Pith. sign in

REVIEW 3 major objections 5 minor 62 references

A Demonstration of ARCANE Reweighting: Reducing the Sign Problem in the MC@NLO Generation of $e^+ e^- \rightarrow q \bar{q} + 1\, jet$ Events

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

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

Pith's one-line read ARCANE reweighting cuts the negative-weight fraction in NLO $e^+e^-\to q\bar{q}+1$ jet generation from 2.25% to $5.77\times 10^{-6}$ while provably preserving all visible distributions.

desk verdict A careful, genuinely useful proof-of-principle that ARCANE reweighting kills negative weights in a concrete MC@NLO process; the no-bias proof is solid but rests on an unverified density-extraction assumption. read the letter →

arxiv 2502.08053 v1 pith:NKVLWIQE submitted 2025-02-12 hep-ph hep-ex

Prasanth Shyamsundar This is my paper · ORCID
classification hep-phhep-ex
keywords ARCANEreweightingnegativeweightssignproblemMC@NLOpartonshowerSudakovvetoalgorithmCatani-SeymoursplittingkernelsNLOeventgeneration
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

ARCANE reweighting is a Monte Carlo technique aimed at the sign problem of NLO event generation: negatively weighted MC@NLO events force collider simulations to generate far more events than the physics alone requires. The technique changes each event's weight by an additive term that shifts contribution between generator histories — standard (S) versus hard-remainder (H) pathways — that end in the exact same visible final state, and the shift is engineered to integrate to zero over the hidden pathway information, so every visible distribution is untouched by construction. For $e^+e^- \to q\bar{q} + 1$ jet events, applying it to a tutorial MC@NLO generator cuts the post-unweighting negative-event fraction from 2.25% to $5.77\times 10^{-6}$, with $\chi^2$ comparisons showing no bias even when the redistribution kernel is deliberately a poor constant approximation. The process is admittedly a mild case of the sign problem, but it has the same structure as harder hadron-collision cases, and the demonstration supplies a concrete template — choose a stopping point, enumerate the pathways to each visible event, redistribute — that the author argues carries over to those cases and to higher orders.

What carries the argument

The load-bearing object is the ARCANE redistribution function, written as $G(V,\mathrm{H}) = -\Lambda(V)$ for hard-remainder events and $G(V,\mathrm{S},\mathrm{rej\text{-}list}) = +\Lambda(V)\,\Phi_{\mathrm{un}}(\mathrm{rej\text{-}list}; V)$ for standard events, where the zero-integral condition $\int d\ell\, G(v,\ell) = 0$ is exactly what forces $F^{(\mathrm{ARCANE})}(V)=F^{(\mathrm{ORIG})}(V)$. The obstacle is that the optimal transfer contains the Sudakov form factor $\Delta_{\{f\}}$, which the veto algorithm exists precisely because it cannot be computed; ARCANE sidesteps this by replacing the splitting kernels $\{f\}$ with fitted redistribution kernels $\{h\}$ inside the pathway probabilities, so that the unit normalization of $\Phi_{\mathrm{un}}$ is exact by construction and only the efficiency, never the correctness, depends on the fit. The fits are two one-dimensional polynomials — $\hat{A}(\ell)$ for the running coupling and $\hat{B}(u)$ for the $z$-integrated Catani–Seymour kernel $\kappa_{f_{qq}}$, with the known limit $B(u)\sim (C_F/6\pi)\,u^{3/2}$ built in — making $\Delta_{\{h\}}$ closed-form and yielding the final weight formulas (80) and (91). The substrate is the generator's own sampling densities $P(V,\mathrm{H},c_{\mathrm{gen}})$ and $P(V,\mathrm{S},\mathrm{rej\text{-}list})$, with their Jacobian factors, which define the pathways the transfer moves weight across.

What would settle it

Run one very large dataset — $10^8$ events or more — twice, with identical random seeds except for the weighting scheme, and subtract the ARCANE-weighted and NBL-weighted histograms bin-by-bin over the joint space of the three invariant masses, thrust, and spherocity; the identity $F^{(\mathrm{ARCANE})}(V) = F^{(\mathrm{NBL})}(V)$ predicts every difference be consistent with statistical noise. The first bin deviating by more than five standard deviations — especially inside the low-$t$, low-$z$ corner where the residual negative events concentrate — would falsify the no-bias claim and point to the generator-density extraction (Equations 36 and 46) as the source.

Watch

Extended reading notes

Core claim

The paper shows that ARCANE reweighting can be engineered inside a realistic MC@NLO generator by treating the hidden event information as two families of histories — the two H-histories labeled by the emission channel $c_{\mathrm{gen}}$, and the continuum of S-histories labeled by the sequence of rejected proposals in the Sudakov veto algorithm — and then (i) merging the two H-histories into one by multichannel-style weighting, giving the NBL scheme, and (ii) transferring a contribution $\Lambda(V)$ from the merged H-branch onto the S-branches that end in the same visible event $V$, apportioned among them by $\Phi_{\mathrm{un}}(\mathrm{rej\text{-}list}; V)$. The central guarantee is that the marginal quasi-density is preserved identically, $F^{(\mathrm{ARCANE})}(V) = F^{(\mathrm{ORIG})}(V)$, whenever the transfer functions satisfy the zero-integral and coverage conditions — even though $\Lambda$ and $\Phi_{\mathrm{un}}$ are built from an approximate Sudakov form factor $\Delta_{\{h\}} \approx \Delta_{\{f\}}$ obtained from two one-dimensional polynomial fits. Empirically, the post-unweighting negative-weight fraction drops from 2.25% under NBL to $5.77\times 10^{-6}$ under ARCANE, and the remaining negative events are shown to sit precisely where the physical quasi-density $F^{(\mathrm{ORIG})}(V)$ itself is negative, concentrated at low emission scale $t$ and low $z$.

Load-bearing premise

The entire no-bias guarantee rests on the generator's internal sampling probabilities being captured exactly by the formulas the reweighting uses, including Jacobian factors and veto-pathway probabilities; the paper takes this extraction from the generator's internals as given, and the chi-square validation is its only check.

Editorial extensions

If this is right

  • The post-unweighting negative-event fraction falls from 2.25% (NBL) to $5.77\times10^{-6}$ (ARCANE), pushing the post-unweighting effective event fraction from 0.912 to 1.00; the paper estimates this cuts the downstream simulation cost by roughly 9% for this process, with larger savings wherever the sign problem is worse.
  • Every visible distribution — invariant masses, thrust, spherocity, flavor-tagged observables, any selection based on $V$ — is identical under ARCANE and ORIG weighting by construction, so the efficiency gain carries no reweighting bias.
  • The correctness of the method survives a deliberately bad redistribution kernel: the constant 'bad' kernel still passes the goodness-of-fit tests and still improves all sign-problem metrics, so imperfect fits degrade performance only, not validity.
  • The remaining negative weights are not a flaw of the reweighting: all 22 observed events with $W^{(\mathrm{ARCANE})}<0$ also have $W^{*}<0$, meaning the physical quasi-density of the visible event is itself negative there, so eliminating them requires redistributing across additional pathways (such as the two start-scale channels at a later stopping point) rather than better kernels.

Reading between the lines

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

  • The deeper trick is that the one object ARCANE never needs to compute is the Sudakov form factor $\Delta_{\{f\}}$ itself: any approximation $\Delta_{\{h\}}$ with the correct unit-normalization structure preserves the no-bias property, so the veto algorithm — the standard reason these weights are hard to control — becomes harmless.
  • The 'hard remainder spreading' variant sketched in the paper (sample only S-events, carry the H-contribution on them) could develop into a standalone replacement generator for processes whose sign problem is dominated by the H–S overlap, removing the resolved-event sampling step altogether.
  • Because fit quality controls only efficiency, the two polynomial fits could be replaced by per-event or per-region learned surrogates for $\kappa_{f_{qq}}$ without touching correctness; the paper's own bound on $W^{*}$ provides a built-in audit of how much efficiency any given surrogate leaves on the table.
  • The residual negatives at low $t$ and low $z$ predict where the physical cross-section itself changes sign; checking that prediction directly with independent high-statistics runs of the unweighted generator would test both the physics and the extraction in one go.
Share X Bluesky LinkedIn Reddit HN

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 paper demonstrates ARCANE reweighting for e+e- → q qbar + 1 jet generation in a tutorial MC@NLO-style generator. The technique additively modifies event weights so that contributions are moved between 'hard remainder' (H) and 'standard' (S) Monte Carlo histories that lead to the same visible event, while provably preserving the marginal visible-event quasi density F(V) (Eqs. (62)-(64)). The implementation requires constructing a redistribution kernel whose Sudakov-type integral can be computed analytically; the paper uses polynomial fits to the z-integrated Catani-Seymour kernel, and also studies a deliberately bad constant kernel. Using a medium dataset, Table 2 reports a post-unweighting negative-event fraction of 5.77e-6 for ARCANE versus 2.25% for the NBL weighting, and Figure 9 compares binned visible distributions to argue that no bias is introduced. The paper concludes that the sign problem is almost completely eliminated for this process, with residual negative weights arising where the physical F(V) itself is negative.

Significance. If the quantitative claims hold, this is a useful and carefully explained proof-of-concept: the marginal-preservation property is derived explicitly rather than assumed, the 'bad kernel' stress test is a genuine falsifiability check, and the code and data are made publicly available. The paper is also honest about several limitations, notably that the technique is applied only in the overlapping region of H- and S-events and that some class-2 H-events with negative weights remain. The central idea is not tied to this specific process, and the discussion in Section 6 gives a fair account of what would be needed for hadron collider applications. The main weakness is that the headline numerical claims are not accompanied by uncertainty estimates, and the no-bias validation is less powerful than the algebraic guarantee appears to require.

major comments (3)
  1. [Table 2 and Eqs. (114a)-(114e)] The proof that F^ARCANE(V)=F^ORIG(V) is conditional on exact knowledge of the generator-internal sampling densities P(V,H,cgen) and P(V,S,rej-list), including the Jacobian factors in Eq. (46) and the veto-pathway probability in Eq. (82). The paper states that these functions 'can be extracted from the internals of the base event generator', but provides no independent check of that extraction. The only empirical validation, Figure 9, has 20 bins per variable, uses the approximate error estimate in Eq. (119), and reports (chi^2 - nbins)/sqrt(2 nbins) = 2.88 for the Mqg distribution, which is attributed to statistical fluctuation. A single 3-sigma among five correlated variables is not decisive by itself, but it is also not a confirmation of exactness. I would like to see a direct code-level verification of at least one of the nontrivial ingredients, for example the normalization identities (43)-(44), the Jacobian factor (1-y_for_scale)/(1-y_gen) * y_for_scale, or a comparison of the extracted P(V,S,rej-list) with the empirical distribution of veto histories. This would substantially strengthen the central no-bias claim.
  2. [Abstract and Section 4] The headline performance metrics are reported without uncertainties. Footnote 36 says the uncertainties are 'expected to be sufficiently small', but no supporting calculation is given. This matters because the post-unweighting negative fraction fneg=5.77e-6 is based on only 22 negative-weight events in the medium dataset, so even a Poisson counting uncertainty alone is at the ~20% level, and the actual uncertainty from weight fluctuations could be larger. The derived quantities SP and f_effective inherit this uncertainty. I recommend adding bootstrap or analytic error bars to Table 2, or at least stating explicitly which entries are to be read as indicative rather than as precise performance benchmarks.
  3. The statement that ARCANE 'almost completely eliminates the negative weights problem' is stronger than what the paper actually demonstrates. The redistribution is performed only for class-1 H-events and class-4 S-events; class-2 H-events keep their original (sometimes negative) weights, as acknowledged in the 'Remaining negatively weighted events' discussion. If Table 2 is restricted to classes 1 and 4, this should be stated in the table caption and in the abstract; if it includes all classes, the contribution of class-2 negative weights to the total fneg should be quantified. Otherwise the reader cannot tell whether the claimed 5.77e-6 applies to the full process or only to the overlap region.
minor comments (5)
  1. [Line before Eq. (70)] The text says 'using the fact that ℓ > ℓ_start for events of class 1 and 4', but for these classes the first shower emission satisfies ℓ_start > ℓ > ℓ_cutoff. The inequality should be ℓ_start > ℓ.
  2. [Figure 9] The third row of Figure 9 is labeled 'Mqg/Ecms', duplicating the second row; based on the text it should be the invariant mass of the qbar-g system, and the row label should be corrected accordingly.
  3. [Reference [12]] The companion ARCANE paper is cited as '2501.XXXXX'. If it has by now been assigned an arXiv number, the placeholder should be updated, since several results of the present paper explicitly depend on definitions from Ref. [12].
  4. [Figure 8] The axis label '1 1 Thrust' appears to be a rendering artifact; it should read 1 - sqrt(1 - Thrust), matching the variable definition in the text.
  5. [Footnote 36] The footnote dismisses metric uncertainties without justification. Even if a full uncertainty propagation is deferred, one sentence indicating the source of the 'expected small' error would help the reader assess the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the visible-distribution preservation is an algebraic identity re-derived in this paper, and the fitted polynomials target a known theory kernel rather than the reported negative-weight metrics.

full rationale

The paper's central no-bias guarantee, F^ARCANE(V) = F^ORIG(V), is not an empirical prediction fitted to data but an algebraic identity: Eqs. (1)-(6) define the ARCANE weight shift through a redistribution function with zero latent-space integral, and Eqs. (62)-(64) re-derive the marginal preservation explicitly for the H/S redistribution used here. The polynomial fits of Section 3.5 approximate the known Catani-Seymour kernel kappa_fqq, not the reported negative-weight fractions or histogram heights; the deliberately bad kernel kappa_h^bad in Eq. (107) is then used to show that the no-bias property holds independently of approximation quality, and the resulting f_neg changes substantially (Table 2), so the headline reduction is not forced by construction. The companion paper [12] by the same author is cited for the general ARCANE framework and for some discussion of dependent resampling methods, but the load-bearing identities used in this demonstration are re-derived within the paper rather than imported as unverified external claims. The main remaining caveat is that the no-bias argument assumes the expressions for generator-internal densities P(V,H,cgen) and P(V,S,rej-list) exactly describe the implemented generator, and the empirical chi-square test has limited power (one ~3-sigma deviation in Mqg is attributed to fluctuations); this is a robustness/correctness concern about an assumption, not a circular reduction of the paper's conclusions to its inputs.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim does not introduce new physics entities. It rests on the standard MC@NLO and Catani-Seymour setup and on the exactness of generator-internal sampling densities. The only ad hoc construction is the approximate redistribution kernel kappa_h, whose coefficients are fitted to known functions and which affects efficiency, not unbiasedness.

free parameters (4)
  • Polynomial coefficients {a_i} for Rhat A(ℓ) = Table 1 (12 values)
    Degree-11 least-squares fit to A(ℓ)=αs(e^ℓ GeV^2) on a chosen ℓ-interval, rounded to 6 significant figures. Used in the kappa_h model (103)-(106); fitted to a known function, not to the reported metrics.
  • Polynomial coefficients {b_i} for Rhat B(u) = Table 1 (12 values)
    Degree-11 least-squares fit to B(u)u^{-3/2}=eκ_fqq(e^{-u/4})u^{-3/2} on a chosen interval, rounded to 6 significant figures. Combined with Rhat A to build the redistribution kernel.
  • Sampling probabilities P(H), P(S) = P(H)=0.25, P(S)=0.75
    User-chosen generator inputs conforming to Ref [16] defaults. They affect the mixture of H and S events but not the final visible quasi density.
  • Bad redistribution kernel constant kappa_h_bad = 0.3 (for M=91.2 GeV)
    Ad hoc constant chosen to demonstrate that an intentionally poor kernel does not introduce bias. Used only for the ARCANE,BAD weighting scheme.
assumptions (5)
  • domain assumption MC@NLO formalism and Catani-Seymour dipole factorization correctly describe e+e- -> q qbar + 1 jet, with spin-averaged splitting kernels and running coupling as implemented in the base generator.
    Invoked throughout Sections 2 and 3; the physics content of the weights is taken from prior literature [13-19] and the tutorial generator [16].
  • domain assumption The base event generator from Ref [16] implements the described pipeline exactly, including the unweighted veto algorithm, proposal kernels g_c, and acceptance probabilities alpha_c=f_c/g_c.
    The ARCANE weight formulas (80), (91) depend on sampling densities P(V,H,cgen), P(V,S,rej-list) and the pathway probabilities (36) being exact; the paper extracts them from internals of the base event generator (Section 3.2).
  • ad hoc to paper The visible parametrization V=(cosθlo, φlo, fl, ℓ, c, z, φ) is complete and non-redundant for both H and S histories, and the Jacobian factors in Eq (46) correctly convert H-event densities to this parametrization.
    Required for the redistribution to preserve the marginal quasi density F(V); any Jacobian error would violate the zero-integral condition (2).
  • standard math The recursive-chain representation of the weighted veto algorithm in Eq (27) and the conditional rejection-pathway density in Eq (36) are exact for the implemented algorithm.
    These identities underpin the factorization of Lambda and Phi_un; they follow from the algorithm definition if the algorithm is implemented as stated.
  • domain assumption The ARCANE framework of Ref [12] is valid, in particular that any G satisfying Eq (2) with the stated coverage conditions preserves the visible distribution and yields finite weights.
    The present paper re-derives the preservation property for its specific G in Eqs (62)-(64), but the general framework and the definition of optimal redistribution are attributed to the unpublished companion paper [12].

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Demonstration of ARCANE Reweighting: Reducing the Sign Problem in the MC@NLO Generation of $e^+ e^- \rightarrow q \bar{q} + 1\, jet$ Events." pith.science (2026). https://pith.science/paper/NKVLWIQE

@misc{pith2026250208053,
  author       = {Pith},
  title        = {Pith review of: A Demonstration of ARCANE Reweighting: Reducing the Sign Problem in the MC@NLO Generation of $e^+ e^- \rightarrow q \barq + 1\, jet$ Events},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NKVLWIQE}},
  note         = {Machine review of arXiv:2502.08053}
}
abstract

Negatively weighted events, which appear in the simulation of particle collisions, significantly increase the computational requirements of collider experiments. A new technique called ARCANE reweighting has been introduced in a companion paper to tackle this problem. This paper demonstrates the technique for the next-to-leading-order generation of $e^+ e^- \rightarrow q \bar{q} + 1\, jet$ events. By redistributing the contributions of "standard" and "hard remainder" pathways in the generator that lead to the same final event, ARCANE reweighting almost completely eliminates the negative weights problem for this process. Some thoughts on implementing the technique in other scenarios are provided.

Figures

Figures reproduced from arXiv: 2502.08053 by the authors.

Figure 1
Figure 1. Flowchart depicting a general overview of the event-generation pipeline [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Flowchart depicting the generation of H-events, until the generation of a resolved event and the choosing of ℓstart for the first parton shower (PS) emission. • The flavor fl of the quark. • The specific sequence of Nrej ∈ {0, 1, . . . } rejected proposals rej-list ≡ [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗
Figure 3
Figure 3. Flowchart depicting the generation of S-events, until performing the first parton shower (PS) emission and choosing ℓstart for the next emission. 3 Walkthrough of the ARCANE Reweighting Implementation A non-technical overview of the implementation of ARCANE reweighting will be provided next, before walking through the technical details of the implementation. 3.1 Overview of the Redistribution Strategy The events pro… view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: A diagram depicting the different Monte Carlo (MC) histories that lead to [PITH_FULL_IMAGE:figures/full_fig_p018_4.png]
Figure 5
Figure 5. Figure 5: Plots depicting the z-independent splitting kernel κfqq , proposal kernel κgqq , and redistribution kernels κhqq and κhbad qq used in this paper. For convenience, the subscript qq has been dropped from the plot labels and legends. The top-left and top-right panels depi…
Figure 6
Figure 6. Figure 6: Heatmaps depicting the z-independent splitting kernel κfqq (left panel) and redistribution kernel κhqq (middle panel), as a function of ℓ and M, for 2 GeV ≤ M ≤ 110 GeV and 0 ≤ ℓ < ln(M2 /(4 GeV2 )). The right panel shows the signed relative error in κhqq with respect …
Figure 7
Figure 7. Figure 7: Two-dimensional unit-normalized histograms of different pairs of event [PITH_FULL_IMAGE:figures/full_fig_p033_7.png]
Figure 8
Figure 8. Figure 8: Each row depicts the one-dimensional weighted distributions of a different [PITH_FULL_IMAGE:figures/full_fig_p039_8.png]
Figure 9
Figure 9. Figure 9: Each row depicts the one-dimensional weighted distributions (for events [PITH_FULL_IMAGE:figures/full_fig_p040_9.png]
Figure 10
Figure 10. Figure 10: The left panel shows a scatter plot of W∗,lb and W∗,ub for events from classes 1 and 4 with W(ARCANE) < 0 in the medium dataset. There are 22 such events in total, marked by orange diamonds. Events from the same classes with W(ARCANE) ≥ 0, which lie within the zoomed-…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 19 canonical work pages

  1. [12]

    Shyamsundar,ARCANE Reweighting: A Monte Carlo Technique to Tackle the Negative Weights Problem in Collider Event Generation (2025), 2501.XXXXX

    P . Shyamsundar,ARCANE Reweighting: A Monte Carlo Technique to Tackle the Negative Weights Problem in Collider Event Generation (2025), 2501.XXXXX

    work page 2025

  2. [1]

    Buckley ,Computational challenges for MC event generation, J

    A. Buckley ,Computational challenges for MC event generation, J. Phys. Conf. Ser.1525(1), 012023 (2020), doi:10.1088 /1742-6596/1525/1/012023, 1908.00167

    arXiv 2020

  3. [2]

    Amoroso et al., Challenges in Monte Carlo Event Generator Software for High-Luminosity LHC, Comput

    S. Amoroso et al., Challenges in Monte Carlo Event Generator Software for High-Luminosity LHC, Comput. Softw. Big Sci.5(1), 12 (2021), doi:10.1007/s41781-021-00055-1, 2004. 13687

    doi:10.1007/s41781-021-00055-1 2021

  4. [3]

    Jadach, W

    S. Jadach, W . Płaczek, S. Sapeta, A. Siódmok and M. Skrzypek,Matching NLO QCD with parton shower in Monte Carlo scheme — the KrkNLO method , JHEP 10, 052 (2015), doi:10.1007/JHEP10(2015)052, 1503.06849

    arXiv 2015

  5. [4]

    Frederix, S

    R. Frederix, S. Frixione, S. Prestel and P . Torrielli, On the reduction of neg- ative weights in MC@NLO-type matching procedures , JHEP 07, 238 (2020), doi:10.1007/JHEP07(2020)238, 2002.12716

    arXiv 2020

  6. [5]

    Danziger, S

    K. Danziger, S. Höche and F . Siegert, Reducing negative weights in Monte Carlo event generation with Sherpa (2021), 2110.15211

    arXiv 2021

  7. [6]

    J. R. Andersen, C. Gütschow, A. Maier and S. Prestel, A Positive Resampler for Monte Carlo events with negative weights , Eur. Phys. J. C 80(11), 1007 (2020), doi:10.1140/epjc/s10052-020-08548-w, 2005.09375

    arXiv 2020

  8. [7]

    Nachman and J

    B. Nachman and J. Thaler, Neural resampler for Monte Carlo reweighting with preserved uncertainties, Phys. Rev. D102(7), 076004 (2020), doi:10.1103/PhysRevD.102.076004, 2007.11586

    arXiv 2020

Show all 62 references
  1. [8]

    J. R. Andersen and A. Maier, Unbiased elimination of negative weights in Monte Carlo samples, Eur. Phys. J. C 82(5), 433 (2022), doi:10.1140 /epjc/s10052-022-10372-3, 2109.07851

    2022 arXiv

  2. [9]

    J. R. Andersen, A. Maier and D. Maître, Efficient negative-weight elimination in large high-multiplicity Monte Carlo event samples , Eur. Phys. J. C 83(9), 835 (2023), doi:10.1140/epjc/s10052-023-11905-0, 2303.15246

    2023 arXiv

  3. [10]

    Frederix and P

    R. Frederix and P . Torrielli, A new way of reducing negative weights in MC@NLO , Eur. Phys. J. C 83(11), 1051 (2023), doi:10.1140 /epjc/s10052-023-12243-x, 2310.04160. 49 SciPost Physics Submission

    2023 arXiv

  4. [11]

    J. R. Andersen, A. Cueto, S. P . Jones and A. Maier, A Cell Resampler study of Negative Weights in Multi-jet Merged Samples (2024), 2411.11651

    2024 arXiv

  5. [13]

    Frixione and B

    S. Frixione and B. R. Webber,Matching NLO QCD computations and parton shower simula- tions, JHEP 06, 029 (2002), doi:10.1088/1126-6708/2002/06/029, hep-ph /0204244

    2002 doi

  6. [14]

    Frixione, P

    S. Frixione, P . Nason and B. R. Webber, Matching NLO QCD and parton showers in heavy flavor production, JHEP 08, 007 (2003), doi:10.1088 /1126-6708/2003/08/007, hep-ph/0305252

    2003 arXiv

  7. [15]

    Frixione and B

    S. Frixione and B. R. Webber, The MC@NLO 2.0 event generator (2003), hep-ph / 0307146

    2003

  8. [16]

    S. Höche, Introduction to parton showers and matching, GitLab repository: gitlab.com / shoeche/tutorials, Commit ID: 5b7fcd43861a95e1fe85958a60b4ac55d8a7e236, Loca- tion in repository: “./ps/” (2018)

    2018

  9. [17]

    S. Höche, Introduction to parton-shower event generators , In Theoretical Advanced Study Institute in Elementary Particle Physics: Journeys Through the Precision Frontier: Amplitudes for Colliders , pp. 235–295, doi:10.1142 /9789814678766_0005 (2015), 1411.4085

    2015 arXiv

  10. [18]

    B. R. Webber, Monte Carlo Simulation of Hard Hadronic Processes, Ann. Rev. Nucl. Part. Sci. 36, 253 (1986), doi:10.1146 /annurev.ns.36.120186.001345

    1986

  11. [19]

    Catani and M

    S. Catani and M. H. Seymour,A General algorithm for calculating jet cross-sections in NLO QCD, Nucl. Phys. B 485, 291 (1997), doi:10.1016/S0550-3213(96)00589-5, [Erratum: Nucl.Phys.B 510, 503–504 (1998)], hep-ph /9605323

    1997 doi

  12. [20]

    Hoeche, F

    S. Hoeche, F . Krauss, M. Schonherr and F . Siegert,A critical appraisal of NLO+PS matching methods, JHEP 09, 049 (2012), doi:10.1007 /JHEP09(2012)049, 1111.1220

    2012 arXiv

  13. [21]

    Platzer and M

    S. Platzer and M. Sjodahl, The Sudakov Veto Algorithm Reloaded, Eur. Phys. J. Plus 127, 26 (2012), doi:10.1140 /epjp/i2012-12026-x, 1108.6180

    2012 arXiv

  14. [22]

    Lönnblad, Fooling Around with the Sudakov Veto Algorithm, Eur

    L. Lönnblad, Fooling Around with the Sudakov Veto Algorithm, Eur. Phys. J. C73(3), 2350 (2013), doi:10.1140 /epjc/s10052-013-2350-9, 1211.7204

    2013 arXiv

  15. [23]

    Kleiss and R

    R. Kleiss and R. Verheyen, Competing Sudakov Veto Algorithms, Eur. Phys. J. C 76(7), 359 (2016), doi:10.1140 /epjc/s10052-016-4231-5, 1605.09246

    2016 arXiv

  16. [24]

    B. R. Webber, Quantum Correlations in QCD Jets , Phys. Lett. B 193, 91 (1987), doi:10.1016/0370-2693(87)90461-8

    1987 doi

  17. [25]

    Bellm, S

    J. Bellm, S. Plätzer, P . Richardson, A. Siódmok and S. Webster, Reweighting Parton Showers, Phys. Rev. D 94(3), 034028 (2016), doi:10.1103 /PhysRevD.94.034028, 1605.08256

    2016 arXiv

  18. [26]

    Mrenna and P

    S. Mrenna and P . Skands,Automated Parton-Shower Variations in Pythia 8, Phys. Rev. D 94(7), 074005 (2016), doi:10.1103 /PhysRevD.94.074005, 1605.08352

    2016 arXiv

  19. [27]

    Bothmann, M

    E. Bothmann, M. Schönherr and S. Schumann, Reweighting QCD matrix- element and parton-shower calculations , Eur. Phys. J. C 76(11), 590 (2016), doi:10.1140/epjc/s10052-016-4430-0, 1606.08753. 50 SciPost Physics Submission

    2016 arXiv

  20. [28]

    Bierlich, P

    C. Bierlich, P . Ilten, T . Menzo, S. Mrenna, M. Szewc, M. K. Wilkinson, A. Youssef and J. Zupan, Reweighting Monte Carlo predictions and automated fragmentation variations in Pythia 8, SciPost Phys. 16(5), 134 (2024), doi:10.21468 /SciPostPhys.16.5.134, 2308. 13459

    2024

  21. [29]

    Kleiss and R

    R. Kleiss and R. Pittau, Weight optimization in multichannel Monte Carlo, Comput. Phys. Commun. 83, 141 (1994), doi:10.1016 /0010-4655(94)90043-4, hep-ph /9405257

    1994

  22. [30]

    Weinzierl, Introduction to Monte Carlo methods (2000), hep-ph /0006269

    S. Weinzierl, Introduction to Monte Carlo methods (2000), hep-ph /0006269

    2000

  23. [31]

    The Sage Developers, SageMath, the Sage Mathematics Software System (Version 10.4) , https://www.sagemath.org (2024)

    2024

  24. [32]

    version 5.47.0 (2023)

    Maxima, Maxima, a computer algebra system. version 5.47.0 (2023)

    2023

  25. [33]

    Meurer, C

    A. Meurer, C. P . Smith, M. Paprocki, O. ˇCertík, S. B. Kirpichev, M. Rocklin, A. Kumar, S. Ivanov, J. K. Moore, S. Singh, T . Rathnayake, S. Viget al., Sympy: symbolic computing in python, PeerJ Computer Science 3, e103 (2017), doi:10.7717 /peerj-cs.103

    2017

  26. [34]

    Brandt, C

    S. Brandt, C. Peyrou, R. Sosnowski and A. Wroblewski, The Principal axis of jets. An Attempt to analyze high-energy collisions as two-body processes, Phys. Lett. 12, 57 (1964), doi:10.1016/0031-9163(64)91176-X

    1964 doi

  27. [35]

    Georgi and M

    H. Georgi and M. Machacek,A Simple QCD Prediction of Jet Structure in e+ e- Annihilation, Phys. Rev. Lett.39, 1237 (1977), doi:10.1103 /PhysRevLett.39.1237

    1977

  28. [36]

    Farhi, A QCD Test for Jets , Phys

    E. Farhi, A QCD Test for Jets , Phys. Rev. Lett. 39, 1587 (1977), doi:10.1103/PhysRevLett.39.1587

    1977 doi

  29. [37]

    Yamamoto, AN EFFICIENT ALGORITHM FOR CALCULATING THRUST IN HIGH MULTIPLICITY REACTIONS, J

    H. Yamamoto, AN EFFICIENT ALGORITHM FOR CALCULATING THRUST IN HIGH MULTIPLICITY REACTIONS, J. Comput. Phys. 52, 597 (1983), doi:10.1016 /0021- 9991(83)90010-4

    1983

  30. [38]

    Butter, T

    A. Butter, T . Plehn and R. Winterhalder, How to GAN Event Subtraction , SciPost Phys. Core 3, 009 (2020), doi:10.21468 /SciPostPhysCore.3.2.009, 1912.08824

    2020 arXiv

  31. [39]

    Backes, A

    M. Backes, A. Butter, T . Plehn and R. Winterhalder, How to GAN Event Unweighting , SciPost Phys. 10(4), 089 (2021), doi:10.21468 /SciPostPhys.10.4.089, 2012.07873

    2021 arXiv

  32. [40]

    Bierlich et al., A comprehensive guide to the physics and usage of PYTHIA 8.3 , SciPost Phys

    C. Bierlich et al., A comprehensive guide to the physics and usage of PYTHIA 8.3 , SciPost Phys. Codeb. 2022, 8 (2022), doi:10.21468 /SciPostPhysCodeb.8, 2203.11601

    2022 arXiv

  33. [41]

    Nason,MINT: A Computer program for adaptive Monte Carlo integration and generation of unweighted distributions (2007), 0709.2085

    P . Nason,MINT: A Computer program for adaptive Monte Carlo integration and generation of unweighted distributions (2007), 0709.2085

    2007 arXiv

  34. [42]

    G. P . Lepage,Adaptive multidimensional integration: VEGAS enhanced , J. Comput. Phys. 439, 110386 (2021), doi:10.1016 /j.jcp.2021.110386, 2009.05112

    2021

  35. [43]

    Bothmann, T

    E. Bothmann, T . Janßen, M. Knobbe, T . Schmale and S. Schumann, Exploring phase space with Neural Importance Sampling , SciPost Phys. 8(4), 069 (2020), doi:10.21468/SciPostPhys.8.4.069, 2001.05478

    2020 arXiv

  36. [44]

    C. Gao, J. Isaacson and C. Krause,i-flow: High-dimensional Integration and Sampling with Normalizing Flows, Mach. Learn. Sci. Tech. 1(4), 045023 (2020), doi:10.1088 /2632- 2153/abab62, 2001.05486. 51 SciPost Physics Submission

    2020 arXiv

  37. [45]

    C. Gao, S. Höche, J. Isaacson, C. Krause and H. Schulz,Event Generation with Normalizing Flows, Phys. Rev. D101(7), 076002 (2020), doi:10.1103/PhysRevD.101.076002, 2001. 10028

    2020 doi

  38. [46]

    Heimel, R

    T . Heimel, R. Winterhalder, A. Butter, J. Isaacson, C. Krause, F . Maltoni, O. Mattelaer and T . Plehn,MadNIS - Neural multi-channel importance sampling, SciPost Phys. 15(4), 141 (2023), doi:10.21468 /SciPostPhys.15.4.141, 2212.06172

    2023 arXiv

  39. [47]

    Heimel, N

    T . Heimel, N. Huetsch, F . Maltoni, O. Mattelaer, T . Plehn and R. Winterhalder,The MadNIS reloaded, SciPost Phys. 17(1), 023 (2024), doi:10.21468 /SciPostPhys.17.1.023, 2311. 01548

    2024

  40. [48]

    Heimel, O

    T . Heimel, O. Mattelaer, T . Plehn and R. Winterhalder,Differentiable MadNIS-Lite, SciPost Phys. 18, 017 (2025), doi:10.21468 /SciPostPhys.18.1.017, 2408.01486

    2025 arXiv

  41. [49]

    Höche and S

    S. Höche and S. Prestel, The midpoint between dipole and parton showers, Eur. Phys. J. C 75(9), 461 (2015), doi:10.1140 /epjc/s10052-015-3684-2, 1506.05057

    2015 arXiv

  42. [50]

    Hoeche, F

    S. Hoeche, F . Krauss, S. Schumann and F . Siegert, QCD matrix elements and truncated showers, JHEP 05, 053 (2009), doi:10.1088 /1126-6708/2009/05/053, 0903.1219

    2009 arXiv

  43. [51]

    Bothmann et al

    E. Bothmann et al. , Event generation with Sherpa 3 , JHEP 12, 156 (2024), doi:10.1007/JHEP12(2024)156, 2410.22148

    2024 arXiv

  44. [52]

    Alwall, R

    J. Alwall, R. Frederix, S. Frixione, V . Hirschi, F . Maltoni, O. Mattelaer, H. S. Shao, T . Stelzer, P . Torrielli and M. Zaro,The automated computation of tree-level and next-to- leading order differential cross sections, and their matching to parton shower simulations , JHE...

    2014 arXiv

  45. [53]

    Bewick et al

    G. Bewick et al. , Herwig 7.3 release note , Eur. Phys. J. C 84(10), 1053 (2024), doi:10.1140/epjc/s10052-024-13211-9, 2312.05175

    2024 arXiv

  46. [54]

    Alioli, P

    S. Alioli, P . Nason, C. Oleari and E. Re, A general framework for implementing NLO cal- culations in shower Monte Carlo programs: the POWHEG BOX , JHEP 06, 043 (2010), doi:10.1007/JHEP06(2010)043, 1002.2581

    2010 arXiv

  47. [55]

    Bothmann, T

    E. Bothmann, T . Childers, W . Giele, F . Herren, S. Hoeche, J. Isaacson, M. Knobbe and R. Wang, Efficient phase-space generation for hadron collider event simulation , SciPost Phys. 15(4), 169 (2023), doi:10.21468 /SciPostPhys.15.4.169, 2302.10449

    2023 arXiv

  48. [56]

    Bothmann, T

    E. Bothmann, T . Childers, W . Giele, S. Höche, J. Isaacson and M. Knobbe, A portable parton-level event generator for the high-luminosity LHC, SciPost Phys.17(3), 081 (2024), doi:10.21468/SciPostPhys.17.3.081, 2311.06198

    2024 arXiv

  49. [57]

    Van Rossum and F

    G. Van Rossum and F . L. Drake Jr,Python reference manual, Centrum voor Wiskunde en Informatica Amsterdam (1995)

    1995

  50. [58]

    C. R. Harris, K. J. Millman, S. J. van der Walt, R. Gommers, P . Virtanen, D. Cournapeau, E. Wieser, J. Taylor, S. Berg, N. J. Smith, R. Kern, M. Picuset al., Array programming with NumPy, Nature 585(7825), 357 (2020), doi:10.1038 /s41586-020-2649-2

    2020

  51. [59]

    J. D. Hunter,Matplotlib: A 2d graphics environment, Computing in Science & Engineering 9(3), 90 (2007), doi:10.1109 /MCSE.2007.55

    2007

  52. [60]

    52 SciPost Physics Submission

    Jupyterlab contributors, Jupyterlab version 4.3.5 (2024). 52 SciPost Physics Submission

    2024

  53. [61]

    tqdm contributors, tqdm version 4.67.1 (2024)

    2024

  54. [62]

    Schemdraw contributors, Schemdraw version 0.19 (2024). 53

    2024

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.