REVIEW 3 major objections 5 minor 15 references
F2F, a model-independent method to determine the mass and width of a particle in the presence of interference
T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read A Fourier-fit method, F2F, recovers a particle's mass and width from its measured mass distribution even when interference hides the resonance, without modeling the signal amplitude or phase.
desk verdict Clever Fourier trick for extracting resonance parameters under interference, but the claimed model-independence is not yet backed by proof or enough toy diversity. 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 load-bearing object is the cosine Fourier coefficient c_k of the effective signal distribution. The analytic identity obtained by contour integration relates c_k to the squared Breit-Wigner propagator: the coefficient is proportional to a damped cosine whose damping rate is set by Gamma and whose oscillation frequency is set by M. This identity is what lets the fit separate the two parameters without knowing the interference phase; the slow-varying envelope f(x) only contributes to low modes, where a polynomial A(k) absorbs it, while the high modes carry the propagator's signature. In practical fits the working formula is c_k(M,Gamma,A,B,C) = A(k) + B e^{-(k'$\sigma$)^2/2 - k' $\sqrt$(gamma_0) sin($\theta$/2)} cos(k' $\sqrt$(gamma_0) cos($\theta$/2) - C), with the detector-resolution factor included.
What would settle it
Generate a toy sample with a known mass and width but with an interference phase that changes sharply across the fit range, apply F2F, and check whether the fitted M and Gamma deviate from the inputs; a clear deviation would map the boundary of the method's validity, while no deviation would mean the slow-envelope assumption is weaker than stated.
Extended reading notes
Core claim
The paper's central claim is that the mass and width of a resonance can be determined from the Fourier coefficients of the data-minus-background mass distribution even when the resonance and background amplitudes interfere with an unknown, possibly x-dependent phase. Writing the effective signal as f(x)/(($x^{2}$ - $M^{2}$)^2 + $M^{2}$ $Gamma^{2}$) and assuming f(x) varies slowly, the cosine coefficients c_k satisfy c_k proportional to (pi/$gamma_0^{{3/2}}$) sin($\theta$) e^{-|k'| $\sqrt$(gamma_0) sin($\theta$/2)} cos(|k'| $\sqrt$(gamma_0) cos($\theta$/2) - $\theta$/2), with gamma_0 = $\sqrt$($M_0^{2}$($M_0^{2}$+$Gamma^{2}$)), tan($\theta$) = Gamma/M_0, and M_0 = M - x_min. This analytic form carries both an exponential decay set by Gamma and an oscillation set by M, so a fit of this form to the measured coefficients yields M and Gamma directly. Detector smearing enters as a multiplicative Gaussian factor $e^{{-(k'sigma)^2/2}}$ per mode, and slowly-varying background mismodeling is absorbed by a low-order polynomial in mode number. In toy experiments with Gamma = 10 GeV and M = 125 GeV, the F2F fits agree with the inputs within uncertainties for all tested interference phases, and in the Higgs-width application the method improves the 90% confidence upper limit from 1.54 GeV with a nominal on-shell fit to 0.92 GeV with a long-range F2F fit.
Load-bearing premise
The load-bearing premise is that f(x), the slowly-varying envelope multiplying the squared Breit-Wigner propagator, does not vary much across the fitted mass range, so its Fourier transform is concentrated at low modes and the high-mode data-minus-background coefficients follow the propagator's analytic form.
Editorial extensions
If this is right
- For a narrow resonance like the Higgs boson, the width sensitivity is set by the balance between the detector-resolution damping and the Gamma-driven damping, giving a crude estimate Gamma approximately k pi sigma^2 / L; the paper shows F2F applied to the four-muon channel in a long mass range yields a smaller 90% C.L. upper limit on Gamma_H than a nominal on-shell fit.
- The same Fourier-coefficient fit can be used as a resonance-search statistic: fitting with and without a resonance hypothesis on a grid of (M,Gamma) and taking the chi-square difference gives local p0 values that identify an injected 1 TeV, 50 GeV resonance.
- The method requires no signal Monte Carlo or model of the interference, so it can be applied when the new particle's couplings and phase are unknown.
- Extending the fitted mass range improves width sensitivity because more modes sample the exponential decay, provided the slow-envelope and MC-statistics conditions are met.
- Using more modes than L/sigma is recommended, so the fit includes the region where the resolution factor has not yet suppressed all information.
Reading between the lines
- A natural stress test is to let the interference phase delta(x) vary rapidly across the fitted range, making f(x) no longer slow; the paper's own assumption predicts biased M and Gamma there, so an adaptive range or mode-truncation rule could be developed.
- The same coefficient formula may generalize to other observables where a Breit-Wigner-like propagator controls a distribution, such as energy-dependent line shapes or time-dependent oscillation envelopes, as long as the slow-envelope condition holds.
- Because background mismodeling enters only through the low-mode polynomial A(k), the method could in principle use a data-driven background estimate rather than Monte Carlo, as long as its high-mode coefficients are negligible.
- The periodicity implied by the cosine series creates boundary artifacts; the paper checks that these are confined to edges, which suggests an explicit edge-exclusion or windowing procedure would make the method more robust in low-statistics searches.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes F2F, a method to extract the mass M and width Gamma of a Breit-Wigner resonance from the Fourier coefficients of an invariant-mass distribution in the presence of coherent signal-background interference. After subtracting a known background, the effective signal is expanded in cosine modes; Eq. (15) models the k-dependence of the coefficients using the exact Fourier transform of the squared propagator, with a polynomial A(k) and a global amplitude B and phase C absorbing model dependence. The method is validated on toy experiments for four constant phase angles and three detector resolutions, and is applied to the SM Higgs width measurement and to a new-resonance search. The contour integration leading from Eq. (5) to Eq. (6) is correct for an idealized constant envelope, and the toy studies are genuine closure tests; the central open question is whether the 'slow-varying f' approximation is quantitatively controlled.
Significance. If the envelope approximation can be justified, F2F is a valuable addition to the hep-ph toolbox: it avoids explicit signal modeling, includes interference at the template level, treats detector smearing analytically, and the Higgs-width application suggests competitive sensitivity. The paper's concrete strengths are the clean contour-integral derivation of Eq. (6), the analytical resolution factor in Eq. (7), and closure tests over several phases and resolutions; no machine-checked proofs or code are provided, but the toy studies are described in sufficient detail to be reproduced. The main limitation is that the central 'model-independent' claim currently rests on an unquantified slow-variation assumption; as written, the significance of the method is conditional on that assumption being satisfied or empirically demonstrated.
major comments (3)
- [Sec. II, Eqs. (2), (5), (6), (15); Sec. III] The transition from Eq. (5) to Eq. (6) simply drops f(x) after the sentence 'Assuming f(x) is a slow-varying function'. For the paper's own model in Eq. (2), f(x) is not slow-varying in the relevant sense: the interference term contains (x^2-M^2), whose variation over the resonance width is O(2M Gamma), which is not small compared with the denominator scale. The exact Fourier coefficient of the product f(x) G(x;M,Gamma) is a convolution of the transforms of f and the squared propagator; a single global amplitude B and phase C in Eq. (15) can absorb only the first few terms of a Taylor expansion of f. No quantitative condition on f, no truncation error bound, and no scan over envelope slope, curvature, or phase gradient are given. The Sec. III toys use one exponential envelope with tau=200 GeV and constant delta, so they do not exercise the claimed independence from the signal model. This is load-bearing: if f has k-dependent amplitude or phase corrections, the fitted M and Gamma will be biased.
- [Secs. II and IV-V, Eqs. (11)-(12)] The method assumes that the background Fourier coefficients are known exactly: c_k(seff) is the difference between data and background-sample coefficients. In the Higgs and search applications, the background MC uncertainty is indicated as a band but is explicitly not included in the fit, and the polynomial A(k) in Eq. (15) is asserted to absorb background mismodeling without demonstration. A background shape error with power at high Fourier modes cannot be absorbed by a low-order polynomial and would be misinterpreted as resonance information. A systematic study of background-modeling uncertainties, or an explicit statement that the method requires essentially perfect background knowledge, is needed before the claimed precision can be taken at face value.
- [Sec. V, Eq. (19), Fig. 14] The search application computes local p0 values with 2 degrees of freedom at each point of a mass-width grid but does not address the trials factor from scanning the grid. The background-only toy in the left panel of Fig. 14 shows an apparent feature near (1.3 TeV, 5 GeV); if this is the maximum local p0 over the scanned grid, the global significance should be reported, or the paper should explicitly restrict the claim to local p0. Without this, the statistical interpretation in a search context may be overstated.
minor comments (5)
- [Throughout] The manuscript contains numerous typographical errors and language issues, including 'reosnant', 'Stardard-Model', 'inteference', 'hardon colliders', 'responce', 'seires', 'indepdent', 'siganl', 'measurment', and 'seaches'. A thorough language editing pass is needed.
- [Ref. [9]] Reference [9] cites a Wikipedia page for contour integration; a standard textbook reference would be more appropriate for a journal submission.
- [Sec. IV] The text says 2000 toy data samples are generated, but Figs. 11 and 12 show distributions from 1000 toys; please clarify the sample sizes and state whether the reported upper limits are based on 1000 or 2000 toys.
- [Eq. (13)] The PDF defined in Eq. (13) is not guaranteed to be positive for all x when the effective signal yield c0 is negative or when the Fourier template dips; the fit should either enforce positivity or restrict the fit range, and this should be stated explicitly.
- [Eq. (17)] The estimate Gamma = k pi sigma^2 / L in Eq. (17) is presented as a sensitivity estimate without specifying the mode k at which the two exponential terms are comparable; please state that this is a heuristic order-of-magnitude estimate and define k.
Circularity Check
No significant circularity: the Fourier-coefficient template is derived from the Breit-Wigner propagator by contour integration, with M and Gamma as free fit parameters, and no load-bearing self-citation chain is present.
full rationale
The derivation chain is self-contained. The paper starts from the exact expansion of the squared amplitude, Eq. (2), defines the effective signal as f(x)/((x^2-M^2)^2 + M^2 Gamma^2), and then derives the Fourier coefficient relation, Eq. (6), by contour integration of the squared propagator. The slow-varying assumption on f(x) is a stated modeling assumption, not a circular reuse of the target result; it is externally testable through the toy experiments and through the fitted B and C parameters, which absorb only the overall amplitude and phase of the slowly varying envelope and do not encode M or Gamma. Equation (15) is a fit template in which A(k), B, and C are nuisance parameters while M and Gamma enter only through the propagator-derived exponential and cosine factors. The toy studies are closure tests: data are generated from known M and Gamma, and the fits recover those inputs; this is validation, not prediction-from-fit. The Higgs-width and search applications also generate toy data and compare fitted results with inputs, so no fitted parameter is renamed as a prediction. There are no uniqueness theorems imported from the author's prior work, no self-citation is load-bearing, and no known result is merely relabeled. Possible concerns about the slow-varying condition being too strong, or about the single global amplitude and phase not covering general production amplitudes, are accuracy or model-dependence risks, not circularity. Therefore the circularity score is 0.
Assumptions & free parameters
free parameters (3)
- B =
floating in fit
- C =
floating in fit
- A(k) polynomial coefficients =
floating in fit
assumptions (3)
- domain assumption The signal amplitude has the relativistic Breit-Wigner propagator form 1/(x^2-M^2+iM*Gamma)
- domain assumption The function f(x) in Eq. 2 is slowly varying, so its Fourier transform is concentrated at low modes and the high-mode coefficients follow the propagator's Fourier transform
- domain assumption The background shape and yield are known precisely (from MC or sidebands), so c_k(Data)-c_k(Bkg) isolates the effective signal
Cite this review
Pith. "Pith review of F2F, a model-independent method to determine the mass and width of a particle in the presence of interference." pith.science (2026). https://pith.science/paper/VV2DP3ZR
@misc{pith2026190804048,
author = {Pith},
title = {Pith review of: F2F, a model-independent method to determine the mass and width of a particle in the presence of interference},
year = {2026},
howpublished = {\url{https://pith.science/paper/VV2DP3ZR}},
note = {Machine review of arXiv:1908.04048}
}
abstract
It is generally believed that any particle to be discovered will have a TeV-order mass. Given its great mass, it must have a large decay width. Therefore, the interference effect will be very common if they and the Standard-Model (SM) particles contribute to the same final state.However, the interference effect could make a new particle show up not like a resonance, and it is difficult to search and measure its properties. In this work, a model-independent method, F2F (Fit To Fourier coefficients), is proposed to search for an unknown resonance and to determine its mass ($M$) and width ($\Gamma$) in the presence of interference. Basically we express the sum of reosnant signal and the interference as a cosine Fourier series and relate the Fourier coefficients with the mass and width. The relation is based on the general propagator form, $1/(x^2-M^2+iM\Gamma)$. Thus it does not need any signal model. Toy experiments show that the obtained mass and width agree well with the inputs with a similar precision as using an explicit signal model. We also show that we can apply this method to measure the Stardard-Model Higgs width and to make statistic interpretation in searching for new resonance allowing for interference.
Figures
Figures from the paper (11 more)
Reference graph
Works this paper leans on
-
[9]
see for example https://en.wikipedia.org/wiki/Contour_integration
-
[1]
BESIII Collaboration, Phys. Rev. D91 (2015) 112001, arXiv:1504.03194
arXiv 2015
-
[2]
ATLAS Collaboration, JHEP 1904 (2019) 048, arXiv:1902.05892
arXiv 2019
-
[3]
ATLAS Collaboration, JHEP 1905 (2019) 164, arXiv:1812.07343
work page Pith review arXiv 2019
-
[4]
CMS Collaboration, JHEP 1806 (2018) 127, Erratum: JHEP 1903 (2019) 128, arXiv:1804.01939
arXiv 2018
-
[5]
CMS Collaboration, CERN-EP-2019-147, arXiv:1908.01115
arXiv 2019
-
[6]
Interference effect on heavy Higgs resonance signal in \gamma\gamma and ZZ channels
S. Jung, Y.W. Yoon, and J. Song, Phys. Rev. D93 (2016) 055035, arXiv:1510.03450
work page Pith review arXiv 2016
-
[7]
A. Djouadi, J. Ellis, A. Popov, J. Quevillon, JHEP 1903 (2019) 119, arXiv:1901.03417 17
arXiv 2019
Show all 15 references
-
[8]
Kauer, A
N. Kauer, A. Lind, P. Maierh¨ ofer, and W. Song, JHEP 1907 (2019) 108, arXiv:1905.03296
2019 arXiv
- [10]
- [11]
-
[12]
Alwell et al., JHEP 1407 (2014) 079, arXiv:1405.0301
J. Alwell et al., JHEP 1407 (2014) 079, arXiv:1405.0301
2014 arXiv
-
[13]
ATLAS Collaboration, Phys. Lett. B 786 (2018) 223, arXiv:1808.01191
2018 arXiv
-
[14]
CMS Collaboration, Phys. Rev. D 99, 112003, arXiv:1901.00174
1901 arXiv
-
[15]
CMS Collaboration, JHEP 11 (2017) 047, arXiv:1706.09936, see additional figures in http: //cms-results.web.cern.ch/cms-results/public-results/publications/HIG-16-041 18
2017 arXiv
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.