REVIEW 2 major objections 5 minor 1 cited by
A Novel On-Shell Recursive Relation
T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash
Pith's one-line read This paper derives on-shell recursion relations for biadjoint scalar and pure Yang-Mills tree amplitudes from double-cover CHY factorization, replacing off-shell currents with on-shell amplitudes, and factorizing BCJ numerators.
desk verdict Genuinely new on-shell recursion with solid low-point checks, but the central projection step—independence of factorized currents from the off-shell mass—is asserted rather than proved for general n. 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 engine of the construction is the "full common set" $K_n = \tilde{K}_n \cup \{s_{13}\}$: a set of $n(n-3)/2 - 1$ independent Mandelstam invariants, arranged in a polygon in the Gram matrix, that contains every pole appearing in the recursive factorization but excludes the effective mass variables $s_{2\ldots i}$ of the off-shell legs. The paper shows that, as a consequence of the pole structure in Eq. (32), each product of amputated currents $J_i \times J_{n+2-i}$ is independent of its own $s_{2\ldots i}$; this translational invariance allows one to fix $s_{2\ldots i}=0$, replacing the currents by on-shell amplitudes evaluated in $K_n$. For the longitudinal sector of pure Yang-Mills, the machinery also includes transmutation operators that rewrite $\mathrm{YM}\oplus\phi^3$ amplitudes as pure-gluon amplitudes, plus completeness relations over transverse polarizations.
What would settle it
Evaluate the derivative $\partial/\partial s_{2\ldots i}$ of the product of amputated currents appearing in Eq. (21) at a generic six- or seven-point kinematic point; if it is nonzero, the on-shell replacement in Eqs. (40) and (79) fails. Alternatively, compute the right-hand side of Eq. (40) at $n=6$ and compare it with the known CHY or Berends-Giele six-point amplitude; any disagreement would disprove the recursion.
Extended reading notes
Core claim
The central result is an on-shell recursion relation for color-ordered tree amplitudes. For biadjoint scalar theory, Eq. (40) expresses the $n$-point amplitude as a sum over channels of products of lower-point on-shell amplitudes, $A_i(2,3,\ldots,\kappa_i)\times A_{n+2-i}(1,\kappa'_i,\ldots,n)$, each evaluated in the common kinematic set $K_n$ with the off-shell mass variable set to zero; the first channel uses the three-point amplitude $A_3(1,2,\zeta)$ times the $(n-1)$-point amplitude. For pure Yang-Mills, Eq. (79) has the same structure, with the longitudinal/contact sector converted into on-shell pure-gluon amplitudes by a linear operator $\mathcal{O}_i$ built from polarization completeness and transmutation operators. The paper shows at four and five points that the BCJ numerators obtained this way take an explicitly factorized form, as products of lower-point BCJ numerators summed over intermediate polarizations, and that spurious poles cancel in the final amplitude.
Load-bearing premise
The load-bearing premise is that every product of off-shell currents $J_i \times J_{n+2-i}$ is independent of the Mandelstam variable $s_{2\ldots i}$ that acts as the effective mass of the off-shell legs; the paper verifies this at four and five points and for the non-longitudinal Yang-Mills identities up to eight points, but does not prove it for arbitrary $n$.
Editorial extensions
If this is right
- Tree amplitudes in Tr(phi^3) and pure YM can be computed from lower-point on-shell amplitudes alone, without deforming momenta or computing boundary integrals.
- BCJ numerators produced by the recursion are manifestly factorized into products of lower-point BCJ numerators, giving a direct on-shell route to color-kinematics-satisfying numerators and hence to gravity amplitudes via the double copy.
- The BCFW boundary contribution is identified with an on-shell amplitude evaluated in the common set (Eq. (47)), so boundary terms can be computed without a large-z analysis.
- The mechanism of spurious-pole cancellation is claimed to be the same as in BCFW, giving a built-in consistency check at every order.
- The construction is stated to extend to arbitrary gauge choices and, in ongoing work, to fermions, non-linear sigma model, and higher-derivative theories.
Reading between the lines
- If the derivative-independence premise holds at all multiplicities, $K_n$ defines a canonical slice of kinematic space where off-shell currents and on-shell amplitudes coincide; that identification may be provable directly from the double-cover CHY pole structure, which would turn the checked recursion into a theorem.
- A natural stress test is to generate six-point BCJ numerators from Eq. (79) and check they obey the BCJ relations and the amplitude master formula; the paper demonstrates factorized numerators only through five points.
- The same common-set independence may hold for NLSM or (DF)^2 currents, whose CHY pole structures are similar; if the projection works there, the recursion would give a uniform on-shell construction across effective field theories.
- Combined with BCFW, the recursion could be used to compute BCFW boundary terms at high points, where the large-z behavior is hard to determine, by evaluating a lower-point amplitude in the common set.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes new on-shell recursion relations for biadjoint scalar (BAS) and pure Yang-Mills (YM) tree amplitudes, derived from the double-cover CHY factorization formulas of the author's earlier work. The key idea is to introduce a 'common set' K_n of independent Mandelstam invariants such that each off-shell factorization product of amputated currents is independent of the remaining Mandelstam variable, namely the effective mass squared of the off-shell legs. Setting that variable to zero projects the currents onto on-shell amplitudes, giving the new recursions (40) and (79). As a byproduct, the paper derives factorized BCJ numerators, which are checked against the literature at five points. The paper also compares the BAS recursion with Berends-Giele and BCFW recursion, and includes four- and five-point checks for BAS and YM, with higher-point support delegated to a GitHub notebook.
Significance. If the central projection property holds, this is a genuinely new way to organize tree-level on-shell recursion and to generate BCJ numerators in factorized form, with potential applications to double-copy and to other theories. The paper's strengths include explicit low-point checks against independent known amplitudes, agreement of the five-point BCJ numerators with Ref. [18], and the provision of machine-checkable support material in Ref. [17]. The construction is not circular: the final recursion is tested against amplitudes and numerators obtained by independent methods. However, the main claim rests on an unproved kinematic-independence property for general multiplicity, which makes the current version incomplete as a general derivation.
major comments (2)
- [Section III (paragraph after Eq. (36)) and Section IX] The on-shell projection is the load-bearing step of the paper, but it is not proven for general n. The text states that J^{φ3}_{n+2−i} × J^{φ3}_i is independent of the Mandelstam variable s_{2...i} (the effective mass of the off-shell legs) and therefore can be set to zero. The only explicit evidence is the four- and five-point checks in Eqs. (37)–(39). The pole-structure analysis in Eq. (32) shows which denominators appear after eliminating k_n, but it does not constrain the residues/numerators, which can in principle depend on s_{2...i} through the shifted scattering equations (19)–(20), where Δ_{ab} contains k_χ^2 = s_{2...i}. Since Eqs. (40) and (79) require this independence for every i = 3, ..., n−1, the recursions are not established for general n. Section IX states ∂_{s_{23...i}}(J_i × J_{n+2−i}) = 0 as a result without providing a proof or a complete higher-point analytic check. Please supply either a general proof of this independence from the double-cover CHY representation, or at minimum a complete analytic six-point verification of every current product appearing in Eq. (14), and state the status of the property (proven versus checked) explicitly.
- [Section VI (Eqs. (78)–(80))] The Yang-Mills longitudinal-sector step is asserted rather than derived. The text says 'Proceeding as in the previous section ... we arrive at the general expressions' (Eq. (78)), but the mechanism by which the transmutation operators T, the subset sums over δ_i and α_i, and the operator O_i in Eq. (80) are selected is not demonstrated in the paper. The four- and five-point examples in Sections V and VIII are consistent, and the six- and seven-point checks are delegated to the external notebook [17], but the central YM recursion (79) is therefore not independently verifiable from the manuscript. Please include either a derivation of Eq. (78) or a self-contained summary of the six-point verification (for example, a table showing that each term of Eq. (79) reproduces the known six-point amplitude) so that the general formula is supported by evidence available in the text.
minor comments (5)
- [Section III (Eq. (33))] The Gram matrix is described with 'blue shading', which is not visible in a monochrome printout; please replace the color cue with an explicit list of the entries in K_n or with a different visual marker.
- [Section II (Eq. (14)) and Section III (Eq. (40))] The first denominator is written as s_{34...n} in Eq. (14) but as s12 in Eq. (40); the equality s_{34...n} = s12 follows from momentum conservation and should be stated explicitly to avoid apparent inconsistency.
- [Section IV (Eq. (54))] The notation |_{1↔2} is used in the last line of Eq. (54) before its definition in the following paragraph; please define it at first use or add a footnote.
- [Section VI (Eq. (77))] The sentence 'These relations have been explicitly verified up to eight-point amplitudes' is not substantiated in the text, since the verification is delegated to an external notebook [17]. Please either include a brief description of the verification method and range, or soften the claim to 'checked numerically' with an explicit reference to the notebook.
- [Throughout] There are several typos and grammatical slips, e.g., 'previouly' in Section V, 'Let us first to recall' in the same section, and a double comma in Eq. (77). Please proofread the manuscript.
Circularity Check
No circularity: the off-shell to on-shell replacement is checked against independent BG/BCFW results and known BCJ numerators, and the common set is fixed by pole structure rather than fitted.
full rationale
The derivation chain starts from the author's earlier double-cover factorization formulae [5,6], but the novel step is the replacement of products of amputated currents by products of on-shell amplitudes restricted to the common set K_n. That replacement is not an input: it is explicitly verified at four and five points for BAS (Eqs. 36-39) and in the Yang-Mills sector (Eqs. 65-66, 73), and the final five-point BCJ numerators are compared with the independent construction of Ref. [18] (Eq. 102). The common set K_n is defined from the pole structure of the recursive currents in Eq. (32), not fitted to reproduce known amplitudes, and the recursion is benchmarked against Berends-Giele and BCFW results (Eqs. 28-31, 46-51). Self-citations to [5,6] are prior published derivations with stated assumptions and are externally checkable; they do not reduce to the target result. The main weakness is that the kinematic independence asserted in Section IX, ∂_{s_{2...i}}(J_i times J_{n+2-i}) = 0, is verified at low points but not proved for general n; however, an unproven hypothesis is a correctness gap, not a circular step. No fitted parameter is renamed as a prediction, and no load-bearing argument reduces by construction to its own input.
Assumptions & free parameters
assumptions (5)
- domain assumption The double-cover CHY factorization formulae (Eq. 14 for BAS, Eq. 54 for YM) correctly compute amputated currents and amplitudes.
- domain assumption Naculich's kinematic-shift prescription (Eqs. 19-20) gives the correct amputated currents with up to three off-shell legs.
- ad hoc to paper The amputated currents are independent of the Mandelstam variables associated with the effective 'mass' of the off-shell legs (partial_{s_{2...i}}(J_i times J_{n+2-i}) = 0).
- domain assumption The Dong-He-Hou universal expansion (Eq. 67) and Cheung-Shen-Wen transmutation operators (Eqs. 70-71) apply to the present off-shell longitudinal currents.
- ad hoc to paper The gauge fixing (pqr|m) = (n12|3) is generic; the recursion extends to arbitrary gauge choices as sketched with the matrix (106).
Cite this review
Pith. "Pith review of A Novel On-Shell Recursive Relation." pith.science (2026). https://pith.science/paper/XMZXULWV
@misc{pith2026250723510,
author = {Pith},
title = {Pith review of: A Novel On-Shell Recursive Relation},
year = {2026},
howpublished = {\url{https://pith.science/paper/XMZXULWV}},
note = {Machine review of arXiv:2507.23510}
}
read the original abstract
We present a novel framework for deriving on-shell recursion relations, with a specific focus on biadjoint and pure Yang-Mills theories. Starting from the double-cover CHY factorization formulae, we identify a suitable set of independent kinematic variables that enables the reconstruction of amputated currents from amplitudes. As a byproduct, this new recursive structure recasts the BCJ numerators into an explicitly on-shell factorized form.
Figures
Forward citations
Cited by 1 Pith paper
-
On differential operators for scalar-scaffolded gluons
Differential operators on scalar-scaffolded variables extract individual phi^3 diagrams from gluon amplitudes, and the independent mixed amplitudes are counted by Catalan numbers.
Reference graph
Works this paper leans on
-
[18]
Unifying Rela- tions for Scattering Amplitudes,
C. Cheung, C. H. Shen and C. Wen, “Unifying Rela- tions for Scattering Amplitudes,” JHEP02(2018), 095 doi:10.1007/JHEP02(2018)095
-
[24]
Understanding zeros and splittings of or- dered tree amplitudes via Feynman diagrams,
K. Zhou, “Understanding zeros and splittings of or- dered tree amplitudes via Feynman diagrams,” JHEP03 (2025), 154 doi:10.1007/JHEP03(2025)154
-
[17]
H. Gomez, “Computer support material”. The Mathematica notebook can be downloaded from the website,https://github.com/humgomzu-dev/ Support-Material-A-novel-on-shell.git
-
[1]
New recursion rela- tions for tree amplitudes of gluons,
R. Britto, F. Cachazo and B. Feng, “New recursion rela- tions for tree amplitudes of gluons,” Nucl. Phys. B715 (2005), 499-522 doi:10.1016/j.nuclphysb.2005.02.030. R. Britto, F. Cachazo, B. Feng and E. Witten, “Direct proof of tree-level recursion relation in Yang- Mills theory,” Phys. Rev. Lett.94(2005) 181602 doi:10.1103/PhysRevLett.94.181602
-
[2]
On Tree Amplitudes in Gauge Theory and Gravity,
N. Arkani-Hamed and J. Kaplan, “On Tree Amplitudes in Gauge Theory and Gravity,” JHEP04(2008), 076 doi:10.1088/1126-6708/2008/04/076
-
[3]
Scattering equations and Kawai-Lewellen-Tye orthogonality,
F. Cachazo, S. He and E. Y. Yuan, “Scattering equations and Kawai-Lewellen-Tye orthogonality,” Phys. Rev. D 90(2014) 6, 065001; doi:10.1103/PhysRevD.90.065001; “Scattering of Massless Particles in Arbitrary Di- mensions,” Phys. Rev. Lett.113(2014) 17, 171601; doi:10.1103/PhysRevLett.113.171601; “Scattering of Massless Particles: Scalars, Gluons and Gravit...
-
[4]
=−1. The numerators are given by, 1 2 X δ∈{∅,{4}} (ϵ2 ·ϵ 5)(ϵ1 ·f δ ·ϵ 3)T (1↔2) [κ4,3,δ,2]N(κ4,2,4,3) = 1 2 (ϵ2 ·ϵ 5)(n(34) ·ϵ 1), 1 2 X δ∈{∅,{4}} (ϵ2 ·ϵ 5)(ϵ1 ·f δ ·ϵ 3)T (1↔2) [κ4,3,δ,2]N(κ4,4,2,3) = 1 2 (ϵ2 ·ϵ 5)(ϵ1 ·ϵ 3)(ϵ4 ·k 13).(100) Finally, by assembling all these on-shell contributions in Eq. (76), we arrive at the expected result, A5(1,2,3,4,5...
-
[5]
Ambitwistor strings and the scattering equations,
L. Mason and D. Skinner, “Ambitwistor strings and the scattering equations,” JHEP1407, 048 (2014). doi:10.1007/JHEP07(2014)048. N. Berkovits, “Infinite Tension Limit of the Pure Spinor Superstring,” JHEP1403, 017 (2014). doi:10.1007/JHEP03(2014)017. H. Gomez and E. Y. Yuan, “N-point tree-level scattering amplitude in the new Berkovits‘ string,” JHEP1404, ...
Show all 31 references
-
[6]
Λscattering equations,
H. Gomez, “Λscattering equations,” JHEP1606(2016)
2016
-
[7]
Smoothly splitting amplitudes and semi-locality,
F. Cachazo, N. Early and B. Giménez Umbert, “Smoothly splitting amplitudes and semi-locality,” JHEP08(2022), 252 doi:10.1007/JHEP08(2022)252
2022 doi
-
[8]
New Factorization Relations for Yang Mills Am- plitudes,
N. E. J. Bjerrum-Bohr, P. H. Damgaard and H. Gomez, “New Factorization Relations for Yang Mills Am- plitudes,” Phys. Rev. D99(2019) no.2, 025014 14 doi:10.1103/PhysRevD.99.025014. H. Gomez, “Scatter- ing equations and a new factorization for ampli- tudes. Part I. Gauge theorie...
2019 doi
-
[9]
Recursive Calculations for Processes with n Gluons,
F. A. Berends and W. T. Giele, “Recursive Calculations for Processes with n Gluons,” Nucl. Phys. B306(1988), 759-808 doi:10.1016/0550-3213(88)90442-7
1988 doi
-
[10]
CHY representations for gauge theory and gravity amplitudes with up to three massive particles,
S. G. Naculich, “CHY representations for gauge theory and gravity amplitudes with up to three massive particles,” JHEP05(2015), 050 doi:10.1007/JHEP05(2015)050
2015 doi
-
[11]
Stringy canonical forms,
N. Arkani-Hamed, S. He and T. Lam, “Stringy canonical forms,” JHEP02, 069 (2021) doi:10.1007/JHEP02(2021)069 [arXiv:1912.08707 [hep- th]]. N. Arkani-Hamed and C. Figueiredo, “All-order splits and multi-soft limits for particle and string ampli- tudes,” [arXiv:2405.09608 [hep-t...
2021 arXiv
-
[12]
Berends-Giele recursion for double- color-ordered amplitudes,
C. R. Mafra, “Berends-Giele recursion for double- color-ordered amplitudes,” JHEP07(2016), 080 doi:10.1007/JHEP07(2016)080 [arXiv:1603.09731 [hep- th]]
2016 arXiv
-
[13]
Proof of the Formula of Cachazo, He and Yuan for Yang-Mills Tree Ampli- tudes in Arbitrary Dimension,
L. Dolan and P. Goddard, “Proof of the Formula of Cachazo, He and Yuan for Yang-Mills Tree Ampli- tudes in Arbitrary Dimension,” JHEP05(2014), 010 doi:10.1007/JHEP05(2014)010
2014 doi
-
[14]
New Relations for Gauge-Theory Amplitudes,
Z. Bern, J. J. M. Carrasco and H. Johansson, “New Relations for Gauge-Theory Amplitudes,” Phys. Rev. D78, 085011 (2008). doi:10.1103/PhysRevD.78.085011. Z. Bern, J. J. M. Carrasco and H. Johansson, “Per- turbative Quantum Gravity as a Double Copy of Gauge Theory,” Phys. Rev. L...
2008 doi
-
[15]
Tree level amplitudes from soft theorems,
K. Zhou, “Tree level amplitudes from soft theorems,” JHEP03(2023), 021 doi:10.1007/JHEP03(2023)021. C. Hu and K. Zhou, “Recursive construction for expansions of tree Yang–Mills amplitudes from soft theorem,” Eur. Phys. J. C84(2024) no.3, 221 doi:10.1140/epjc/s10052-024-12517-y
2023 doi
-
[16]
Universal expan- sions of scattering amplitudes for gravitons, glu- ons, and Goldstone particles,
J. Dong, S. He and L. Hou, “Universal expan- sions of scattering amplitudes for gravitons, glu- ons, and Goldstone particles,” Phys. Rev. D105 (2022) no.10, 105007 doi:10.1103/PhysRevD.105.105007 [arXiv:2111.10525 [hep-th]]
2022 arXiv
-
[19]
In preparation
N. E. J. Bjerrum-Bohr, H. Gomez, “In preparation”
-
[20]
Expan- sion of Einstein-Yang-Mills Amplitude,
C. H. Fu, Y. J. Du, R. Huang and B. Feng, “Expan- sion of Einstein-Yang-Mills Amplitude,” JHEP09, 021 (2017) doi:10.1007/JHEP09(2017)021 [arXiv:1702.08158 [hep-th]]. Y. J. Du and F. Teng, “BCJ numera- tors from reduced Pfaffian,” JHEP04(2017), 033 doi:10.1007/JHEP04(2017)033. ...
2017 arXiv
-
[21]
New factorizations of Yang-Mills am- plitudes,
Y. Zhang, “New factorizations of Yang-Mills am- plitudes,” Phys. Rev. D111(2025) no.8, 085004 doi:10.1103/PhysRevD.111.085004
2025 doi
-
[22]
A univer- sal splitting of tree-level string and particle scatter- ing amplitudes,
Q. Cao, J. Dong, S. He and C. Shi, “A univer- sal splitting of tree-level string and particle scatter- ing amplitudes,” Phys. Lett. B856(2024), 138934 doi:10.1016/j.physletb.2024.138934. Q. Cao, J. Dong, S. He, C. Shi and F. Zhu, “On universal splittings of tree-level particle...
2024
-
[23]
Hidden Zeros and2- split via BCFW Recursion Relation,
B. Feng, L. Zhang and K. Zhou, “Hidden Zeros and2- split via BCFW Recursion Relation,” [arXiv:2504.14215 [hep-th]]
-
[25]
Generalized2-split for higher- derivative YM and GR amplitudes at tree-level,
L. Zhang and K. Zhou, “Generalized2-split for higher- derivative YM and GR amplitudes at tree-level,” [arXiv:2601.00297 [hep-th]]
-
[26]
On- shell representation and further instances of the 2- split behavior of amplitudes,
T. Azevedo, H. Gomez and R. Lipinski Jusinskas, “On- shell representation and further instances of the 2- split behavior of amplitudes,” JHEP06(2026), 001 doi:10.1007/JHEP06(2026)001 [arXiv:2512.20790 [hep- th]]
2026 arXiv
-
[27]
Ambitwistor strings and the scattering equations on AdS3×S3,
K. Roehrig and D. Skinner, “Ambitwistor strings and the scattering equations on AdS3×S3,” JHEP02(2022), 073 doi:10.1007/JHEP02(2022)073
2022 doi
-
[28]
Scat- tering equations in AdS: scalar correlators in arbitrary dimensions,
L. Eberhardt, S. Komatsu and S. Mizera, “Scat- tering equations in AdS: scalar correlators in arbitrary dimensions,” JHEP11(2020), 158 doi:10.1007/JHEP11(2020)158
2020 doi
-
[29]
Cosmologi- cal Scattering Equations,
H. Gomez, R. L. Jusinskas and A. Lipstein, “Cosmologi- cal Scattering Equations,” Phys. Rev. Lett.127(2021) no.25, 251604 doi:10.1103/PhysRevLett.127.251604. H. Gomez, R. Lipinski Jusinskas and A. Lipstein, “Cos- mologicalscatteringequationsattree-levelandone-loop,” JHEP07(202...
2021 doi
-
[30]
BCJ amplitude relations for Anti-de Sitter bound- ary correlators in embedding space,
P. Diwakar, A. Herderschee, R. Roiban and F. Teng, “BCJ amplitude relations for Anti-de Sitter bound- ary correlators in embedding space,” JHEP10 (2021), 141 doi:10.1007/JHEP10(2021)141. A. Herder- schee, R. Roiban and F. Teng, “On the differ- ential representation and color-k...
2021 doi
-
[101]
doi:10.1007/JHEP06(2016)101
2016 doi
Reviewed August 6, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.