REVIEW 2 major objections 9 references
Sequential Correct Screening and Post-Screening Inference
T0 review · 2 major / 0 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read Sequential Correct Screening guarantees that its reported subsets contain the true top-m variables at every stopping time, and its intervals control the false coverage rate.
desk verdict The full text is a different paper; the abstract promises a useful top-m screening/PSI method, but I can't review what isn't there. 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
Two mechanisms carry the proposal. Sequential Correct Screening (SCS) is the screening engine: it examines candidate variables over successive rounds, discards any variable whose evidence rules it out of the top-m, and returns nested survivor subsets; its defining property is anytime validity, so the probability that the true top-m set is covered at every reported time stays at the nominal level regardless of the stopping rule used. The companion device is the post-screening inference (PSI) procedure, which constructs confidence intervals for the retained parameters and is engineered to control the false coverage rate (FCR) — the expected proportion of reported intervals that miss their true
What would settle it
Run repeated simulations of top-m screening in which the candidate estimates are serially dependent or heavy-tailed, stop the procedure adaptively (for example, as soon as the retained set stops shrinking), and measure the empirical fraction of runs in which the final subset misses the true top-m set, and the fraction of reported intervals that miss their targets. If either fraction exceeds the level the method promises — with the same procedure on independent, well-behaved data as a control — the anytime-coverage or false-coverage claim is refuted.
Extended reading notes
Core claim
The paper's claim is that screening for the top-m variables can be made anytime-valid: SCS produces a sequence of variable subsets such that, with probability at least the nominal level, every subset in the sequence contains the true top-m variables simultaneously across all stopping times. This is stronger than a fixed-sample guarantee, because the user may stop after any number of screening rounds and the subset on the table is still covered. The companion PSI procedure constructs confidence intervals for the parameters of the variables that survive screening, and is designed to control the false coverage rate whenever it is conducted, meaning that across repeated selections the expected p
Load-bearing premise
The load-bearing premise is that the statistics used to rank candidate variables behave as the proof requires, enough to make the anytime coverage and false-coverage guarantees hold, and the abstract does not state those conditions nor can they be located here, as the full text bundled with this submission is an unrelated manuscript on turbulent pipe flow.
Editorial extensions
If this is right
- An analyst who must stop early — because of budget, time, or a fixed deadline — can report the current subset with the same nominal coverage guarantee instead of needing a pre-specified sample size.
- Reported intervals, taken as a package at the moment of stopping, carry a controlled false coverage rate, so the analyst does not need a separate correction for having selected the variables from the data.
- The method targets the concrete task of picking the m largest population parameters, so it applies wherever practitioners choose the best features, treatments, or units from a larger pool, as in the paper's suicide-rate application.
- The paper's simulation studies and data application indicate that the guarantees are attainable in finite samples, not only in the limit.
Reading between the lines
- If the proof technique generalizes, SCS-style anytime-valid screening could extend to other structured selection tasks — top-k groups, clusters, or discovery rules — and to data streams that continue indefinitely; the paper itself does not construct those extensions.
- The full text bundled with this submission is an unrelated manuscript on turbulent pipe flow, so the proofs and the exact regularity conditions of SCS and PSI could not be verified here; a reader should confirm what assumptions the proof places on the candidate estimators — for example independence, moment, or mixing conditions — before applying the method to serially dependent or heavy-tailed dat
- A natural end-to-end extension the paper does not build is combining SCS with online multiple-testing corrections so that variables receive error-controlled significance labels as data accumulate, not just a covered top-m set and intervals.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The submission (arXiv:2508.14596) presents an abstract claiming a new Sequential Correct Screening (SCS) methodology that returns a sequence of variable subsets containing the true top-m set with high probability at any stopping time, and a post-screening inference (PSI) procedure that controls the false coverage rate (FCR) whenever conducted. The abstract further states that theoretical guarantees, simulation studies, and a suicide-rate application are provided. The supplied full text is not that manuscript. It is arXiv:2508.14593, a Journal of Fluid Mechanics submission by Kranz, Morón, and Avila on Bayesian optimisation of energy consumption in turbulent pipe flow. The full text contains pipe-flow equations, DNS results, and optimisation tables; it contains no SCS algorithm, no PSI procedure, no theorem statements, no proofs, no simulation study of screening, and no suicide-rate data analysis. Consequently, the central claims of the paper are not present in the submitted manuscript and cannot be evaluated.
Significance. If the claimed results held, the paper would address a valuable problem: anytime-valid selection of top-m variables and post-selection FCR control at arbitrary stopping times are genuinely underdeveloped in the sequential testing literature. The specific target 'FCR whenever it is conducted' is a meaningful and nontrivial goal. However, because the submitted full text is unrelated to these claims, I cannot assess correctness, novelty, or assumptions. There are no machine-checked proofs, no reproducible code, and no derivations to credit. The manuscript as submitted therefore has no auditable statistical content beyond the abstract, and its significance cannot be established.
major comments (2)
- [Full text, all sections] The entire submitted body is a different paper. Sections 1-5 and Appendices A-D describe direct numerical simulations of turbulent pipe flow and Bayesian optimisation of pulsatile driving waveforms, with equations (1.1)-(2.9) and Table 1; none of the SCS/PSI content from the abstract appears. There is no definition of top-m, no sequential algorithm, no filtration or stopping time, no theorem statement, and no proof of anytime validity or FCR control. The central claims are therefore entirely unsubstantiated in the supplied manuscript.
- [Abstract vs. body] The abstract promises theoretical guarantees and a real-data application on suicide rates. The supplied text contains no regularity conditions under which the anytime-valid coverage holds, no derivation of the FCR bound, and no simulation or data section matching that description. I cannot identify any hidden assumption or circular step because the statistical argument itself is absent; this absence is load-bearing and prevents any substantive review of the methodology.
Circularity Check
No circularity detectable; the supplied full text is an unrelated manuscript (arXiv:2508.14593), so the SCS/PSI derivation chain cannot be audited.
full rationale
The abstract of arXiv:2508.14596 claims a novel 'Sequential Correct Screening' methodology with anytime-valid guarantees and post-screening inference controlling FCR, but the submitted full text is a completely different paper (arXiv:2508.14593v1, 'Bayesian minimisation of energy consumption in turbulent pipe flow via unsteady driving' by Kranz, Morón and Avila). There is no overlap in authors, topic, equations, or results. Consequently, there is no presented derivation chain for SCS or PSI to walk: the claimed theorems, assumptions, proofs, simulation details, and data analysis are all absent from the supplied text. Circularity requires exhibiting a specific reduction where an output equals its input by construction, or where a load-bearing premise is justified only by a self-citation chain. No such reduction can be exhibited because the essential statistical content is missing entirely. The mismatched submission is a serious verifiability and integrity problem, but it is not a circularity problem. The pipe-flow paper that was actually provided is, on its face, self-contained empirical/computational work whose conclusions are based on DNS and optimization, not on fitting a prediction to its own inputs. Therefore, the correct circularity finding is 0: no specific circular step is identifiable from the available evidence.
Assumptions & free parameters
assumptions (1)
- domain assumption The screening statistics at each sequential step have sampling properties (e.g., known or tightly bounded estimation error, or a martingale-type structure) that make the anytime-valid coverage proof valid; these conditions are not stated in the abstract.
Cite this review
Pith. "Pith review of Sequential Correct Screening and Post-Screening Inference." pith.science (2026). https://pith.science/paper/7PV6NADM
@misc{pith2026250814596,
author = {Pith},
title = {Pith review of: Sequential Correct Screening and Post-Screening Inference},
year = {2026},
howpublished = {\url{https://pith.science/paper/7PV6NADM}},
note = {Machine review of arXiv:2508.14596}
}
abstract
Selecting the top-$m$ variables with the $m$ largest population parameters from a larger set of candidates is a fundamental problem in statistics. In this paper, we propose a novel methodology called Sequential Correct Screening (SCS), which sequentially screens out variables that are not among the top-$m$. A key feature of our method is its anytime validity; it provides a sequence of variable subsets that, with high probability, always contain the true top-$m$ variables. Furthermore, we develop a post-screening inference (PSI) procedure to construct confidence intervals for the selected parameters. Importantly, this procedure is designed to control the false coverage rate (FCR) whenever it is conducted -- an aspect that has been largely overlooked in the existing literature. We establish theoretical guarantees for both SCS and PSI, and demonstrate their performance through simulation studies and an application to a real-world dataset on suicide rates.
Reference graph
Works this paper leans on
-
[1]
Pacific Journal of Mathematics 16 (1), 1–3
������� ����� 1966 Minimization of functions having lipschitz continuous first partial derivatives. Pacific Journal of Mathematics 16 (1), 1–3. ������� ��� ������ ��� ������ ��� ����������� �� � �������� ��2010 Experimental assessment of drag reduction by traveling waves in a turbulent pipe flow.Physics of Fluids 22 (11). ������ ����� �������� ������ � ��...
work page 1966
-
[2]
IMA Journal of Applied Mathematics 6 (1), 76–90
general considerations. IMA Journal of Applied Mathematics 6 (1), 76–90. ����� �������� ���������� ���� ��� ��������� ����� ��� ��������� ������ � ������ ��������� �� 2024 Acceleration is the key to drag reduction in turbulent flow. Proceedings of the National Academy of Sciences 121 (43). ��������� ������� ��� ���� ������ � ������ ���� 2020 Spatiotempora...
work page 2024
-
[6]
phase-averaged analysis.Journal of Fluid Mechanics 766, 272–296. �������� ������ ������� ������ �� � ��������� ���� ��2019 Stabilisation and drag reduction of pipe flows by flattening the base profile. Journal of Fluid Mechanics 863, 850–875. ������� �������� �� 2010 Controlling turbulence. Science 327 (5972), 1462–1463. ������� � 1972 On bayes methods fo...
work page 2019
-
[46]
The Computer Journal 13 (3), 317–322
��������� ��1970 A new approach to variable metric algorithms. The Computer Journal 13 (3), 317–322. ����� ����� ������� ������ ����������� ������ ����� �� � �������� ��������2023a On–off pumping for drag reduction in a turbulent channel flow.Journal of Fluid Mechanics
work page 1970
-
[948]
Journal of Physics: Conference Series 2753 (1), 012009
��� ���� ������� ������������� ������� � ������ ���� 2024 Predictability of decay events in transitional wall-bounded flows. Journal of Physics: Conference Series 2753 (1), 012009. �������� �������� � ������ ������ 2004 Critical assessment of turbulent drag reduction through spanwise wall oscillations. Journal of Fluid Mechanics 521, 251–271. ��������� ��...
work page 2024
-
[966]
����� ����� ������� ������ ����������� ������ ����� �� � �������� ��������2023b Saving energy in turbulent flows with unsteady pumping. Scientific Reports 13 (1). ��������� ����2001 Pump life cycle costs — a guide to LCC analysis for pumping systems. Brussels, Belgium: Hydraulic Institute & Europum, iSBN: 978-1-935762-89-8. ���� ������������ ������� � ���...
arXiv 2001
-
[979]
� ������ ���� ������� ��������� ������� ������ ������� ��������� ���� �������� ���� ����� � ������ ���� 2020 nscouette– a high-performance code for direct numerical simulations of turbulent taylor–couette flow. SoftwareX 11, 100395. ��� �������� ���� ��� ���� ���� � ��� ���2024 Improved sqp and slsqp algorithms for feasible path-based process optimisation...
work page 2020
-
[1966]
is used in the line search to guarantee an appropriate step size ��𝑗�. A detailed description of the SLSQP method can be found in Schittkowski (1982), Kraft (1988) or Ma et al. (2024) and the references therein. We assessed the performance in terms of feasibility, computational effort and robustness of the SLSQP method using the restricted (uni-variant) t...
work page 1982
Show all 9 references
-
[1982]
The Armijio condition (Armijo
arising from a factorization of a Broyden-Fletcher-Goldfarb-Shanno-approximation to the Hessian (BFGS, Broyden (1970); Fletcher (1970); Goldfarb (1970); Shanno (1970)). The Armijio condition (Armijo
1970
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.