Pith. sign in

REVIEW 1 major objections 5 minor 25 references

NA62 measures the ultra-rare K+→π+νν̄ decay to 20% precision and finds it consistent with the Standard Model.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 21:00 UTC pith:35CG6ZNJ

load-bearing objection A genuine new NA62 measurement that doubles the K→πνν̄ sample and reaches 20% precision; the dominant upstream-background CDA extrapolation is the one place I'd push referees to dig hard. the 1 major comments →

arxiv 2607.16413 v1 pith:35CG6ZNJ submitted 2026-07-17 hep-ex

Measurement of the branching ratio of the K⁺rightarrowπ⁺νbar{ν} decay

NA62 Collaboration This is my paper
classification hep-ex PACS 13.20.Eb
keywords K+ to pi+ nu nubarbranching ratioNA62rare kaon decayStandard Modelflavour physicsCKMupstream background
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

This paper reports a new measurement of the branching ratio of the ultra-rare kaon decay K+→π+νν̄ using NA62 data from 2023–2024, combined with the earlier 2016–2022 dataset. The combined result is B(K+→π+νν̄) = (9.6 +1.9 −1.8)×10^-11, a relative precision of 20%. With 84 observed candidates against 30 expected background events, the background-only hypothesis is excluded at more than 6σ. The value matches the Standard Model prediction of about 8×10^-11, and the increased sample roughly doubles the effective statistics, making this the most precise determination of this golden-mode decay.

Core claim

Using the full 2016–2024 dataset, the NA62 collaboration measures B(K+→π+νν̄) = (9.6 +1.8 −1.6 [stat] +0.8 −0.6 [syst])×10^-11 = (9.6 +1.9 −1.8)×10^-11. For the 2023–2024 sample alone, 33 signal candidates are observed where 22.9±1.1 Standard Model signal events and 12.0 +2.9 −2.2 background events are expected, giving B = (7.2 +2.3 −2.1)×10^-11. The combination reaches, for the first time, an expected significance for the Standard Model observation exceeding 5σ, and the observed significance against background-only exceeds 6σ. The measurement is compatible with Standard Model predictions and with the most stringent direct limit on the related decay KL→π0νν̄.

What carries the argument

The analysis is anchored by the single-event sensitivity: the number of kaon decays is normalised to a large sample of K+→π+π0 events, and the expected signal is scaled from the Standard Model branching ratio. A profile likelihood-ratio test over six independent pion-momentum bins extracts the branching ratio, with the expected signal and background per bin as inputs. The dominant background—upstream interactions and accidental matches—is estimated with a fully data-driven template fit to the closest-distance-of-approach (CDA) distribution of an upstream reference sample, extrapolated into the signal region by a fitted factor f_CDA = 0.19 +0.05 −0.04.

Load-bearing premise

The upstream background estimate, which contributes 7.5 of the 12.0 expected background events, rests on the assumption that the fitted fraction f_CDA = 0.19, determined from a template fit to events with CDA > 4 mm, remains valid for the signal region CDA < 4 mm.

What would settle it

If a future, larger upstream reference sample yields a measured f_CDA for CDA < 4 mm that differs from 0.19 by more than the current uncertainties, the background estimate and thus the central branching ratio shift by the corresponding 7.5-event contribution. Concretely, a 2025–2026 dataset with roughly twice the current sample should reproduce f_CDA within its quoted uncertainty; a discrepancy would falsify the present central value.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The measured rate agrees with Standard Model predictions, so no new physics is indicated in s→d neutrino transitions at the current precision.
  • Combined with the KOTO upper limit on KL→π0νν̄, this result sharpens the constraints in the Grossman–Nir plane and excludes a wide class of beyond-Standard-Model scenarios proposed to explain flavour anomalies.
  • Doubling the effective sample size cuts the statistical uncertainty, making the measurement competitive with the theoretical prediction uncertainty of about 8%, so future progress will depend on improving both statistics and theory inputs.
  • NA62 expects to collect 50% more data in 2025–2026, so the final dataset should push the measurement toward roughly 15% precision and provide a sharper test of the Standard Model.

Where Pith is reading between the lines

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

  • The new central value (9.6×10^-11) is lower than the 2016–2022 measurement (13.3×10^-11); if the 2025–2026 data continue this trend toward the Standard Model, the agreement would be even stronger, but if the value rises again, the upstream background model may need closer scrutiny.
  • The 20% experimental precision now approaches the theoretical prediction uncertainty, so improvements in long-distance charm contributions and CKM inputs will be as important as extra statistics for future interpretations.
  • A direct test of the CDA-extrapolation assumption would be to measure f_CDA below 4 mm using a much larger upstream reference sample from the 2025–2026 run; a value significantly different from 0.19 would shift the background estimate and hence the branching ratio.
  • The machine-learning techniques introduced here—transformer-based beam tracking and CNN-based calorimetric identification—reduce backgrounds and improve efficiency, suggesting that similar approaches could benefit other high-rate fixed-target flavour experiments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The paper reports the NA62 measurement of B(K+→π+ννbar) using data collected in 2023–2024 and the combination with the earlier 2016–2022 sample. After describing detector and trigger upgrades, machine-learning-based tracking and particle identification, and a blinded signal selection, the analysis estimates all backgrounds with data-driven methods and extracts the branching ratio with a profile likelihood fit in momentum bins. The quoted results are B(2023–2024) = (7.2+2.3−2.1)×10^-11 and B(2016–2024) = (9.6+1.9−1.8)×10^-11, consistent with Standard Model predictions.

Significance. If the result holds, this is the most precise measurement of the golden K→πννbar mode, reaching 20% relative precision and strengthening constraints on new physics in flavour-changing neutral currents. Strengths of the paper include the blinded analysis, the use of statistically independent control regions with global p-values of 0.65 and 0.79, the data-driven background estimation, and the transparent breakdown of statistical and systematic uncertainties. The central extraction is not circular: the SM benchmark enters only to compute expected event counts, while the branching ratio is fitted directly to observed counts.

major comments (1)
  1. [Sec. 7.2, Eq. (7.1), Fig. 5b] The upstream background is the largest background (7.5 of 12.0 expected events). The factor f_CDA=0.19 is obtained from a template fit to the upstream reference sample only for CDA>4 mm and then applied to the signal region CDA<4 mm. The templates contain events below 4 mm, but the fit does not validate their shapes in that region under the full signal selection, and the 15% systematic from varying template conditions is not a closure test of this transfer. Because a 30% error in the upstream estimate would shift B by about 0.8e-11, comparable to the total uncertainty, this is load-bearing. Please provide a closure test (e.g. a control sample with CDA<4 mm and signal suppressed by an independent criterion) or otherwise demonstrate the validity of the sub-4 mm template extrapolation; also clarify whether the Fig. 6 validation samples exercise this region.
minor comments (5)
  1. [Sec. 7.2] The text states a 15% relative statistical uncertainty and an additional 15% systematic uncertainty on f_CDA, yet quotes f_CDA=0.19+0.05−0.04 (+26%/−21%). Please state how these components are combined to obtain the asymmetric quoted uncertainty.
  2. [Sec. 8] In Eqs. (8.1) and (8.2), specify explicitly that statistical and systematic uncertainties are added in quadrature to obtain the total uncertainty.
  3. [Fig. 2] The multiple axes in Fig. 2a are difficult to read; clarify which axis corresponds to efficiency, false track rate, and purity, and what the error bands represent.
  4. [Sec. 5] The π0 rejection inefficiency is quoted as (1.24±0.04)×10^-8; please state the control sample and method used to measure it.
  5. [Tables 2a and 3] The central values for total background in Table 2a and the sum of background entries in Table 3 differ slightly (12.0 vs 12.34). If this is due to rounding, please say so; otherwise reconcile the values.

Circularity Check

0 steps flagged

No significant circularity: the branching ratio is extracted directly from a profile-likelihood fit to observed event counts, normalized by the externally measured K+→π+π0 branching fraction; the SM benchmark is not used to define the result.

full rationale

The central claim is a direct experimental measurement, not a derivation from the Standard Model. The branching ratio is obtained from a profile-likelihood fit to observed event counts in six momentum bins for 2023–2024 data and twenty-one bins for the 2016–2024 combination (Eqs. 8.1 and 8.2). The normalization uses the external PDG value B(K+→π+π0) = (20.67±0.08)% in Eq. 6.1, and the signal sensitivity in Eq. 6.2 uses the SM benchmark B_SM = 8.4×10−11 only to compute the expected number of SM signal events for sensitivity and consistency checks; the fitted branching ratio is a free parameter, so the measurement is not forced to reproduce the SM input. Background estimates are data-driven from control samples and simulations (Sec. 7), with validations shown in control regions. References to the collaboration's previous paper [16] are methodological self-citations for established selection, fitting, and validation procedures and for the previously published 2016–2022 dataset; the 2023–2024 analysis has its own normalisation, acceptances, trigger efficiencies, and background estimates, and the combination with [16] is an explicit combination of independent datasets. The largest background, upstream background (7.5 of 12.0 expected events), relies on a template fit to CDA>4 mm and an f_CDA = 0.19 extrapolation to CDA<4 mm (Sec. 7.2). This is a systematic assumption and a possible source of bias, but it is not circular: f_CDA is a data-derived scale factor, and the branching-ratio fit does not impose the upstream background estimate by construction. No equation reduces to its own input, no fitted parameter is renamed as a prediction, and no load-bearing argument depends solely on a self-citation. Minor self-citations are methodological and do not undermine the independent content of the measurement. Therefore no significant circularity is present.

Axiom & Free-Parameter Ledger

5 free parameters · 3 axioms · 0 invented entities

The measurement is a standard counting experiment using an external normalization channel and Monte Carlo simulation. The free parameters listed are nuisance parameters measured from control samples; they are not fitted to the signal region. No new particles, forces, or dimensions are proposed.

free parameters (5)
  • f_CDA = 0.19^{+0.05}_{-0.04}
    Extrapolation factor from the template fit to the CDA distribution of the upstream reference sample, used to compute the upstream background (section 7.2).
  • δ_match = -0.13 ± 0.13 (syst.)
    Correction factor to the K/π matching probability in the upstream background estimate, extracted from control samples (section 7.2).
  • Kinematic tail fraction for K⁺→π⁺π⁰ = (9.4±1.5)×10⁻⁴
    Measured from a control sample to estimate the K⁺→π⁺π⁰ background in the signal regions (section 7.1).
  • Kinematic tail fraction for K⁺→μ⁺ν = (1.4±0.5)×10⁻⁵
    Measured from a control sample to estimate the K⁺→μ⁺ν background in the signal regions (section 7.1).
  • Kinematic tail fraction for K⁺→π⁺π⁺π⁻ = (6±1)×10⁻⁶
    Measured from a control sample to estimate the K⁺→π⁺π⁺π⁻ background in the signal regions (section 7.1).
axioms (3)
  • domain assumption The Monte Carlo simulation of the NA62 detector accurately reproduces the detector response for signal and background processes.
    Used to compute acceptances, tail fractions (partly), and to validate the CNN/PID. The paper provides validation on control samples, but the simulation is not independently verified.
  • domain assumption The CDA distribution of the upstream reference sample is a linear superposition of the interaction and accidental templates.
    Section 7.2: the template fit is performed to the CDA distribution of the URS over 4–32 mm; the extrapolation to the signal region assumes the same two-component model holds for CDA<4 mm.
  • domain assumption The external world-average values from the PDG (e.g., B(K⁺→π⁺π⁰) = 20.67±0.08%) are correct.
    Used in Eq. 6.1 for the normalisation and thus for the single event sensitivity. The uncertainty is propagated to the final result.

pith-pipeline@v1.3.0-alltime-deepseek · 15293 in / 7075 out tokens · 70816 ms · 2026-08-01T21:00:12.601836+00:00 · methodology

0 comments
read the original abstract

The NA62 experiment has measured the branching ratio of the rare decay $K^+\to\pi^+\nu\bar{\nu}$ using data collected in 2023 and 2024 at the CERN SPS, obtaining ${\rm B}_{\rm 2023-2024}(K^+\to\pi^+\nu\bar{\nu}) = \left(7.2^{+2.3}_{-2.1}\right)\times10^{-11}$. With results from data collected between 2016 and 2022, the analysis gives ${\rm B}_{\rm 2016-2024}(K^+\to\pi^+\nu\bar{\nu}) = \left(9.6^{+1.9}_{-1.8}\right)\times10^{-11}$, corresponding to a relative precision of $20\%$. This value is consistent with Standard Model predictions.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

25 extracted references · 24 linked inside Pith

  1. [1]

    Buras and E

    A.J. Buras and E. Venturini,The exclusive vision of rare K and B decays and of the quark mixing in the standard model,Eur. Phys. J. C82(2022) 615 [2203.11960]

    Pith/arXiv arXiv 2022

  2. [2]

    Anzivino et al.,Workshop summary: Kaons@CERN 2023,Eur

    G. Anzivino et al.,Workshop summary: Kaons@CERN 2023,Eur. Phys. J. C84(2024) 377 [2311.02923]

    Pith/arXiv arXiv 2023

  3. [3]

    Allwicher et al.,Probing third-generation New Physics withK→πν¯νandB→K (∗)ν¯ν, Phys

    L. Allwicher et al.,Probing third-generation New Physics withK→πν¯νandB→K (∗)ν¯ν, Phys. Lett. B861(2025) 139295 [2410.21444]

    Pith/arXiv arXiv 2025

  4. [4]

    J. Brod, M. Gorbahn and E. Stamou,Two-Loop Electroweak Corrections for theK→πν¯ν Decays,Phys. Rev. D83(2011) 034030 [1009.0947]

    Pith/arXiv arXiv 2011

  5. [5]

    Buras, D

    A.J. Buras, D. Buttazzo and R. Knegjens,K→πν νandε’/εin simplified new physics models,JHEP11(2015) 166 [1507.08672]

    Pith/arXiv arXiv 2015

  6. [6]

    Bordone, D

    M. Bordone, D. Buttazzo, G. Isidori and J. Monnard,Probing Lepton Flavour Universality withK→πν¯νdecays,Eur. Phys. J. C77(2017) 618 [1705.10729]

    Pith/arXiv arXiv 2017

  7. [7]

    Bobeth and A.J

    C. Bobeth and A.J. Buras,Leptoquarks meetε ′/εand rare Kaon processes,JHEP02(2018) 101 [1712.01295]

    Pith/arXiv arXiv 2018

  8. [8]

    Aebischer, A.J

    J. Aebischer, A.J. Buras and J. Kumar,Another SMEFT story:Z ′ facing new results on ϵ′/ϵ,∆M K andK→πν ν,JHEP12(2020) 097 [2006.01138]

    Pith/arXiv arXiv 2020

  9. [9]

    Deppisch, K

    F.F. Deppisch, K. Fridell and J. Harz,Constraining lepton number violating interactions in rare kaon decays,JHEP12(2020) 186 [2009.04494]

    Pith/arXiv arXiv 2020

  10. [10]

    Descotes-Genon, S

    S. Descotes-Genon, S. Fajfer, J.F. Kamenik and M. Novoa-Brunet,Implications ofb→sµµ anomalies for future measurements ofB→K (∗)ν¯νandK→πν¯ν,Phys. Lett. B809(2020) 135769 [2005.03734]

    Pith/arXiv arXiv 2020

  11. [11]

    Marzocca, S

    D. Marzocca, S. Trifinopoulos and E. Venturini,From B-meson anomalies to Kaon physics with scalar leptoquarks,Eur. Phys. J. C82(2022) 320 [2106.15630]

    Pith/arXiv arXiv 2022

  12. [12]

    Crosas et al.,Flavor non-universal vector leptoquark imprints inK→πν¯νand∆F= 2 transitions,Phys

    `O.L. Crosas et al.,Flavor non-universal vector leptoquark imprints inK→πν¯νand∆F= 2 transitions,Phys. Lett. B835(2022) 137525 [2207.00018]

    Pith/arXiv arXiv 2022

  13. [13]

    Gorbahn et al.,The anatomy ofK + →π +ν¯νdistributions,Eur

    M. Gorbahn et al.,The anatomy ofK + →π +ν¯νdistributions,Eur. Phys. J. C84(2024) 680 [2312.06494]

    Pith/arXiv arXiv 2024

  14. [14]

    Buras, J

    A.J. Buras, J. Harz and M.A. Mojahed,Disentangling new physics inK→πν νand B→K(K ∗)ν νobservables,JHEP10(2024) 087 [2405.06742]. [15]KOTOcollaboration,Search for theK L →π 0ν¯νDecay at the J-PARC KOTO Experiment, Phys. Rev. Lett.134(2025) 081802 [2411.11237]. [16]NA62collaboration,Observation of theK + →π +ννdecay and measurement of its branching ratio,JH...

    Pith/arXiv arXiv 2024

  15. [19]

    Anderson et al.,FELIX: a High-Throughput Network Approach for Interfacing to Front End Electronics for ATLAS Upgrades,J

    J. Anderson et al.,FELIX: a High-Throughput Network Approach for Interfacing to Front End Electronics for ATLAS Upgrades,J. Phys. Conf. Ser.664(2015) 082050

    2015

  16. [20]

    Ammendola et al.,The integrated low-level trigger and readout system of the CERN NA62 experiment,Nucl

    R. Ammendola et al.,The integrated low-level trigger and readout system of the CERN NA62 experiment,Nucl. Instrum. Meth. A929(2019) 1 [1903.10200]. 16 [21]NA62collaboration,Performance of the NA62 trigger system,JHEP03(2023) 122 [2208.00897]

    Pith/arXiv arXiv 2019

  17. [22]

    Vaswani et al.,Attention is all you need,NIPS30(2017) 5999 [1706.03762]

    A. Vaswani et al.,Attention is all you need,NIPS30(2017) 5999 [1706.03762]

    Pith/arXiv arXiv 2017

  18. [23]

    Lin et al.,Focal loss for dense object detection,ICCV(2017) 2980 [1708.02002]

    T.-Y. Lin et al.,Focal loss for dense object detection,ICCV(2017) 2980 [1708.02002]. [24]NA62collaboration,Improved calorimetric particle identification in NA62 using machine learning techniques,JHEP11(2023) 138 [2304.10580]

    Pith/arXiv arXiv 2017

  19. [25]

    K. He, X. Zhang, S. Ren and J. Sun,Deep residual learning for image recognition,CVPR (2016) 770 [1512.03385]

    Pith/arXiv arXiv 2016

  20. [26]

    Misra,Mish: A self regularized non-monotonic activation function,BMVC(2020) [1908.08681]

    D. Misra,Mish: A self regularized non-monotonic activation function,BMVC(2020) [1908.08681]

    Pith/arXiv arXiv 2020

  21. [27]

    Loshchilov and F

    I. Loshchilov and F. Hutter,Decoupled Weight Decay Regularization,ICLR(2019) [1711.05101]

    Pith/arXiv arXiv 2019

  22. [28]

    Ghosh, T

    A. Ghosh, T. Schaaf and M.R. Gormley,AdaFocal: Calibration-aware Adaptive Focal Loss, NeurIPS35(2022) 1583 [2211.11838]

    Pith/arXiv arXiv 2022

  23. [29]

    Punzi,Sensitivity of searches for new signals and its optimization,PHYSTAT(2003) MODT002 [physics/0308063]

    G. Punzi,Sensitivity of searches for new signals and its optimization,PHYSTAT(2003) MODT002 [physics/0308063]. [30]Particle Data Groupcollaboration,Review of particle physics,Phys. Rev. D110(2024) 030001

    Pith/arXiv arXiv 2003

  24. [31]

    Buras, D

    A.J. Buras, D. Buttazzo, J. Girrbach-Noe and R. Knegjens,K + →π +ννandK L →π 0ννin the Standard Model: status and perspectives,JHEP11(2015) 033 [1503.02693]. [32]NA62collaboration,An investigation of the very rareK + →π +ννdecay,JHEP11(2020) 042 [2007.08218]. [33]BNL-E949collaboration,Study of the decayK + →π +ν¯νin the momentum region 140< Pπ <199MeV/c,P...

    Pith/arXiv arXiv 2015

  25. [35]

    Ettore Pancini

    Y. Grossman and Y. Nir,K L →π 0ν¯νbeyond the Standard Model,Phys. Lett. B398(1997) 163 [hep-ph/9701313]. 17 The NA62 Collaboration UCLouvain, Centre for Cosmology , Particle Physics and Phenomenology , CP3, Louvain-La-Neuve, Belgium S. Alibocus , B. Bloch-Devaux 1 , E. Cortina Gil , N. Lurkin TRIUMF, V ancouver, British Columbia, Canada T. Numao , V. Shan...

    Pith/arXiv arXiv 1997