Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Neutrino flavor composition using High Energy Starting Events with IceCube

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Twelve years of IceCube data constrain the astrophysical neutrino flavor ratio at Earth to 0.19 electron : 0.43 muon : 0.38 tau, consistent with the earlier 7.5-year result.

desk verdict A candid ICRC proceedings with a genuinely new 12-year HESE flavor measurement; the tau fraction rests on five fragile double-cascade events, but the paper is upfront about it. read the letter →

arxiv 2507.06835 v1 pith:5Z6MEIRU submitted 2025-07-09 astro-ph.HE astro-ph.IM

Neha Lad , Thijs Juan van Eeden , Markus Ackermann (for the IceCube Collaboration) This is my paper · ORCID
classification astro-ph.HEastro-ph.IM
keywords neutrinoflavorcompositionHigh-EnergyStartingEventsastrophysicalneutrinostaudoublecascadeIceCuberatio
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to measure the flavor composition of the astrophysical neutrino flux at Earth using 97 high-energy starting events collected by the IceCube detector over 12 years, all above 60 TeV. It reports a best-fit flavor ratio $f_{\nu_e}: f_{\nu_\mu}: f_{\nu_\tau} = 0.19:0.43:0.38$, a spectral index of $2.84^{+0.19}_{-0.18}$, and consistency with the earlier 7.5-year HESE analysis when the same likelihood is used. The tau fraction is carried almost entirely by five double-cascade candidate events, four of which fall in the background-dominated region and one of which resembles a misclassified muon. A reliable flavor ratio matters because it encodes how cosmic accelerators produce neutrinos and whether neutrino propagation is standard.

What carries the argument

The analysis is carried by a forward-folding likelihood that fits signal and nuisance parameters simultaneously to three reconstructed topologies. Tracks and single cascades enter through two-dimensional distributions in reconstructed energy and zenith angle, while double cascades—tau charged-current interactions whose tau decays to an electron or hadrons, leaving two separated showers—enter through energy and reconstructed tau decay length. Double-cascade candidates must satisfy a chain of classification criteria: a converged double-cascade reconstruction, cascade energies above 1 TeV, vertices inside or within 50 m of the detector, a reconstructed tau length of at least 10 m, energy confinement above 0.99, and an energy asymmetry between -0.98 and 0.3. Detector systematics enter through a Monte Carlo reweighting scheme that varies them event-by-event, which is what lets the fit handle the sparse double-cascade simulation sample.

What would settle it

Reclassify the five double-cascade candidates with the improved reconstruction and updated ice model the paper proposes, and compare against a classifier calibrated with flasher-generated double-cascade signatures; if four of the five are reassigned to single cascades or muon tracks, the best-fit tau fraction $f_{\nu_\tau}=0.38$ would drop well below the central value and the measured flavor ratio would change materially.

Watch

Extended reading notes

Core claim

Using a forward-folding maximum-likelihood fit over three event topologies—tracks, cascades, and double cascades—the analysis reports the 12-year HESE flavor composition at Earth as $f_{\nu_e}=0.19^{+0.26}_{-0.15}$, $f_{\nu_\mu}=0.43^{+0.27}_{-0.17}$, $f_{\nu_\tau}=0.38^{+0.37}_{-0.24}$, with spectral index $2.84^{+0.19}_{-0.18}$ and all-flavor normalization $5.94^{+5.64}_{-4.28}\times10^{-18}\,\mathrm{GeV}^{-1}\mathrm{s}^{-1}\mathrm{sr}^{-1}\mathrm{cm}^{-2}$. The best-fit model predicts 63.4 cascades, 23.4 tracks, and 4.1 double cascades against observed counts of 64, 28, and 5. The result is presented as consistent with the previous 7.5-year measurement and as a step toward resolving the tau-neutrino component of the astrophysical flux.

Load-bearing premise

The result collapses if the five double-cascade events are not really tau neutrinos as the Monte Carlo models them, since four of the five fall in the background-dominated region and one resembles a misclassified muon.

Editorial extensions

If this is right

  • The 12-year flavor ratio becomes the most current HESE-based constraint on the astrophysical neutrino flavor composition at Earth.
  • Agreement with the 7.5-year measurement under a comparable likelihood indicates the selection and fit have not drifted over successive data releases.
  • The tau fraction remains limited by double-cascade statistics: five events set the uncertainty, so growth of that sample is the path to a sharper tau-neutrino result.
  • The optimized summary statistic is projected to tighten the flavor contours by folding in more observables without diluting the sparse Monte Carlo.
  • Combining HESE with the northern track and cascade samples is expected to improve the flavor measurement beyond what a single sample can deliver.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: if the tau-rich central value survives improved classification, a flavor ratio close to equal thirds is exactly what standard oscillation of a pion-decay source predicts, so the measurement would reinforce the conventional picture of cosmic-ray neutrino production.
  • Editorial inference: the paper's caveat that four of five candidates lie in the background-dominated region suggests the tau fraction is the first place to look for movement when the ice model or reconstruction is updated; a robust double-cascade classifier is the decisive upgrade.
  • Editorial inference: the planned flasher-based calibration of double-cascade-like signatures offers a direct way to measure the classifier's false-positive rate, so the same five events could become a validated tau sample rather than a topology-driven hint.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. This proceedings contribution reports an updated measurement of the astrophysical neutrino flavor composition with IceCube's 12-year HESE sample. Events are classified as tracks, cascades, or double cascades using reconstruction likelihoods and hard selection cuts (Section 2). A forward-folding likelihood fit over energy, zenith, and tau-length observables (Section 3) yields a best-fit composition f_nu_e:f_nu_mu:f_nu_tau = 0.19+0.26-0.15 : 0.43+0.27-0.17 : 0.38+0.37-0.24, spectral index 2.84+0.19-0.18, and normalization 5.94+5.64-4.28 x 10^-18 GeV^-1 s^-1 sr^-1 cm^-2 (Section 4). The sample contains 97 events, including five double-cascade candidates. The paper also presents planned improvements using optimized summary statistics, BDT classification, and combined samples (Section 5).

Significance. The measurement is a legitimate update to the IceCube flavor program: it uses established analysis machinery, includes SnowStorm detector systematics, and provides a direct comparison with the 7.5-year result under a matched likelihood. The quoted uncertainties are large and the result is consistent with standard expectations, so the paper's main value is as a status report and methods-development contribution. It is commendably transparent about the fragility of the double-cascade channel. However, the central tau-flavor constraint rests on only five candidate events whose classification is acknowledged to be unstable under reconstruction and ice-model changes; unless that fragility is quantified, the flavor-composition claim cannot be considered fully supported. The paper's strengths are the careful application of established statistical methods, the matched comparison to the prior result, and the explicit limitation statements.

major comments (3)
  1. [4 (Table 1 and Fig. 2)] The tau-flavor result is carried almost entirely by five double-cascade events: Table 1 predicts only 3.8 astrophysical and 0.3 conventional double cascades, while five are observed, and Figure 2 places four of the five in the 68% background region, with the 96 m/77 TeV event described as typical for misclassified muons. The paper itself states (Section 4) that double-cascade identification and reconstructed observables have limited robustness under changes to reconstruction algorithms and ice-model updates. This makes the classification purity and efficiency a load-bearing systematic, yet no quantitative test of its impact is provided. Please add a robustness test that removes or reclassifies the suspicious event, varies the double-cascade selection cuts (L_reco, E_C, E_A) within their resolution, or introduces a nuisance parameter for classification uncertainty, and report how f_nu_tau and the flavor contours change.
  2. [3 (likelihood and binning)] The double-cascade PDF uses 13 energy bins by 10 length bins for an expected yield of about 4 events, and the paper states that sparse Monte Carlo statistics limit the analysis. The limited-MC likelihood of Ref. [8] mitigates Poisson and weighting effects, but the coverage of the quoted 68% contours under sparse templates and the chosen binning is not demonstrated. Please include an ensemble of pseudo-experiments showing that the confidence intervals have proper coverage, and quote the contribution of limited simulation statistics to the flavor uncertainties separately from detector systematics.
  3. [4 (normalization and comparison)] The quoted all-flavor normalization has highly asymmetric uncertainties (Phi = 5.94+5.64-4.28 x 10^-18), and the text asserts that the uncertainty growth relative to the 7.5-year result is expected because extra flavor parameters are introduced, but no supporting comparison is shown. Please present the fit with and without the flavor parameters, and break down the uncertainty into statistical, atmospheric-background, and detector-systematic components, so that the reader can see the source of the degraded constraint.
minor comments (5)
  1. [4 (Fig. 1 caption)] The caption reads "The solid and lines represent the 68% confidence regions"; this should be "solid and dashed lines".
  2. [5] The sentence "the optimized binning was obtained by training on the model on E_tot, l_tau, theta, phi" contains a duplicated preposition; it should read "training on the model with E_tot, l_tau, theta, phi".
  3. [4 (comparison with HESE-7.5)] The phrase "extended likelihood that includes the information from resimulating the tau neutrino candidates" is not defined; please add a sentence explaining how that likelihood differs from the one used here and whether the difference affects the comparison.
  4. [References] Reference [17] lists "A. Balagopal, A. Karle, , and V. Basu" with an extra comma; the author list should be cleaned.
  5. [3 (likelihood definition)] The notation L(n|theta,xi) = L_Double L_Single L_Track would benefit from a short definition of n, theta, and xi, and from a statement that each per-topology likelihood is a product over binned Poisson terms (or a product over events).

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the 12-year HESE flavor measurement is a forward-folding likelihood fit to observed events, and the cited prior work is reused methodology rather than a self-referential derivation.

full rationale

This paper is an experimental measurement, not a derivation from first principles. The reported flavor fractions and spectral index are free parameters in a forward-folding likelihood fit to 97 observed HESE events, with the total likelihood given as L(n|theta,xi)=LDoubleLSingleLTrack. The selection criteria (L_reco>=10 m, E_C>0.99, -0.98<=E_A<=0.3) are fixed conditions derived from simulated signal and background distributions; they are not fitted to the data's flavor ratio, so the measured flavor composition is not forced by construction. The paper's heavy use of IceCube references—for the classification chain (Ref. [6]), the likelihood (Ref. [8]), atmospheric nuisance parameters (Ref. [9]), and detector systematics (Ref. [10])—is reuse of established analysis tools, not a chain in which the central claim reduces to an unverified self-citation. The comparison with the 7.5-year HESE result uses an independent, earlier data set and is explicitly framed as consistency, not as an input to the 12-year fit. Table 1's expected event counts are labeled 'assuming the best-fit flavor composition,' so they are a goodness-of-fit summary rather than an independent prediction masquerading as a result. The paper itself concedes a real limitation: 'the double cascade identification and reconstructed observables have shown limited robustness under changes to reconstruction algorithms and updates to the ice model,' and four of the five double-cascade candidates fall in the 68% background region, with one described as 'typical for misclassified muons.' These are important systematic and modeling concerns that affect the reliability of the tau-flavor fraction, but they do not make the measurement circular: the flavor fractions are not defined in terms of the double-cascade classifier or the background model, they are fitted to the observed data under those assumptions. No equation-level reduction, self-definitional step, or fitted input relabeled as a prediction is present, so the appropriate circularity score is 0.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The central measurement rests on several fitted parameters (flux normalization, spectral index, two flavor fractions, atmospheric nuisance parameters) and on hand-chosen selection thresholds. The analysis also assumes a single power-law flux model, accurate atmospheric background modeling via H4a and SIBYLL2.3c, correct Monte Carlo detector response including the SnowStorm ice model, and validity of Wilks' theorem with sparse Monte Carlo statistics. No new physical entities are introduced.

free parameters (8)
  • All-flavor neutrino flux normalization Phi = 5.94e-18 GeV^-1 s^-1 sr^-1 cm^-2
    Fitted in the likelihood (Section 4); the measurement is strongly tied to this parameter.
  • Spectral index gamma = 2.84 (+0.19 / -0.18)
    Fitted single power-law index in Section 4.
  • Electron flavor fraction f_nu_e = 0.19 (+0.26 / -0.15)
    One of the two free flavor fractions fitted in Section 4 (f_e + f_mu + f_tau = 1).
  • Muon flavor fraction f_nu_mu = 0.43 (+0.27 / -0.17)
    Second free flavor fraction fitted in Section 4.
  • Prompt atmospheric neutrino normalization Phi_prompt = 0 (best fit)
    Nuisance parameter in the atmospheric background fit; Section 4 reports best-fit normalization of zero.
  • Conventional atmospheric neutrino and muon nuisance parameters = not specified
    Atmospheric flux nuisance parameters from Ref [9] are varied in the fit but their best-fit values are not reported.
  • HESE selection thresholds = E_tot > 60 TeV, L_reco >= 10 m, E_C > 0.99, -0.98 <= E_A <= 0.3
    Hand-chosen from simulated signal and background distributions (Section 2); these cuts influence which events enter the flavor fit.
  • Binning choices = 13x10 double cascade bins; 21x10 track and cascade bins; future 80-bin optimized summary
    Chosen to balance MC statistics and sensitivity; affects the likelihood and resulting contours.
assumptions (6)
  • domain assumption Atmospheric backgrounds are modeled by the H4a cosmic ray model with SIBYLL2.3c hadronic interactions, including conventional and prompt neutrino fluxes.
    Invoked in Section 2; a wrong atmospheric background normalization directly shifts the fitted astrophysical flavor fractions and spectral index.
  • domain assumption The astrophysical neutrino flux is a single power law in energy with a flavor composition that is constant over the fitted energy range.
    The likelihood in Section 3 fits normalization, spectral index, and two flavor fractions under this parametrization; deviations from a single power law are not modeled.
  • domain assumption Monte Carlo simulations, including the SnowStorm ice model with anisotropic scattering, accurately describe detector response and event reconstruction.
    Section 3 states detector systematics are handled with SnowStorm and the ice model from Ref [7]; the authors warn that double cascade identification is not robust to ice model updates (Section 4).
  • domain assumption The likelihood defined in Ref [8] correctly accounts for limited Monte Carlo statistics.
    The likelihood is taken from prior work and is not derived in this paper; the reliability of the result depends on its validity for sparse bins.
  • standard math Wilks' theorem provides valid confidence regions despite limited Monte Carlo statistics.
    Used to derive 68% contours in Section 4 and Figure 1; the authors acknowledge sparse MC in the double cascade sample, so asymptotic assumptions may be strained.
  • domain assumption Events are fully described by the three topology hypotheses (track, cascade, double cascade) with the stated branching fractions.
    Section 2 classification; events not fitting these topologies could bias the fit.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Neutrino flavor composition using High Energy Starting Events with IceCube." pith.science (2026). https://pith.science/paper/5Z6MEIRU

@misc{pith2026250706835,
  author       = {Pith},
  title        = {Pith review of: Neutrino flavor composition using High Energy Starting Events with IceCube},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5Z6MEIRU}},
  note         = {Machine review of arXiv:2507.06835}
}
abstract

Astrophysical neutrinos provide crucial insights into their sources and play a key role in multi-messenger astronomy. The neutrino flavor composition at Earth allows us to probe the mechanisms of neutrino production and cosmic ray acceleration, as well as the properties of the environments in which they originate. Understanding the flavor composition also offers a unique opportunity to test new physics in the neutrino sector. The IceCube Neutrino Observatory consists of 1 km$^3$ of ice instrumented with photomultipliers that detect neutrinos through Cherenkov radiation from their interaction products. Different neutrino interactions result in distinct event topologies, such as tracks, cascades, and double cascade events, which allow for the identification of the interacting neutrino type and measurement of the flavor composition of the astrophysical neutrino flux. In this contribution we present the results of the measurement of the flavor ratio of the High-Energy Starting Event Selection based on 12 years of data, a high-purity sample of neutrino interactions that occur inside the detector. In addition, we discuss various methods that could further improve the analysis in the future.

Figures

Figures reproduced from arXiv: 2507.06835 by the authors.

Figure 1
Figure 1. The best-fit flavor com￾position using 12 years of HESE data compared with the results using the same likelihood and 7.5 years of HESE data [7]. The solid and lines represent the 68% confidence regions obtained using Wilk’s theo￾rem. Expected flavor compositions on Earth from different astrophysi￾cal sources are also marked. length and the vertical line indicates the region that contains 68% of the background. Doubl… view at source ↗
Figure 2
Figure 2. Expected num￾ber of double cascade candi￾date events according to the best-fit model as a function of reconstructed energy and tau length. The 5 observed data events are marked with black markers. The sig￾nal (white) and background (dark blue) dominated re￾gions are marked with their respective percentiles. 5. Improvements & Outlook The observables for the double cascade sample in this analysis were the reconstructe… view at source ↗
Figure 3
Figure 3. Projected sensitivity using normal (blue) and optimized (red) anal￾ysis approach. The solid (dashed) lines depict the corresponding 68% (90%) con￾straints derived using Wilks’ theorem. Additional observables such as the energy asymmetry, the energy confinement and the zenith angle were not included because it would increase the dimensionality of the analysis histogram. Limited Monte Carlo statistics hinder an accura… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The median and quartiles of the reconstructed energy asymmetry 𝐴rec (left) and length 𝐿rec (right) resolution for double cascades in the HESE sample. Further improvements are anticipated by replacing the current particle identification method with a boosted decision tr…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Paucity of downward UHE neutrino tracks in IceCube versus unexpected huge KM3-230213A event: solving the puzzles?

    astro-ph.HE 2025-12 reject novelty 3.0 of 10

    The record-energy KM3-230213A event is more likely a tilted-array misreconstruction of an atmospheric muon than a genuine ultra-high-energy neutrino, according to a comparison with IceCube and Auger observations.

Reference graph

Works this paper leans on

20 extracted references · 19 canonical work pages · cited by 1 Pith paper

  1. [8]

    C. A. Argüelles, A. Schneider, and T. Yuan,JHEP2019 no. 6, (2019) 1–18

    work page 2019

  2. [1]

    C. A. Argüelles, T. Katori, and J. Salvado,Phys. Rev. Lett.115 no. 16, (2015) 161303

    work page 2015

  3. [2]

    IceCubeCollaboration, M. G. Aartsenet al., JINST12 no. 03, (2017) P03012

    2017

  4. [3]

    IceCubeCollaboration, M. G. Aartsenet al., Science342 no. 6161, (2013) 1242856

    work page 2013

  5. [4]

    T. K. Gaisser,Astropart. Phys.35no. 12, (2012) 801–806

    work page 2012

  6. [5]

    Riehn, H

    F. Riehn, H. P. Dembinski, R. Engel, A. Fedynitch, T. K. Gaisser, and T. Stanev,PoSICRC (2017) 301

    work page 2017

  7. [6]

    Usner and the IceCube Collaboration,PoSICRC(2017) 974

    IceCubeCollaboration, M. Usner and the IceCube Collaboration,PoSICRC(2017) 974

    work page 2017

  8. [7]

    IceCubeCollaboration, M. G. Aartsenet al., Eur. Phys. J. C82 no. 11, (2022) 1031

    work page 2022

Show all 20 references
  1. [9]

    IceCubeCollaboration, M. G. Aartsenet al., Astrophys. J.928 no. 1, (2022) 50

    2022

  2. [10]

    IceCubeCollaboration, M. G. Aartsenet al.,JCAP2019no. 10, (2019) 048

    2019

  3. [11]

    IceCubeCollaboration, M. G. Aartsenet al.,Phys. Rev. D104 no. 2, (2021) 022002

    2021

  4. [12]

    De Castro and T

    P. De Castro and T. Dorigo,Comput. Phys. Commun.244 (2019) 170–179

    2019

  5. [13]

    Janik,PoSICRC(2025) 1132

    IceCubeCollaboration, O. Janik,PoSICRC(2025) 1132

    2025

  6. [14]

    Chen,PoSICRC(2025) 1011

    IceCubeCollaboration, Z. Chen,PoSICRC(2025) 1011

    2025

  7. [15]

    IceCubeCollaboration, M. G. Aartsenet al.,Phys. Rev. Lett.125 no. 12, (2020) 121104

    2020

  8. [16]

    IceCubeCollaboration, M. G. Aartsenet al.,Phys. Rev. D91 no. 2, (2015) 022001

    2015

  9. [17]

    Balagopal, A

    IceCubeCollaboration, A. Balagopal, A. Karle, , and V. Basu,PoSICRC(2025) 983

    2025

  10. [18]

    Ishiharaet al., PoSICRC(2019) 1177

    IceCubeCollaboration, A. Ishiharaet al., PoSICRC(2019) 1177

    2019

  11. [19]

    Usner,Search for astrophysical tau-neutrinos in six years of high-energy starting events intheIceCubedetector

    M. Usner,Search for astrophysical tau-neutrinos in six years of high-energy starting events intheIceCubedetector . PhDthesis,Humboldt-UniversitätzuBerlin,Berlin, Germany,2018

    2018

  12. [20]

    IceCubeCollaboration, M. G. Aartsen and the IceCube Collaboration,PoSICRC(2013) 0580. 8 Neutrino flavor composition using High Energy Starting Events with IceCube Full Author List: IceCube Collaboration R. Abbasi16, M. Ackermann63, J. Adams17, S. K. Agarwalla39, a, J. A. Aguil...

    2013

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.