REVIEW 3 major objections 4 minor 4 references
Properties of the cumulated deficient binary digit sum
T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The cumulated deficient binary digit sum equals the cardinality of an explicit set family, proving a prior conjecture.
desk verdict A small but genuine conjecture is proved; the main proof has a gap in the induction base that is easily patched, and the paper is worth a serious referee. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the set family $S_n$ defined by the congruence condition above; its cardinalities obey identity 2.3, a recurrence in which adding $2^k$ to $n$ adds a predictable block of elements. The other main mechanism is the Takagi function $\tau$, linked to $A268289$ by identity 2.1; the paper combines the two recurrences to obtain functional equations for $\tau$ and then converts those back into index-transformation identities for $A268289$. The named identity doing the work is the cardinality recurrence 2.3 together with its digit-sum counterpart 2.2.
What would settle it
Directly enumerate $S_n$ from its definition and compare with the cumulated binary digit-sum difference for all $n$ up to $2^{16}$; any mismatch would falsify identity 2.4. A cheaper check is the recurrence 2.3 itself: for a fixed $k$, verify that $|S_{n+2^k}| - |S_n|$ equals $(n+1)(\lfloor \log_2 n\rfloor - k + 2) + 2^k - 2^{\lfloor \log_2 n\rfloor + 1}$ for every $n < 2^k$; the first failure refutes the proof's central step.
Extended reading notes
Core claim
The core discovery is identity 2.4: for every nonnegative integer $n$, $A268289_n = |S_n|$, where $S_n = \{m : 1 \le m \le n,\ ((n-m) \bmod 2^{\lfloor \log_2 m\rfloor+1}) < 2^{\lfloor \log_2 m\rfloor}\}$. The proof proceeds by showing that both sequences are generated from the same three initial values, $(0,1,1)$, by the same recurrence: the cardinality transition $|S_{n+2^k}| = |S_n| + (n+1)(\lfloor \log_2 n\rfloor - k + 2) + 2^k - 2^{\lfloor \log_2 n\rfloor+1}$ (identity 2.3) matches the corresponding cumulated digit sum transition (identity 2.2). Once the equality is in hand, the paper combines it with the classical link between $A268289$ and the Takagi function $\tau$ to produce a chain of identities: scaling and reflection formulas for $\tau$, closed forms for $A268289$ at compound indices, and infinite families of index transformations (identities 5.1–5.9).
Load-bearing premise
The proof of the central equality leans on identity 2.3, whose single-sentence justification asserts without case-by-case verification that the congruence defining $S_n$ splits exactly as claimed when the index jumps by a power of two; if that recurrence miscounts any block of elements, the equality with $A268289$ collapses.
Editorial extensions
If this is right
- $A268289$ gains a combinatorial interpretation: its $n$th term counts integer points $m \le n$ satisfying a simple congruence, so bounds like $n/2 \le A268289_n \le n$ become counting statements.
- Every identity proved for $|S_n|$ transfers to $A268289$ and vice versa; in particular the recurrences 4.3–4.8 give fast ways to evaluate $A268289$ at indices near powers of two.
- The Takagi functional equations (3.2, 5.7, 5.8) yield new exact values, including $A268289_{5\cdot 4^m/3 - 2/3} = 4^m$ for every nonnegative integer $m$.
- Between consecutive powers of two there is exactly one $n$ with $A268289_n = n/2$, located before $3\cdot 2^{\lfloor \log_2 n\rfloor -1} - 1$, and those indices follow the recurrence $a_j = a_{j-1} + 2a_{j-2} + 2$, matching sequence A026644.
- The minimum of $A268289$ between consecutive powers of two is bounded below by the corresponding term of sequence A000975 and is never attained at an index greater than the relevant A026644 term.
Reading between the lines
- The equality with $|S_n|$ suggests that the running digit-sum deficit can be computed by a simple incremental test on each $m$, which might be turned into an $O(n)$ streaming algorithm using only $O(\log n)$ state; this is not tested in the paper.
- Because $S_n$ arose from Karatsuba's recursion tree, the identities may imply structural properties of that tree's node counts under index shifts; tracing this connection is an extension the paper does not make.
- The self-similar recurrences for $\tau$ in 5.7–5.8 are likely instances of a more general multiplicative periodicity of the Takagi function under rational scaling; a testable extension would be to search for analogues with denominators other than 4.
- The bound $A268289_n \ge n/2$ and the uniqueness result suggest that the sequence's local minima have a recursive, dragon-curve-like structure; one could test whether the same uniqueness holds for $A268289_n = c\,n$ for other rational $c$.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This manuscript studies the OEIS sequence A268289, the cumulated difference between the number of 1s and 0s in the binary expansions of 1, ..., n. The central claim is that A268289_n equals the cardinality of a congruence-defined set family S_n introduced in equation (1), thereby resolving a conjecture from the author's earlier work. The paper also derives recurrence identities for A268289 (2.2), matching recurrences for |S_n| (2.3), and combines these with Lagarias's Takagi-function formula to prove several new identities involving the Takagi function (Section 3), a long list of index recurrences (Sections 4 and 5), and inequalities for A268289 (Section 6). The exposition is very compressed, with several derivations summarized as 'after some simplification.'
Significance. The identity 2.4, if fully established, gives a clean combinatorial interpretation of A268289 and a concrete bridge between this sequence and the Takagi function through explicit, parameter-free formulas. The later recurrences in Sections 4 and 5 are numerous, elementary, and individually checkable, and the bounds in Section 6 are stated as theorems rather than observations. The paper does not rely on numerical fitting or circular definitions: identity 2.4 is approached via set enumeration, and formula 2.1 is cited from Lagarias's theorem, so the Takagi-based identities are not self-referential. The main weakness is not the overall strategy but the incomplete and occasionally incorrect proof details, which are local and repairable.
major comments (3)
- [§2.4] The proof of the central identity is incomplete as written. Identities 2.2 and 2.3 are stated for positive n with n < 2^k, so the recursion step with index n + 2^k can never produce an index that is itself a power of two: for N = 4 the only candidate representations are n = 2, k = 1 (excluded by n < 2) and n = 0, k = 2 (excluded by n > 0), and similarly for all N = 2^k. Starting from the three stated initial terms (0,1,1), one can generate A_3, A_5, A_6, A_7, but not A_4, and A_12 would require A_4. The text says that the proof of 2.2 computes A_{2^k-1} = 2^{k-1}; this is already false at k = 2, since A_3 = 3. Moreover, no analogous computation of |S_{2^k}| is supplied, although such a base value is needed if the induction is repaired. Please state a complete base family that includes all powers of two, correct the erroneous value A_{2^k-1}, and prove the matching |S_{2^k}| = 2^k - k.
- [§2.3] The proof of the set decomposition in identity 2.3 is too compressed for a result on which the main theorem depends. For m in (n, 2^k), membership in S_{n+2^k} is governed by the congruence (n + 2^k - m) mod 2^{floor(log2 m) + 1} < 2^{floor(log2 m)}. The displayed union of the blocks {2^j + n + 1, ..., 2^{j+1} - 1} requires a verification that the congruence reduces correctly for each such m; this is true because 2^k is divisible by every relevant modulus, but the argument is not given. I checked the small cases and the decomposition appears correct, but identity 2.3 is used in the induction for 2.4, so the missing congruence check should be supplied explicitly.
- [§2.5, §3.2, §3.3] Identity 2.5 has an invalid case at n = 1. For n = 1, floor(log2 n) = 0, so the index n + 2^{floor(log2 n) - 1} = 1 + 2^{-1} is not an integer, while evaluating the right-hand side gives A_1 + 2(2) - 2^1 = 1 + 4 - 2 = 3, which is not A_2 = 1. The domain must be restricted to n >= 2, or the n = 1 case must be handled separately. Since identity 2.5 is the basis for the derivations in 3.2 and 3.3, those proofs should also state explicitly that the dyadic points corresponding to n = 1 lie outside the intended interval xi in [0, 1/2] and are therefore not needed for the continuity argument, or provide a separate argument for them.
minor comments (4)
- [§2.2, §2.3] The statements read 'positive integers such than n < 2^k'; 'than' should be 'that'.
- [§3.1] In the proof of identity 3.1, the dyadic-rational reduction starts with xi = k/2^m, but k = 0 makes floor(log2 k) undefined. Exclude xi = 0 from that step and obtain the endpoint by continuity.
- [§3.3.1] Equation 3.3.1 contains the redundant expression '1 + xi - 1'; this should be simplified to xi to avoid confusion.
- [§6.1] The induction in the proof of 6.1 is described informally ('the relation being true for the initial terms'); please specify the exact base case and induction hypothesis for the blocks between consecutive powers of two.
Circularity Check
No significant circularity: the central identity 2.4 is proved by independent recurrences, and the only self-citation supplies a definition rather than load-bearing support.
full rationale
The paper's central claim, identity 2.4, is not circular. The proof does not assume A268289_n = |S_n| as an input; it separately derives recurrences 2.2 and 2.3 for A and |S| with identical right-hand sides and identical initial terms (0, 1, 1). Identity 2.2 is proved by elementary enumeration of binary expansions, and identity 2.3 is proved directly from the set definition (1) via a union decomposition. Identity 2.1 uses theorem 9.1 of Lagarias, an external published result, as its input; the later Takagi identities are derived from this external formula and the independently established recurrences, not from the target identity. Sections 3 through 5 are algebraic combinations of 2.1-2.5 and standard Takagi functional equations; these are rearrangements or application steps, not self-referential definitions. The self-citation [2] only supplies the definition of S_n and records the earlier empirical conjecture; it is not invoked as an unproved uniqueness theorem or as the reason the identity holds. No parameter is fitted, and no 'prediction' is statistically forced by a subset of the data. The only visible weakness is a rigor gap in the induction base for powers of two: recurrences 2.2 and 2.3 are stated for positive n, so they do not by themselves generate indices of the form 2^k from the three stated initial values, and identity 2.3 lacks an explicit base computation for |S_{2^k}|. That is a correctness/completeness concern about the proof exposition, not an instance of a conclusion reducing to its own inputs, so it does not raise the circularity score.
Assumptions & free parameters
assumptions (3)
- domain assumption Identity 2.1: A268289_n = (n+1)(m-k+1) - (2+tau(xi))2^m + 2^{k+1} - 1
- standard math Takagi function self-similarity: tau(xi)=tau(1-xi) and tau(xi/2)=xi/2+tau(xi)/2
- standard math Continuity of the Takagi function
Cite this review
Pith. "Pith review of Properties of the cumulated deficient binary digit sum." pith.science (2026). https://pith.science/paper/YWJJAHOP
@misc{pith2026190802250,
author = {Pith},
title = {Pith review of: Properties of the cumulated deficient binary digit sum},
year = {2026},
howpublished = {\url{https://pith.science/paper/YWJJAHOP}},
note = {Machine review of arXiv:1908.02250}
}
read the original abstract
The sequence A268289 from the On-Line Encyclopedia of Integer Sequences, namely the cumulated differences between the number of digits 1 and the number of digits 0 in the binary expansion of consecutive integers, is studied here. This sequence happens to match a sequence of cardinalities of some specific sets. Furthermore, it can also be expressed by using the Takagi function. The three different definitions have their own properties and combining them together lead to some new identities.
Reference graph
Works this paper leans on
-
[1]
The On-Line Encyclopedia of Integer Sequences , published electronically at https://oeis.org
-
[2]
Flattening Karatsuba's recursion tree into a single summation
Thomas Baruchel, Flattening Karatsuba’s recursion tree into a single sum- mation, in Computer Science, ed. Springer Nature, accepted for publication. Also available at https://arxiv.org/abs/1902.08982
work page Pith review arXiv 1902
-
[3]
Jeffrey C. Lagarias, The Takagi function and its properties , in Functions in number theory and their probabilistic aspects, 153–189, RIMS Kk yroku Bessatsu, B34, Res. Inst. Math. Sci. (RIMS), Kyoto, 2012. MR3 014845
work page 2012
-
[4]
Teiji Takagi, A simple example of the continuous function without deriva- tive, Tokyo Sugaku-Butsurigakkwai Hokoku, vol. 1, 176–177, 1901. 1The two sequences A026644 and A000975 are deeply related to the Chinese Rings puzzle. 8
work page 1901
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.