joint work of [ B– ] 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. Then we answer a conjecture posed by [ BDMMM ] on the high dimensional box fragmentation model. Using tools from order statistics, we prove that under some mild conditions, faces of any arbitrary dimension of the box have total volume converging to strong Benford’s behavior. 1. Introduction At the beginning of the 20 th century, astronomer and mathematician Simon Newcomb [ Ne ] noticed an unusual pattern in the logarithmic tables he used at work. The early pages of the books were far more worn than the later ones, suggesting that numbers starting with smaller digits were consulted more often. From this, Newcomb inferred that people had a tendency to encounter numbers beginning with 1 more frequently. Specifically, he found that 1 appeared as the leading digit about 30\% of the time, 2 about 17\% , with the frequency decreasing for larger digits. Although he formulated a mathematical explanation for this curious phenomenon, his discovery initially went largely overlooked. It took another 57 years after Newcomb’s discovery for physicist Frank Benford to make the exact same observation as Newcomb: the first pages of logarithmic tables were used far more than others. He formulated this law as follows. Definition 1.1 . [ Ben , Page 554] Data exhibits (weak) Benford behavior base B if the frequency F_{d} of leading digit d is \displaystyle F_{d}\ =\ \log_{B}\left(\frac{d+1}{d}\right). (1.1) Nowadays, Benford’s Law is used in detecting many different forms of fraud, and its prevalence in the world fascinates not only mathematicians, but many other scientists as well (to learn more about Benford’s law and its many applications, we recommend [ BeHi , Nig , Mil1 , NiMi ] to name a few). In 1986, Lemons [ Lem ] proposed using Benford’s law to analyze the partitioning of a conserved quantity. Since then, driven by the potential application to nuclear fragmentation, mathematicians and physicists have taken an interest in the Benford behavior of various fragmentation processes. Among these processes of interest is stick fragmentation . In a 1-dimensional stick fragmentation model, one begins with a stick of length L . Draw p_{1} from a probability distribution on (0,1) . This fragments the stick into two sub-sticks of lengths p_{1}L and (1-p_{1})L . For each sub-stick, draw two probabilities p_{2} and p_{3} from the same distribution. We perform this fragmentation process for a total of N stages, where at each stage we fragment all the sub-sticks created during the previous stage according to probabilities. Hence, by the end of N stages, there are 2^{N} sticks. Of particular interest is whether or not this fragmentation process converges to Benford’s behavior. 1.1. Previous Work on Fragmentation An important definition when studying a more precise statistical version of Benford behavior is the notion of the significand of a real number, i.e., its leading digits in scientific notation. Definition 1.2 . Given a positive real number x , we say that its significand base B>1 , denoted S_{B}(x) , is the unique real number S_{B}(x)\in[1,B) such that x=S_{B}(x)\cdot B^{k} , where k is an integer. As is common practice with these techniques involving proofs of Benford behavior, we define a more general version of Benford behavior that allows for processes that do not necessarily exhibit Benford behavior at first but converge to it in the limit. Definition 1.3 . We say that a sequence of random variables X^{(n)} converges to strong Benford behavior base B if and only if \mathbb{P}(S_{B}(X^{(n)})\leq s)\to\log_{B}(s) (1.2) for all s\in[1,B] . Notice by compactness that this implies uniform convergence of ( 1.2 ). An equivalent formulation to the above is the Uniform Distribution Characterization , which is especially suited for investigation of products of random variables; see [ Dia ] for a proof. Proposition 1.4 (Uniform Distribution Characterization) . [ Dia ] A sequence of random variables X^{(n)} converges to strong Benford behavior in base B if and only if \displaystyle\mathbb{P}\left(\log_{B}(X^{(n)})\textup{ mod }1\leq t\right)\rightarrow t, (1.3) for all t\in[0,1] . If ( 1.4 ) is satisfied, then we say that \log_{B}(X^{(n)}) converges to being equidistributed mod 1 . We may now state the previous results on stick fragmentation. Becker et al. [ B– ] proved a theorem regarding fixed proportion 1-dimensional stick fragmentation (compare with their Theorem 1.11). Theorem 1.5 . [ B– ] Choose any p\in(0,1) . In Stage 1, cut a given stick at proportion p to create two pieces of length p and 1-p . In Stage 2, cut each resulting piece into two pieces at the same proportion p . Perform this process for a total of N stages, generating 2^{N} sticks with N+1 distinct lengths (assuming p\neq 1/2 ) given by \displaystyle x_{1} \displaystyle\ =\ Lp^{N} \displaystyle x_{2} \displaystyle\ =\ Lp^{N-1}(1-p) \displaystyle x_{3} \displaystyle\ =\ Lp^{N-2}(1-p)^{2} \displaystyle\quad\vdots \displaystyle x_{N} \displaystyle\ =\ Lp(1-p)^{N-1} \displaystyle x_{N+1} \displaystyle\ =\ L(1-p)^{N}, (1.4) where the frequency of x_{n} is \binom{N}{n}/2^{N} . Choose y so that B^{y}=(1-p)/p , which is the ratio of adjacent lengths (i.e., x_{i+1}/x_{i} ). The decomposition process results in stick lengths that converge to strong Benford’s behavior base B if and only if y\not\in\mathbb{Q} . If y has finite irrationality exponent, the convergence rate can be quantified in terms of that exponent, and there is a power savings. Theorem 1.5 suggests the possibility of examining another fixed proportion 1-dimensional stick fragmentation model, where we cut a piece into an arbitrary m\geq 2 number of pieces according m-1 fixed proportions at each stage. Using new techniques involving multinomial coefficients to reduce the problem into the original fixed proportion 1-dimensional stick fragmentation model in [ B– ] , this paper establishes the following result (see Section 2 ). Theorem 1.6 . 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 all sticks according to proportions p_{1},p_{2},\dots,p_{m-1} to create m pieces. After stage N , we have m^{N} sticks in total, of lengths \displaystyle A_{k_{1},k_{2},\dots,k_{m}}\ =\ p_{1}^{k_{1}}p_{2}^{k_{2}}\cdots p_{m}^{k_{m}}, (1.5) 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_{B}(p_{i}/p_{i+1}) for 1\leq i\leq m-1 . Then the decomposition process results in stick lengths that converge to strong Benford’s behavior base B if and only if y_{i}\not\in\mathbb{Q} for some 1\leq i\leq m-1 . Another possibility is to consider fragmentation processes in higher dimensions. This was considered by Durmić and Miller in [

Read the complete text