Pith. sign in

REVIEW 3 major objections 5 minor 34 references

This paper argues that binary orbital motion has a negligible effect on the population-averaged caustic-crossing rate in microlensing, while predicting that 6.3±0.2% of all events with impact parameter below one Einstein radius should exhib

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

T0 review · deepseek-v4-flash

2026-08-01 06:31 UTC pith:T74O5PL6

load-bearing objection Useful and mostly solid—the 6.3% rate is a genuine prediction for Roman—but the 'orbital motion is negligible' claim rests on an explicit, untested extrapolation that should be either hardened or softened before publication. the 3 major comments →

arxiv 2607.21845 v1 pith:T74O5PL6 submitted 2026-07-23 astro-ph.SR astro-ph.EPastro-ph.GA

Rate of caustic crossing microlensing events in stellar binary lenses with significant orbital motion

classification astro-ph.SR astro-ph.EPastro-ph.GA
keywords microlensingcaustic crossingbinary lensorbital motionevent ratecross sectiongalactic bulgepopulation synthesis
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper tackles a question that matters for interpreting upcoming microlensing surveys: do binary stars that move during an event produce noticeably more caustic-crossing events? It shows that face-on binaries with orbital periods comparable to the event timescale can have caustic cross sections up to four times larger than static binaries, but that inclined orbits shrink the cross section enough to cancel most of this gain. Averaging over a realistic synthetic sample of ~2800 Milky Way binary lenses, the net fractional increase in the caustic-crossing cross section from orbital motion is only ~0.1%, so treating binaries as static remains adequate for rate predictions. As a separate, usable prediction, the paper estimates that 6.3±0.2% of microlensing events with impact parameter u0<1 should show caustic crossings, a number that depends on binary separation, mass ratio, and multiplicity distributions and can be tested with dense photometric monitoring.

Core claim

The central discovery is that, despite large per-geometry variations, orbital motion does not change the total rate of caustic-crossing microlensing events. For face-on circular binaries with P~t_E, rotating caustics sweep larger area and can increase the cross section by up to ~4x, while inclined orbits reduce the time-averaged projected separation and can make the cross section smaller than a static face-on binary. In a sample of 188 simulated binary events selected to maximize the orbital-motion effect, the mean fractional change in cross section is only 1.5±0.8%, and extrapolating to the full ~2800-event population gives ~0.1%. The paper also derives a caustic-crossing event rate of 6.3±

What carries the argument

The key object is the angle-averaged caustic width (linear cross section) w, proportional to the caustic-crossing event rate. For static caustics, w equals the perimeter of the convex hull of the caustic divided by π; for orbiting binaries it is computed numerically by time-evolving the binary axis and projected separation and averaging over orbital phase and source-trajectory angle. The cancellation between face-on enhancement and inclined suppression is the mechanism that produces the small net effect.

Load-bearing premise

The 0.1% headline assumes that binary systems outside the selected cut (P/t_E<12, s_max>0.15) contribute on average zero change to the cross section, even though only 188 of ~2800 events were actually computed with orbital motion; if those uncomputed systems systematically enhance the cross section, the negligible-rate conclusion would not hold.

What would settle it

Run the full sample of ~2800 synthetic binary events through the orbital-motion cross-section code, computing the mean fractional change for all systems, not just the 188 with P/t_E<12 and s_max>0.15; a population-mean change larger than about 0.3% would refute the 'negligible' conclusion.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Static-lens rate calculations are sufficient for survey planning; no need to model orbital motion in population rate estimates.
  • The 6.3±0.2% rate provides a nearly detection-efficiency-independent observable for future high-cadence surveys.
  • Because the rate is sensitive to binary separation, mass ratio, and multiplicity distributions, deviations from it would signal incorrect demographic assumptions.
  • Face-on binaries with short periods should be over-represented among caustic-crossing events, even though the overall rate is unchanged; detecting such systems in future data would confirm the mechanism.
  • The cross section is more sensitive to the projected separation distribution than the mass-ratio distribution, so rate comparisons constrain separation distributions most strongly.

Where Pith is reading between the lines

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

  • If real, the 0.1% result implies that retro-fitting orbital motion in individual events is a modeling detail, not a survey-statistics concern; statistical samples can be interpreted with static models.
  • The paper's dependence on a synthetic galaxy means the 6.3% number is only as good as the underlying multiplicity and separation distributions; a measured rate outside the quoted range might point to the population model rather than to orbital motion.
  • The cancellation could be sensitive to the assumed circular orbits; eccentric binaries may break the symmetry, so quantifying eccentricity effects is a natural next step.
  • The rate prediction could be tested immediately with existing high-cadence data by measuring the caustic-crossing fraction among events with well-constrained impact parameters.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. This paper develops a numerical method for computing caustic-crossing cross sections of binary microlenses, including binaries with significant orbital motion. The method is benchmarked against static-lens calculations: it reproduces the Mao & Paczynski (1991) rate, and the authors show that the larger cross sections of Baltz & Gondolo (2001) arise from using caustic perimeter rather than convex-hull perimeter (Appendix A). The authors apply the method to ~2800 binary-lens events from PopSyCLE, computing orbital-motion cross sections for 188 systems with P/t_E < 12 and s_max > 0.15. They report a mean per-system fractional increase of 1.5 ± 0.8% and extrapolate to a full-population correction of ~0.1%, concluding that orbital motion has a negligible effect on the total caustic-crossing rate. They also compute a static-lens caustic-crossing rate of 6.3 ± 0.2% for events with u0 < 1 through a mass-dependent binary fraction weighting.

Significance. If the conclusions hold, the results are practically important: Roman-era rate calculations can safely adopt static lenses, and the 6.3% caustic-crossing rate becomes a clean demographic test of binary separation, mass ratio, and multiplicity distributions. The paper is strongest in its static cross-section methodology: the explicit convex-hull distinction resolves a known discrepancy with Baltz & Gondolo; the benchmark against Mao & Paczynski is reassuring; and the numerical convergence checks (20 vs 100 α values, doubled q/α sampling in Appendix B, binary searches with quoted tolerances) are reported. The main limitation is that the central orbital-motion conclusion rests on an extrapolation from a non-representative subset of systems, and the statistic used to make that extrapolation is not the statistic needed for a rate change. This needs strengthening before the headline claim is accepted.

major comments (3)
  1. [§5.2 and §6] The central quantitative claim—that orbital motion changes the total caustic-crossing rate by only ~0.1%—relies on an untested zero-correction assumption for systems outside the P/t_E < 12, s_max > 0.15 cuts. The text states: 'If we assume that the average fractional change in cross section for all binaries outside the bounds of our sample is close to 0, then the mean fractional increase ... is ~0.1%.' This is not a derived result. Only 188 of ~2787 binary-lens events are computed with orbital motion, and these are not representative of the rate: their mean static cross section is 0.077 θ_E, six times smaller than the 0.423 θ_E full-sample average (§5.2, §5.4). Moreover, Fig. 3 shows that at P/t_E = 10—just inside the cut—face-on close-topology systems still show ~20% enhancements, so the effect does not obviously vanish for P/t_E just above 12. I request either computation of a random o
  2. [§5.2, 'average fractional increase'] The statistic used for the extrapolation is not the rate correction. The paper reports the unweighted mean of per-system fractional changes, ⟨Δw/w⟩ = (1.5 ± 0.8)%, whereas the fractional change in the total rate is (ΣΔw_i)/(Σw_i). For the selected sample, this directly relevant ratio is 4.9%—a factor of three larger. The 4.9% is dismissed because of correlated errors, but the unweighted 1.5% is then used to make the full-population estimate. The ~0.1% is obtained by scaling 1.5% by the number fraction of selected systems (188/2787 ≈ 7%) and assuming zero outside corrections; it is not scaled by the cross-section fraction, which is only ~1% (188×0.077 vs 2787×0.423). Thus even under the zero-outside assumption, the quoted 0.1% does not follow from the stated numbers. Please report (ΣΔw_i)/(Σw_i) for the selected sample and a rate-weighted estimate for the full sample.
  3. [§5.3] All orbital-motion cross sections are computed for circular orbits, although PopSyCLE includes eccentric binaries. Section 5.3 gives only qualitative arguments and concludes that the eccentric-orbit case requires detailed calculations. This is an unquantified systematic on the central claim. Since the selected sample is small, the eccentricity distribution of the computed systems should at least be reported, and the sensitivity of the mean correction to eccentricity should be tested or bounded. This is secondary to the extrapolation problem but should be addressed in a revision.
minor comments (5)
  1. [§5.2] Typo: 'orbital motion does not shave a large impact' should read 'does not have a large impact.'
  2. [§2.3 / Software section] The lens-equation solver is called 'VBBLensing' in §2.3 but 'VBMicrolensing' in the Software list; the cited code is VBMicrolensing. Please harmonize.
  3. [Table 1 / §3] The table layout is garbled: 'T able 1', the column header 'log uniform s log normal s', and the placement of the µ, σ values make the table hard to read. The log-normal parameters are defined only in the following paragraph; consider moving them into the table caption.
  4. [References] The author name 'Jaroszyński' is rendered inconsistently as 'Jaroszynski' and 'Jaroszy´ nski' across in-text citations and the reference list.
  5. [§5.4] The quoted uncertainty 6.3 ± 0.2% appears to include only sampling variance. The binary fraction, mass-ratio, and separation distribution choices will introduce additional systematic uncertainty; this should be stated explicitly or incorporated.

Circularity Check

0 steps flagged

No circularity: the paper's rate calculations are self-contained and benchmarked against external results.

full rationale

The derivation chain is not circular. The caustic-crossing cross sections are computed directly from the binary lens equation (Witt 1990) using numerical source-plane shooting, and the static-lens results are benchmarked against Mao & Paczynski (1991) and Baltz & Gondolo (2001); the paper reproduces Mao & Paczynski's 6.5% rate as 6.3%. The 6.3% event-rate prediction comes from plugging these independently computed cross sections into the standard rate formalism Γ_cc ∝ μ_rel w θ_E, with the PopSyCLE-simulated binary population and binary fractions from Offner et al. (2023). No fitted parameter is renamed as a prediction. The full-population ~0.1% orbital-motion correction rests on an explicit zero-mean extrapolation for systems outside P/t_E < 12 and s_max > 0.15, but that is an unverified modeling assumption, not a definitional identity or a self-citation chain; it does not make the conclusion equal to an input by construction. The paper's own statement 'If we assume that the average fractional change in cross section for all binaries outside the bounds of our sample is close to 0' transparently labels the extrapolation, and the computed subset shows a 1.5±0.8% effect. This is a correctness/robustness concern, not circularity. Self-citations to PopSyCLE (Lam et al. 2020; Abrams et al. 2025) are external simulation infrastructure, not unverified uniqueness arguments, and the central derivation remains independently testable.

Axiom & Free-Parameter Ledger

4 free parameters · 7 axioms · 0 invented entities

The central rate numbers rest on standard lensing theory plus two paper-specific simplifications: zero assumed correction outside the selected kinematic cuts and circular orbits for all computed systems. These assumptions are flagged but not tested, which is why soundness is 6 rather than higher. No new physical entities are introduced.

free parameters (4)
  • Static cross-section small-s power-law coefficients = w ≈ 2.81 s^2.58 q^0.50
    Fit to the paper's numerical cross sections for s < 0.5 and used to extrapolate the rate integral in §3; not independently derived.
  • Static cross-section large-s power-law coefficients = w ≈ 6.87 s^-2.67 q^0.49
    Fit to the paper's numerical cross sections for s > 2 and used in the rate integral for large separations.
  • Log-normal s distribution parameters for Table 1 = μ = -0.57, σ = 0.88
    Fit to the PopSyCLE s distribution; enters the alternative rate estimates in Table 1, not the primary 6.3% estimate.
  • Sampling-domain cutoffs (β and torus radii) = β = 5 (s ≤ 0.5), β = 2 (0.5 < s ≤ 2), β = 3 (2 < s ≤ 6); torus r_in = 3θ_E, r_out = 6θ_E
    Hand-chosen to balance runtime against completeness; the paper shows caustics are missed at low s and for q ≪ 1 at large s, so these choices directly affect the numerical cross sections.
axioms (7)
  • standard math The binary lens equation (Witt 1990, Eq. 1) accurately describes the lens mapping.
    Used throughout §2.1; standard gravitational lensing result, independent of this paper.
  • domain assumption Caustic-crossing rate is proportional to angle-averaged caustic width; event rate Γ_cc ∝ μ_rel w θ_E (Eq. 6).
    Assumes source trajectories are uniformly distributed in impact parameter and angle; standard microlensing rate formalism (Gaudi 2012).
  • standard math Mean width of a closed concave caustic equals the perimeter of its convex hull divided by π (Appendix A).
    Used to explain the discrepancy with Baltz & Gondolo; from integral geometry (Santaló 2004).
  • domain assumption PopSyCLE output faithfully represents the Milky Way binary population relevant to Roman microlensing.
    The 2787-event sample and binary fractions rely on PopSyCLE; if the population synthesis is wrong, the 6.3% rate shifts.
  • ad hoc to paper All selected binaries are modeled with circular orbits; eccentric-orbit effects are argued qualitatively (§5.3).
    The population sample contains eccentric systems, but orbital-motion cross sections are computed for circular orbits. This simplification could bias the 1.5% average and the 0.1% conclusion.
  • ad hoc to paper The average orbital-motion correction for systems outside the P/t_E < 12 and s_max > 0.15 cuts is zero (§5.2).
    Stated explicitly as an assumption; it converts the 1.5% selected-sample mean into the 0.1% full-population conclusion. No test is provided.
  • domain assumption Binary fractions from Offner et al. (2023), with half the triple/higher-order fraction added, are applicable to the lens population (§5.4).
    Used to weight cross sections and produce the 6.3% rate; uncertainties in these weights are not propagated.

pith-pipeline@v1.3.0-alltime-deepseek · 17065 in / 14842 out tokens · 126634 ms · 2026-08-01T06:31:39.045698+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Rate of caustic crossing microlensing events in stellar binary lenses with significant orbital motion." pith.science (2026). https://pith.science/paper/T74O5PL6

@misc{pith2026260721845,
  author       = {Pith},
  title        = {Pith review of: Rate of caustic crossing microlensing events in stellar binary lenses with significant orbital motion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T74O5PL6}},
  note         = {Machine review of arXiv:2607.21845}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Binary lens microlensing events in which the source crosses a caustic produce sharp, distinctive magnification peaks and can therefore be readily identified. In this paper, we explore the importance of binary orbital motion for the binary lens caustic crossing cross section. If the orbital timescale of the binary system is smaller than the Einstein ring crossing timescale, the caustics sweep out a larger area in the source plane, generally enhancing the cross section. We find that face-on binaries in circular orbits exhibit a substantial increase (up to 4$\times$) in the cross section for caustic crossings. However, highly inclined orbits produce a net decrease relative to a static face-on binary with the same semi-major axis. Using a sample of ~$2800$ synthetic binary microlensing events drawn from a realistic Milky Way population-synthesis model, we calculate the average change in the caustic crossing cross section for individual systems with and without orbital motion. Although orbital motion can significantly alter the cross section for specific geometries, we find that, when averaged over the full population, it produces only a small, negligible increase of ~$0.1\%$ in the average cross section. We also compute the overall rate of caustic crossing binary events in the simulated sample and find that, with sufficiently dense photometric sampling, $6.3 \pm 0.2\%$ of microlensing events with impact parameter $u_0 < 1$ should exhibit caustic crossings. This rate depends on the distributions of binary separation, mass ratio, and multiplicity of stars and compact objects, and can therefore be used to test our understanding of these underlying properties.

Figures

Figures reproduced from arXiv: 2607.21845 by Arjun Murlidhar, B. Scott Gaudi, Todd. A. Thompson.

Figure 1
Figure 1. Figure 1: An example of the cross section calculation for close (red curves), resonant (orange curve), and wide (green curves) caustics for q = 0.6. The center of mass of the binary lens is located at the origin, and both lens components lie on the X-axis, with the more massive companion positioned to the left of the origin. The black dashed lines show the width of the caustics for a source trajectory angle α = 0°, … view at source ↗
Figure 2
Figure 2. Figure 2: Cross section of caustics for a static binary lens as a function of the separation of lens components (in units of θE) for 20 q values between 0.05 and 1 plotted using colored lines. The average cross section over these 20 q values is plotted in red. In [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (Left) Caustic cross section for face-on binary stars in a circular orbit as a function of s for different values of the ratio of orbital period of the binary to the Einstein ring crossing time. Each curve is an average over 20 q values. The gray dashed curve shows the cross section for a face-on binary system with an infinitely small orbital period and is equal to the maximum distance between two points o… view at source ↗
Figure 4
Figure 4. Figure 4: Plots of primary mass of the binary vs maximum projected separation for different lens distances (Dl) and P/tE values. The source distance Ds is fixed at 9.76 kpc. These relations are obtained by averaging Eq. 9 over a uniform mass ratio (q) distribution. The colored dashed lines correspond to the median lens-source relative proper motion value derived from PopSyCLE (µrel,med = 6.65 mas yr−1 ), and the sha… view at source ↗
Figure 5
Figure 5. Figure 5: (Left) Caustic cross section as a function of smax (a/rE) for stellar binary lenses in circular orbits inclined to our line of sight at different angles. Each point is the average cross section over 20 q values and 10 values of orbital phase ϕ. This plot compares the cross sections for 3 different inclinations when P/tE = 3, and shows the cross section for one value of the inclination when P/tE = 10. Binar… view at source ↗
Figure 6
Figure 6. Figure 6: Distribution of parameters for the full sample of 2787 binary lens microlensing events from PopSyCLE to make a valid comparison between the static and or￾biting cross sections, we must also account for the fact that, depending on the phase ϕ, the projected separa￾tion s during the event will typically be smaller than smax. Therefore, even when computing the cross sec￾tions for static binaries, we average o… view at source ↗
Figure 7
Figure 7. Figure 7: Caustic cross section for simulated binary lens events produced by PopSyCLE which have P/tE < 12 and smax > 0.15. Plot on the left is a scatter plot of smax and P/tE for all lenses in this sample. Orange lines show the cuts that have been applied. Right shows the cross section as a function of smax, where the systems have been binned in smax. The points are the mean cross section values of the binned syste… view at source ↗
Figure 8
Figure 8. Figure 8: Mean fractional change in the cross section due to orbital motion effects as a function of P/tE. The systems have been binned in P/tE and the x error bar represents the bin width. The y error bar is the error in the mean. Fig. 7b shows the cross section as a function of smax, where the binary lens systems have been binned in smax and the mean cross section in each bin has been plotted. The green curve show… view at source ↗
Figure 9
Figure 9. Figure 9: Fractional change in cross section due to orbital motion relative to a static binary lens for each system in Fig. 7a. Plot on the left shows the differential distribution of the fractional change across the sample. Plot on the right shows the fractional change as a function of inclination, where the systems have been binned in cosine of the inclination angle. The horizontal error bars show the bin widths a… view at source ↗
Figure 10
Figure 10. Figure 10: Comparison of static binary lens cross sections calculated by E. A. Baltz & P. Gondolo (2001) (green solid line) with the cross section calculated in this work (red dots and dashes). While the two curves are qualitatively similar, the cross sections we calculate are smaller than those in the Baltz and Gondolo paper for all s values. E. A. Baltz & P. Gondolo (2001) define the “width” of a caustic, w as (eq… view at source ↗
Figure 11
Figure 11. Figure 11: (Left) Caustic cross sections for a static lens and an orbiting P/tE = 3 in the resonant caustic regime. The green curve is the original calculation, and the orange curve is the calculation with twice the number of q and α values. (Right) Caustic structure (blue) and source trajectories for a binary lens in the resonant regime with s = 1.1 and q = 0.9. The red dotted line is the actual source trajectory, … view at source ↗
Figure 12
Figure 12. Figure 12: Caustics for q = 1, s = 6.5, 7, and 8 and q = 0.1, s = 6. The shaded region represents the torus with rin = 3 and rout = 6. Caustics for equal mass binaries with s > 6 lie in the torus. For very unequal mass ratios and large s (see q = 0.1, s = 6), one of the two diamond caustics is outside the torus region [PITH_FULL_IMAGE:figures/full_fig_p016_12.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

34 extracted references · 10 canonical work pages · 5 internal anchors

  1. [1]

    S., Lu, J

    Abrams, N. S., Lu, J. R., Lam, C. Y., et al. 2025, ApJ, 980, 103, doi: 10.3847/1538-4357/ada5f9

  2. [2]

    D., Beaulieu, J.-P., Caldwell, J

    Albrow, M. D., Beaulieu, J.-P., Caldwell, J. A. R., et al. 2000, ApJ, 534, 894, doi: 10.1086/308798

  3. [3]

    A., Alves, D., et al

    Alcock, C., Allsman, R. A., Alves, D., et al. 2000, ApJ, 541, 270, doi: 10.1086/309393

  4. [4]

    H., Albrow, M

    An, J. H., Albrow, M. D., Beaulieu, J.-P., et al. 2002, ApJ, 572, 521, doi: 10.1086/340191

  5. [5]

    A., & Gondolo, P

    Baltz, E. A., & Gondolo, P. 2001, ApJ, 559, 41, doi: 10.1086/322402

  6. [6]

    P., Rhie, S

    Bennett, D. P., Rhie, S. H., Nikolaev, S., et al. 2010, ApJ, 713, 837, doi: 10.1088/0004-637X/713/2/837

  7. [7]

    Perturbative analysis in planetary gravitational lensing

    Bozza, V. 1999, A&A, 348, 311, doi: 10.48550/arXiv.astro-ph/9904297

  8. [9]

    2013, ApJ, 768, 129, doi: 10.1088/0004-637X/768/2/129

    Choi, J.-Y., Han, C., Udalski, A., et al. 2013, ApJ, 768, 129, doi: 10.1088/0004-637X/768/2/129

  9. [10]

    1999, A&A, 349, 108, doi: 10.48550/arXiv.astro-ph/9903014

    Dominik, M. 1999, A&A, 349, 108, doi: 10.48550/arXiv.astro-ph/9903014

  10. [11]

    Gaudi, B. S. 2012, ARA&A, 50, 411, doi: 10.1146/annurev-astro-081811-125518

  11. [12]

    S., Bennett, D

    Gaudi, B. S., Bennett, D. P., Udalski, A., et al. 2008, Science, 319, 927, doi: 10.1126/science.1151947

  12. [13]

    1991, ApJ, 366, 412, doi: 10.1086/169575

    Griest, K. 1991, ApJ, 366, 412, doi: 10.1086/169575

  13. [14]

    A., et al

    Han, C., Udalski, A., Bond, I. A., et al. 2024, A&A, 686, A234, doi: 10.1051/0004-6361/202349063

  14. [15]

    Binary Lenses in OGLE-II 1997-1999 Database. A Preliminary Study

    Jaroszynski, M. 2002, AcA, 52, 39, doi: 10.48550/arXiv.astro-ph/0203476

  15. [16]

    Binary Lenses in OGLE-III EWS Database. Seasons 2002--2003

    Jaroszynski, M., Udalski, A., Kubiak, M., et al. 2004, AcA, 54, 103, doi: 10.48550/arXiv.astro-ph/0408243

  16. [17]

    Binary Lenses in OGLE-III EWS Database. Season 2004

    Jaroszynski, M., Skowron, J., Udalski, A., et al. 2006, AcA, 56, 307, doi: 10.48550/arXiv.astro-ph/0701919 Jaroszy´ nski, M., Skowron, J., Udalski, A., et al. 2010, AcA, 60, 197, doi: 10.48550/arXiv.1009.5563

  17. [18]

    Y., Lu, J

    Lam, C. Y., Lu, J. R., Hosek, Jr., M. W., Dawson, W. A., & Golovich, N. R. 2020, ApJ, 889, 31, doi: 10.3847/1538-4357/ab5fd3

  18. [19]

    D., Yee, J

    Malpas, A., Albrow, M. D., Yee, J. C., et al. 2022, AJ, 164, 102, doi: 10.3847/1538-3881/ac7d4c

  19. [20]

    1991, ApJL, 374, L37, doi: 10.1086/186066

    Mao, S., & Paczynski, B. 1991, ApJL, 374, L37, doi: 10.1086/186066

  20. [21]

    Offner, S. S. R., Moe, M., Kratter, K. M., et al. 2023, in Astronomical Society of the Pacific Conference Series, Vol. 534, Protostars and Planets VII, ed. S. Inutsuka, Y. Aikawa, T. Muto, K. Tomida, & M. Tamura, 275, doi: 10.48550/arXiv.2203.10066

  21. [22]

    1986, ApJ, 304, 1, doi: 10.1086/164140

    Paczynski, B. 1986, ApJ, 304, 1, doi: 10.1086/164140

  22. [23]

    2009, ApJ, 690, 1772, doi: 10.1088/0004-637X/690/2/1772

    Pejcha, O., & Heyrovsk´ y, D. 2009, ApJ, 690, 1772, doi: 10.1088/0004-637X/690/2/1772

  23. [24]

    T., Gaudi, B

    Penny, M. T., Gaudi, B. S., Kerins, E., et al. 2019, ApJS, 241, 3, doi: 10.3847/1538-4365/aafb69

  24. [25]

    T., Kerins, E., & Mao, S

    Penny, M. T., Kerins, E., & Mao, S. 2011a, MNRAS, 417, 2216, doi: 10.1111/j.1365-2966.2011.19403.x

  25. [26]

    T., Mao, S., & Kerins, E

    Penny, M. T., Mao, S., & Kerins, E. 2011b, MNRAS, 412, 607, doi: 10.1111/j.1365-2966.2010.17933.x

  26. [27]

    T., Kerins, E., Rattenbury, N., et al

    Penny, M. T., Kerins, E., Rattenbury, N., et al. 2013, MNRAS, 434, 2, doi: 10.1093/mnras/stt927

  27. [28]

    2018, MulensModel: Microlensing light curves modeling,, Astrophysics Source Code Library, record ascl:1803.006 http://ascl.net/1803.006

    Poleski, R., & Yee, J. 2018, MulensModel: Microlensing light curves modeling,, Astrophysics Source Code Library, record ascl:1803.006 http://ascl.net/1803.006

  28. [29]

    Poleski, R., & Yee, J. C. 2019, Astronomy and Computing, 26, 35, doi: 10.1016/j.ascom.2018.11.001 Santal´ o, L. A. 2004, Integral Geometry and Geometric

  29. [30]

    1986, A&A, 164, 237

    Schneider, P., & Weiss, A. 1986, A&A, 164, 237

  30. [31]

    2012, arXiv e-prints, arXiv:1203.1034, doi: 10.48550/arXiv.1203.1034

    Skowron, J., & Gould, A. 2012, arXiv e-prints, arXiv:1203.1034, doi: 10.48550/arXiv.1203.1034

  31. [32]

    Binary Lenses in OGLE III EWS Database. Season 2005

    Skowron, J., Jaroszynski, M., Udalski, A., et al. 2007, AcA, 57, 281, doi: 10.48550/arXiv.0802.3704

  32. [33]

    2014, AJ, 147, 87, doi: 10.1088/0004-6256/147/4/87

    Tokovinin, A. 2014, AJ, 147, 87, doi: 10.1088/0004-6256/147/4/87

  33. [34]

    Witt, H. J. 1990, A&A, 236, 311

  34. [35]

    2000, ApJ, 541, 728, doi: 10.1086/309468

    Zheng, Z., & Gould, A. 2000, ApJ, 541, 728, doi: 10.1086/309468