arxiv: 2605.01005 · v1 · submitted 2026-05-01 · 🌌 astro-ph.EP · astro-ph.IM

Recognition: unknown

A Coordinate System for Dynamical Instabilities in Hierarchical Systems in REBOUND

Garett Brown, Tiger Lu

Pith reviewed 2026-05-09 18:03 UTC · model grok-4.3

classification 🌌 astro-ph.EP astro-ph.IM

keywords N-body integrationwide binary systemshierarchical systemsdynamical instabilitieshybrid integratorsclose encountersREBOUNDTRACE

0 comments

The pith

A new coordinate system in the TRACE integrator lets REBOUND accurately simulate instabilities in wide binary hierarchical systems where other hybrid methods fail.

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

The paper implements a coordinate transformation suited to wide binary systems inside the TRACE hybrid integrator of the REBOUND N-body package. This targets the regime in which standard hybrid integrators produce poor results during close encounters in hierarchical configurations such as planet-planet scattering or Kozai-Lidov cycles. The authors describe the implementation, which supports encounters between any pair of bodies, and then benchmark it against other REBOUND integrators on several test problems. In those tests the new coordinates yield qualitatively correct behavior when alternatives fail and produce results statistically similar to the high-precision IAS15 integrator while delivering speedups of up to 9 times. Guidelines are supplied for choosing when the coordinates are appropriate.

Core claim

Transforming to wide binary coordinates inside TRACE enables the hybrid integrator to handle close encounters between any bodies in hierarchical systems, delivering qualitatively correct results in regimes where other hybrid methods fail and statistically similar outcomes to IAS15 with up to 9 times the speed.

What carries the argument

Wide binary coordinate system that supports close encounters between any pair of bodies and is embedded in the TRACE hybrid integrator.

If this is right

Planet-planet scattering in wide binaries can be modeled reliably without switching to a slower integrator.
Stellar flybys and Kozai-Lidov oscillations become accessible to hybrid integration with maintained accuracy.
Statistical outcomes comparable to IAS15 become available at up to 9 times lower computational cost.
Users receive concrete guidelines for selecting these coordinates in new hierarchical integrations.

Where Pith is reading between the lines

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

The same transformation idea could be ported to other hybrid N-body codes that currently struggle with wide binaries.
Large-scale surveys of exoplanet architectures around binary stars become more feasible because of the speed gain.
The approach may reduce reliance on full-precision integrators for initial exploration of long-term hierarchical dynamics.

Load-bearing premise

The new coordinate transformation and its close-encounter handling must stay numerically stable and free of systematic biases across all mass ratios, separations, and encounter geometries that appear in real hierarchical systems.

What would settle it

A side-by-side run of the planet-planet scattering or ZLK test cases at extreme mass ratios or very small separations that produces large divergences from IAS15 results or visible numerical instability would falsify the performance claim.

Figures

Figures reproduced from arXiv: 2605.01005 by Garett Brown, Tiger Lu.

**Figure 1.** Figure 1: Schematic representation of a wide binary system, where a system of planets orbits around one component of a binary star system. Orbits and physical sizes not to scale. while occasionally switching over to more flexible integration techniques when close encounters call for it. Wide binary systems, in which a planetary system orbits one component of a stellar binary (see view at source ↗

**Figure 2.** Figure 2: Comparison of energy errors in a parameter space of binary companions, in semimajor axis-mass space. For each companion, the ratio of energy errors between an integration with WB coordinates and an integration with DHC coordinates is plotted, both with the TRACE integrator. Blue corresponds to WB performing better, and red the opposite. The gray contour roughly corresponds to the limit where WB outperform… view at source ↗

**Figure 3.** Figure 3: Distributions of semimajor axis, eccentricity, and inclination from results of three-body scattering experiments including a binary perturber. Results from IAS15 are plotted as a solid green histogram, and taken to be the gold standard for numerical integration. Results from TRACE DHC and TRACE WB are plotted as blue and purple step histograms, respectively. Rasio (1996); D. Malmberg et al. (2011) for the … view at source ↗

**Figure 4.** Figure 4: Effect of a stellar flyby as a function of perturber mass on a single planet system’s energy error (top), eccentricity excitation (middle) and inclination excitation (bottom). We compare results from IAS15 (green), TRACE WB (violet) and TRACE DHC (blue) to analytic predictions (black) view at source ↗

**Figure 5.** Figure 5: Binary inclination - timestep space in which TRACE DHC and WB are preferred over IAS15 for ZLK integrations. Gray squares indicate where TRACE does not reproduce IAS15 results. Hatched colored squares indicate where TRACE reproduces IAS15 results, but is slower. Unhatched colored squares represent parameter space where TRACE is preferred over IAS15. WB coordinates greatly expand the region of parameter spa… view at source ↗

**Figure 6.** Figure 6: Evolution of Jupiter’s eccentricity for a system with a Sun-like star, and Earth-like planet on a 0.1 AU orbit, a Jupiter-like planet on a 5 AU orbit, and a 0.5 M⊙ companion on a 100 AU orbit inclined by 70◦ . We compare results from IAS15 (green), TRACE WB (purple) and TRACE DHC (blue). For clarity, only a representative slice of the evolution from 0.5 to 1.5 Myr is plotted. IAS15 takes ∼20× longer to … view at source ↗

read the original abstract

We implement coordinates suitable for studying wide binary systems in TRACE, a hybrid integrator in the widely used open-source N-body integration package REBOUND. This is a regime in which traditional hybrid integrators perform poorly. The coordinate system supports close encounters between any pair of bodies in the system. We describe the implementation of this coordinate system and benchmark its performance against other integrators in the REBOUND ecosystem. In tests of planet-planet scattering, stellar flybys, and ZLK oscillations. TRACE in wide binary coordinates is qualitatively correct when other hybrid methods fail, and in many cases returns statistically similar results to the high-precision IAS15 integrator with up to 9x speedups. We also provide some guidelines for when use of these coordinates are appropriate.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

REBOUND's TRACE now has wide-binary coordinates that handle some hierarchical close-encounter cases faster than prior hybrid options.

read the letter

This paper adds a coordinate system for wide binary hierarchies to the TRACE integrator in REBOUND. The key point is that it lets users run simulations of distant companions with inner planets or close encounters more efficiently than before, in cases where other hybrid methods break down. They describe the implementation and test it against existing REBOUND integrators on planet-planet scattering, stellar flybys, and ZLK oscillations. The results show that TRACE with these coordinates stays qualitatively correct where others fail and often matches the high-accuracy IAS15 runs statistically, with speedups reaching 9 times. The guidelines on appropriate use cases are a nice practical touch. The main limitation is the scope of the benchmarks. They cover three instability types but do not appear to systematically vary mass ratios down to very small values or test extreme encounter geometries. Without those, it's hard to know if systematic biases creep in outside the tested slice. The paper relies on direct numerical comparisons rather than analytic properties of the coordinates, which is fine for an implementation but leaves the robustness claim somewhat provisional. This work is for REBOUND users who model hierarchical planetary systems, especially wide binaries. Anyone running large numbers of such simulations will find the new option and the reported timings useful. I think it should go to peer review. The addition is real, the tests are relevant, and the community benefits from having this documented even if further validation on edge cases would help.

Referee Report

1 major / 2 minor

Summary. The paper implements a coordinate system for wide binary systems in the TRACE hybrid integrator within the REBOUND N-body package. This targets regimes where traditional hybrid integrators struggle with close encounters in hierarchical configurations. The authors describe the implementation, which supports encounters between any pair, and benchmark it against other REBOUND integrators in three classes of tests: planet-planet scattering, stellar flybys, and ZLK oscillations. They report that TRACE in wide-binary coordinates yields qualitatively correct results where other hybrids fail, produces statistically similar outcomes to the high-precision IAS15 integrator in many cases, achieves speedups up to 9x, and provide usage guidelines.

Significance. If the benchmarks are robust, the work supplies a practical, open-source tool for efficient simulation of dynamical instabilities in astrophysically common hierarchical systems such as wide binaries and multi-planet architectures. Integration into the widely used REBOUND ecosystem and the inclusion of explicit usage guidelines are strengths that could facilitate broader adoption for parameter surveys. The contribution is primarily methodological and engineering-focused rather than introducing new physical insights or parameter-free derivations.

major comments (1)

[Benchmarking section] Benchmarking section (planet-planet scattering, stellar flybys, and ZLK tests): The explored parameter space does not systematically include extreme mass ratios (10^{-8} to 1), separations spanning 10–10^6 AU, or grazing high-velocity encounters. The central claim that results remain 'qualitatively correct' and 'statistically similar' to IAS15 without systematic biases therefore rests on an unverified extrapolation; explicit conservation checks, divergence metrics, or additional runs in these regimes are needed to substantiate generality across relevant hierarchical configurations.

minor comments (2)

[Abstract and methods] The abstract and methods description omit details on error-bar computation, exact data-exclusion rules, and statistical tests underlying the 'statistically similar' comparisons to IAS15; these should be added for reproducibility.
[Results] Runtime and accuracy comparison tables or figures would benefit from explicit listing of integrator settings (timestep, tolerance) and number of realizations per test class to allow direct assessment of the reported speedups.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their thoughtful review of our manuscript. We address the major comment on the benchmarking section below. We have made partial revisions to clarify the scope of our tests and add conservation checks.

read point-by-point responses

Referee: [Benchmarking section] Benchmarking section (planet-planet scattering, stellar flybys, and ZLK tests): The explored parameter space does not systematically include extreme mass ratios (10^{-8} to 1), separations spanning 10–10^6 AU, or grazing high-velocity encounters. The central claim that results remain 'qualitatively correct' and 'statistically similar' to IAS15 without systematic biases therefore rests on an unverified extrapolation; explicit conservation checks, divergence metrics, or additional runs in these regimes are needed to substantiate generality across relevant hierarchical configurations.

Authors: We appreciate the referee pointing out the limited parameter space in our benchmarks. The tests presented were chosen to demonstrate the method's performance in regimes where other hybrid integrators fail, specifically for typical planet-planet scattering events, stellar flybys in wide binaries, and ZLK oscillations in hierarchical triples. These cover a range of mass ratios and separations relevant to astrophysical applications, though not the most extreme ends. We agree that the claims of qualitative correctness and statistical similarity are based on the tested cases, and we do not claim universality without further verification. In the revised version, we will expand the discussion to explicitly state the explored parameter ranges, add energy conservation plots for all test cases, and include a note on potential limitations for extreme grazing encounters or very low mass ratios. We have performed additional runs for a few more extreme cases (e.g., mass ratio 10^{-6} and larger separations) and will include them in the revised manuscript to better support the generality. However, a full systematic exploration of the entire parameter space (10^{-8} to 1, 10 to 10^6 AU, all encounter types) is beyond the scope of this paper but could be the subject of future work. revision: partial

Circularity Check

0 steps flagged

No significant circularity: implementation and benchmarking paper with external validation

full rationale

The paper describes the implementation of wide-binary coordinates in the TRACE hybrid integrator and benchmarks its performance against independent high-precision integrators such as IAS15. No derivation chain, first-principles predictions, or fitted parameters are presented that reduce to the inputs by construction. Central performance claims rest on direct numerical comparisons to external reference integrators rather than quantities defined from the new coordinates themselves. No self-citations, ansatzes, or uniqueness theorems are invoked as load-bearing elements. The work is self-contained against external benchmarks, consistent with a score of 0.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work is a numerical implementation that inherits standard Newtonian gravity, symplectic integration properties, and hybrid switching logic from the existing REBOUND and TRACE codebase. No new physical axioms or free parameters are introduced beyond the coordinate choice itself.

axioms (2)

standard math Newtonian point-mass gravity governs the motion of all bodies
Implicit in all N-body integrators in REBOUND; stated in the context of the hierarchical system tests.
domain assumption Hybrid integrators can switch between fast and slow methods without introducing unacceptable errors when close encounters are properly detected
Core assumption of the TRACE integrator that the new coordinates extend.

pith-pipeline@v0.9.0 · 5418 in / 1457 out tokens · 27146 ms · 2026-05-09T18:03:16.535279+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

50 extracted references · 44 canonical work pages · 1 internal anchor

[1]

Antognini, J. M. O. 2015, MNRAS, 452, 3610, doi: 10.1093/mnras/stv1552

work page doi:10.1093/mnras/stv1552 2015
[2]

2022, MNRAS, 515, 5942, doi: 10.1093/mnras/stac1763

Brown, G., & Rein, H. 2022, MNRAS, 515, 5942, doi: 10.1093/mnras/stac1763

work page doi:10.1093/mnras/stac1763 2022
[3]

2024, https://airball.gbrown.ca/

Brown, G., Rein, H., Mohsin, H., et al. 2024, https://airball.gbrown.ca/

2024
[4]

, keywords =

Cai, M. X., Kouwenhoven, M. B. N., Portegies Zwart, S. F., & Spurzem, R. 2017, MNRAS, 470, 4337, doi: 10.1093/mnras/stx1464

work page doi:10.1093/mnras/stx1464 2017
[5]

Lissauer, J. J. 2002, AJ, 123, 2884, doi: 10.1086/340074

work page doi:10.1086/340074 2002
[6]

E., & Wetherill, G

Chambers, J. E., & Wetherill, G. W. 1998, Icarus, 136, 304, doi: 10.1006/icar.1998.6007

work page doi:10.1006/icar.1998.6007 1998
[7]

, keywords =

Chatterjee, S., Ford, E. B., Matsumura, S., & Rasio, F. A. 2008, ApJ, 686, 580, doi: 10.1086/590227

work page doi:10.1086/590227 2008
[8]

2025, AJ, 169, 308, doi: 10.3847/1538-3881/adc933

Christian, S., Vanderburg, A., Becker, J., et al. 2025, AJ, 169, 308, doi: 10.3847/1538-3881/adc933

work page doi:10.3847/1538-3881/adc933 2025
[9]

2014, in Protostars and Planets VI, ed

Davies, M. B., Adams, F. C., Armitage, P., et al. 2014, in Protostars and Planets VI, ed. H. Beuther, R. S. Klessen, C. P. Dullemond, & T. Henning, 787–808, doi: 10.2458/azu uapress 9780816531240-ch034

work page doi:10.2458/azu 2014
[10]

AJ 116(4), 2067–2077 (1998)

Duncan, M. J., Levison, H. F., & Lee, M. H. 1998, AJ, 116, 2067, doi: 10.1086/300541

work page doi:10.1086/300541 1998
[11]

ApJ 796, 23 (2014)

Grimm, S. L., & Stadel, J. G. 2014, ApJ, 796, 23, doi: 10.1088/0004-637X/796/1/23

work page doi:10.1088/0004-637x/796/1/23 2014
[12]

Heggie, D. C. 1975, Monthly Notices of the Royal Astronomical Society, 173, 729, doi: 10.1093/mnras/173.3.729

work page doi:10.1093/mnras/173.3.729 1975
[13]

Heggie, D. C. 2006, in Few-Body Problem: Theory and Computer Simulations, 20, doi: 10.48550/arXiv.astro-ph/0512504

work page doi:10.48550/arxiv.astro-ph/0512504 2006
[14]

C., & Rasio, F

Heggie, D. C., & Rasio, F. A. 1996, Monthly Notices of the Royal Astronomical Society, 282, 1064, doi: 10.1093/mnras/282.3.1064

work page doi:10.1093/mnras/282.3.1064 1996
[15]

Hernandez, D. M. 2026, arXiv e-prints, arXiv:2603.24456, doi: 10.48550/arXiv.2603.24456

work page internal anchor Pith review doi:10.48550/arxiv.2603.24456 2026
[16]

M., & Dehnen, W

Hernandez, D. M., & Dehnen, W. 2017, MNRAS, 468, 2614, doi: 10.1093/mnras/stx547

work page doi:10.1093/mnras/stx547 2017
[17]

M., & Dehnen, W

Hernandez, D. M., & Dehnen, W. 2023, MNRAS, 522, 4639, doi: 10.1093/mnras/stad657

work page doi:10.1093/mnras/stad657 2023
[18]

M., Zeebe, R

Hernandez, D. M., Zeebe, R. E., & Hadden, S. 2022, MNRAS, 510, 4302, doi: 10.1093/mnras/stab3664

work page doi:10.1093/mnras/stab3664 2022
[19]

Hunter, J. D. 2007, Computing in Science & Engineering, 9, 90, doi: 10.1109/MCSE.2007.55 Juri´ c, M., & Tremaine, S. 2008, ApJ, 686, 603, doi: 10.1086/590047

work page doi:10.1109/mcse.2007.55 2007
[20]

A., Raymond, S

Kaib, N. A., Raymond, S. N., & Duncan, M. 2013, Nature, 493, 381, doi: 10.1038/nature11780

work page doi:10.1038/nature11780 2013
[21]

1991, Celestial Mechanics and Dynamical Astronomy (ISSN 0923-2958), vol

Kinoshita, H., Yoshida, H., & Nakai, H. 1991, Celestial Mechanics and Dynamical Astronomy (ISSN 0923-2958), vol. 50, no. 1, 1991, p. 59-71., 50, 59

1991
[22]

F., & Duncan, M

Levison, H. F., & Duncan, M. J. 2000, AJ, 120, 2117, doi: 10.1086/301553

work page doi:10.1086/301553 2000
[23]

Li, G., & Adams, F. C. 2015, MNRAS, 448, 344, doi: 10.1093/mnras/stv012

work page doi:10.1093/mnras/stv012 2015
[24]

2014a, ApJ, 791, 86, doi: 10.1088/0004-637X/791/2/86

Li, G., Naoz, S., Holman, M., & Loeb, A. 2014a, ApJ, 791, 86, doi: 10.1088/0004-637X/791/2/86

work page doi:10.1088/0004-637x/791/2/86
[25]

2014b, ApJ, 785, 116, doi: 10.1088/0004-637X/785/2/116

Li, G., Naoz, S., Kocsis, B., & Loeb, A. 2014b, ApJ, 785, 116, doi: 10.1088/0004-637X/785/2/116

work page doi:10.1088/0004-637x/785/2/116
[26]

Lidov, M. L. 1962, Planetary and Space Science, 9, 719

1962
[27]

M., & Rein, H

Lu, T., Hernandez, D. M., & Rein, H. 2024, MNRAS, 533, 3708, doi: 10.1093/mnras/stae1982

work page doi:10.1093/mnras/stae1982 2024
[28]

2023, The Astrophysical Journal, 948, 41, doi: 10.3847/1538-4357/acc06d

Lu, T., Rein, H., Tamayo, D., et al. 2023, The Astrophysical Journal, 948, 41, doi: 10.3847/1538-4357/acc06d

work page doi:10.3847/1538-4357/acc06d 2023
[29]

2010, MNRAS, 408, 2115, doi: 10.1111/j.1365-2966.2010.17291.x

Malmberg, D., Davies, M. B., & Heggie, D. C. 2011, MNRAS, 411, 859, doi: 10.1111/j.1365-2966.2010.17730.x

work page doi:10.1111/j.1365-2966.2010.17730.x 2011
[30]

M., Lithwick, Y ., Rasio, F

Teyssandier, J. 2013, MNRAS, 431, 2155, doi: 10.1093/mnras/stt302

work page doi:10.1093/mnras/stt302 2013
[31]

A., Hinkley, S., et al

Ngo, H., Knutson, H. A., Hinkley, S., et al. 2016, ApJ, 827, 8, doi: 10.3847/0004-637X/827/1/8

work page doi:10.3847/0004-637x/827/1/8 2016
[32]

, keywords =

Parker, R. J., & Quanz, S. P. 2012, MNRAS, 419, 2448, doi: 10.1111/j.1365-2966.2011.19911.x 10

work page doi:10.1111/j.1365-2966.2011.19911.x 2012
[33]

Pham, D., Rein, H., & Spiegel, D. S. 2024, The Open Journal of Astrophysics, 7, 1, doi: 10.21105/astro.2401.02849 Poincar´ e, H. 1890, Acta Math, 13, 270

work page doi:10.21105/astro.2401.02849 2024
[34]

Flannery, B. P. 2002, Numerical recipes in C++ : the art of scientific computing

2002
[35]

2026, arXiv e-prints, arXiv:2601.08019, doi: 10.48550/arXiv.2601.08019

Rein, H., Dey, K., & Tamayo, D. 2026, arXiv e-prints, arXiv:2601.08019, doi: 10.48550/arXiv.2601.08019

work page doi:10.48550/arxiv.2601.08019 2026
[36]

Rein, H., & Liu, S. F. 2012, A&A, 537, A128, doi: 10.1051/0004-6361/201118085

work page doi:10.1051/0004-6361/201118085 2012
[37]

Rein, H., & Spiegel, D. S. 2015, MNRAS, 446, 1424, doi: 10.1093/mnras/stu2164

work page doi:10.1093/mnras/stu2164 2015
[38]

1988,A&A, 202,

Rein, H., & Tamayo, D. 2015, MNRAS, 452, 376, doi: 10.1093/mnras/stv1257

work page doi:10.1093/mnras/stv1257 2015
[39]

2019, Research Notes of the AAS, 3, 16, doi: 10.3847/2515-5172/aaff63

Rein, H., & Tamayo, D. 2019, Research Notes of the AAS, 3, 16, doi: 10.3847/2515-5172/aaff63

work page doi:10.3847/2515-5172/aaff63 2019
[40]

2019a, MNRAS, 489, 4632, doi: 10.1093/mnras/stz2503

Rein, H., Tamayo, D., & Brown, G. 2019a, MNRAS, 489, 4632, doi: 10.1093/mnras/stz2503

work page doi:10.1093/mnras/stz2503
[41]

M., Tamayo, D., et al

Rein, H., Hernandez, D. M., Tamayo, D., et al. 2019b, MNRAS, 485, 5490, doi: 10.1093/mnras/stz769

work page doi:10.1093/mnras/stz769
[42]

1976, Bulletin of the Astronomical Institutes of Czechoslovakia, 27, 92

Rickman, H. 1976, Bulletin of the Astronomical Institutes of Czechoslovakia, 27, 92

1976
[43]

2003, Celestial Mechanics and Dynamical Astronomy, 87, 411, doi: 10.1023/B:CELE.0000006767.34371.2f

Roy, A., & Haddow, M. 2003, Celestial Mechanics and Dynamical Astronomy, 87, 411, doi: 10.1023/B:CELE.0000006767.34371.2f

work page doi:10.1023/b:cele.0000006767.34371.2f 2003
[44]

Schoettler, C., & Owen, J. E. 2024, MNRAS, 533, 3484, doi: 10.1093/mnras/stae1900

work page doi:10.1093/mnras/stae1900 2024
[45]

, keywords =

Spurzem, R., Giersz, M., Heggie, D. C., & Lin, D. N. 2009, Astrophysical Journal, 697, 458, doi: 10.1088/0004-637X/697/1/458

work page doi:10.1088/0004-637x/697/1/458 2009
[46]

Sullivan, K., & Gilbert, G. J. 2026, arXiv e-prints, arXiv:2603.21897, doi: 10.48550/arXiv.2603.21897

work page doi:10.48550/arxiv.2603.21897 2026
[47]

Sundman, K. F. 1913, Acta mathematica, 36, 105

1913
[48]

E., et al

Virtanen, P., Gommers, R., Oliphant, T. E., et al. 2020, Nature methods, 17, 261 von Zeipel, H. 1910, Astronomische Nachrichten, 183, 345, doi: 10.1002/asna.19091832202

work page doi:10.1002/asna.19091832202 2020
[49]

2015, , 150, 127, 10.1088/0004-6256/150/4/127

Wisdom, J. 2015, AJ, 150, 127, doi: 10.1088/0004-6256/150/4/127

work page doi:10.1088/0004-6256/150/4/127 2015
[50]

, keywords =

Wisdom, J., & Holman, M. 1991, AJ, 102, 1528, doi: 10.1086/115978

work page doi:10.1086/115978 1991