Generalized Inner–Outer (Offset) Degeneracy

• Generalized inner–outer degeneracy is defined by matching the magnification along the lens axis, causing distinct planetary configurations to yield nearly identical light curves.
• The framework unifies classical close–wide and inner–outer degeneracies, extending the concept to resonant topologies through analytic and numerical methods.
• The derived offset mapping offers a predictive, scalable tool for modeling degenerate lens configurations in both two-body and multi-planet microlensing systems.

Searching arXiv for the cited papers to ground the article in the relevant literature. Generalized inner–outer (offset) degeneracy is a magnification degeneracy in two-body planetary microlensing in which two distinct $2\mathrm{L}1\mathrm{S}$ lens configurations with the same planetary mass ratio $q\ll1$ but different projected separations $s_A$ and $s_B$ produce nearly identical light curves because the source trajectory passes through, or sufficiently close to, a lens-axis null where their magnifications coincide. In this formulation, the degeneracy is defined by magnification matching along the star–planet axis rather than by caustic-shape symmetry, and it unifies the classical close–wide and inner–outer degeneracies while extending through resonant topologies (Zhang et al., 2021).

1. Microlensing definition and geometric content

In the planetary regime, the host star is placed at the origin and the planet on the real axis at $x=s$, with source-plane coordinates measured in units of the Einstein radius $\theta_E$. The standard notation is

$$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$

and the complex binary-lens equation is written as

$z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$

The generalized inner–outer, or offset, degeneracy compares two lenses with the same $q$ and different separations $s_A$ and $q\ll1$0, and studies the magnification difference

$q\ll1$1

A central empirical object is the null: a ring-like curve in the source plane on which $q\ll1$2. Along the star–planet axis, this null intersects at

$q\ll1$3

In the planetary regime, the numerically measured null location deviates from this expression by less than about $q\ll1$4 over a wide range of separations, except at very extreme $q\ll1$5, and is essentially independent of $q\ll1$6. The source trajectory is characterized by impact parameter $q\ll1$7 and angle $q\ll1$8, so its intercept on the lens axis is

$q\ll1$9

The degeneracy occurs when $s_A$0. This “source–null matching principle” is the operational definition of the offset degeneracy (Zhang et al., 2021).

The quantity $s_A$1 is also the perturbative axis location of the planetary caustic,

$s_A$2

so $s_A$3 can be interpreted as the midpoint between the approximate planetary-caustic positions of the two lenses. This suggests that the generalized inner–outer language reflects a shift from caustic-centered reasoning to axis-centered magnification matching.

2. Relation to close–wide and classical inner–outer degeneracies

Historically, the close–wide degeneracy was formulated as the approximate invariance of the central caustic under

$s_A$4

for $s_A$5 and $s_A$6. The classical inner–outer degeneracy was derived in the Chang–Refsdal approximation for planetary caustics, where source trajectories passing the inner and outer sides of a caustic can yield nearly identical perturbations. Both were therefore understood as caustic-symmetry degeneracies.

The offset formulation changes the level of description. If $s_A$7, then

$s_A$8

and therefore

$s_A$9

In that case the null ring collapses to a single point at the origin, so the magnification difference vanishes only at that point rather than along an extended axis segment. The close–wide relation is therefore a transition point of the offset family, not the generic observed degeneracy. This is the basis for the statement that the commonly invoked close–wide degeneracy “essentially never arises in actual events”: real events have finite impact parameter, and the source rarely passes exactly through the origin (Zhang et al., 2021).

The same framework also subsumes the classical inner–outer degeneracy. In the planetary-caustic limit, $s_B$0 encodes the axis position of the planetary caustic, and the offset condition expresses magnification matching through the midpoint of two such positions. However, the offset degeneracy does not require the Chang–Refsdal approximation, does not require symmetry of a single caustic, and remains applicable when the event lies in resonant or semi-resonant topology where the standard inner–outer derivation is not strictly valid. A central consequence is the distinction between caustic degeneracy and magnification degeneracy: degenerate caustics do not strictly result in degenerate magnifications (Zhang et al., 2021).

3. Mathematical formulation and conditions of exactness

A mathematical treatment derives the offset degeneracy directly from the binary-lens equation and from the magnification polynomial along the lens axis. With the source on the axis, $s_B$1, the lens equation can be reduced to a fifth-degree polynomial in image position, and the image magnifications can be encoded by a fifth-degree polynomial in $s_B$2. This factorization separates the off-axis and on-axis images and permits analytic treatment of the total magnification on the lens axis (Zhang et al., 2022).

For two lenses with equal $s_B$3, the lens-axis null is defined by equality of total magnifications. The resulting derivation yields, to zeroth order in $s_B$4,

$s_B$5

Inside caustics, this expression is exact at zeroth order in $s_B$6. Outside caustics, the equality condition becomes a quartic in $s_B$7, and the same zeroth-order solution is exact whenever

$s_B$8

The analysis therefore establishes the offset degeneracy as a mathematical degeneracy rather than a purely numerical pattern. It also clarifies the status of the close–wide limit: the relation $s_B$9 is compatible with the offset condition only when $x=s$0, that is, when the source crosses the primary (Zhang et al., 2022).

The same treatment introduces an effective-primary coordinate shift,

$x=s$1

to align central and resonant caustics consistently. This does not alter the zeroth-order term of $x=s$2, but it regularizes the first-order correction in $x=s$3. The paper also examines oblique trajectories numerically. The analytic result guarantees exact matching on the lens axis to the stated order, whereas off-axis persistence depends on trajectory angle and on how rapidly the magnification field changes away from the null.

4. Predictive mapping and construction of degenerate counterparts

For practical modeling, the offset degeneracy provides a direct mapping between paired separations. If one begins with a known planetary solution $x=s$4 having parameters $x=s$5, one first computes the lens-axis intercept

$x=s$6

Equating this with the null position gives

$x=s$7

which can be solved for the degenerate separation: $x=s$8 The positive root is taken so that $x=s$9. In the planetary limit, one sets $\theta_E$0, because the null location and shape are nearly independent of $\theta_E$1 for $\theta_E$2 (Zhang et al., 2021).

This formula is the paper’s explicit predictive rule for generalized inner–outer degeneracy. It applies to close, wide, and resonant combinations, and it recovers the close–wide transformation as the special case $\theta_E$3. The formal treatment recasts the same condition with $\theta_E$4, emphasizing that the relevant observable is the point where the source trajectory crosses the lens axis near the anomaly. A plausible implication is that the parameter relation is best viewed as a trajectory-conditioned mapping in separation space, rather than as a topology-conditioned caustic symmetry (Zhang et al., 2022).

5. Empirical discovery, reanalysis, and extension beyond two bodies

The offset degeneracy was discovered through amortized posterior estimation on simulated $\theta_E$5 events. The analysis used a neural density estimator trained on 691,257 simulated $\theta_E$6 light curves in a Roman-like survey configuration, followed by a power transformation, HDBSCAN clustering to identify multimodal posteriors, and LBFGS-B optimization to refine each posterior mode. Inspection of multimodal posteriors revealed not only inner–outer-like and close–wide-like pairs but also resonant degeneracies that did not satisfy the classical relations, which led to the identification of the null-ring structure and the offset mapping (Zhang et al., 2021).

A systematic reanalysis of 23 planetary microlensing events with exactly two degenerate $\theta_E$7 solutions showed that all 23 events obey the null condition and the separation-mapping formula. In the $\theta_E$8 versus $\theta_E$9 plane, the source trajectories pass very close to the predicted null. In the $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$0 versus $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$1 plane, the published values and the values predicted from the offset formula essentially coincide, even for events lying significantly off the pure close–wide diagonal. The resulting classification includes cases traditionally labeled inner–outer, resonant-close or resonant-wide, and events that are not obviously close–wide or inner–outer in the classical sense. The empirical conclusion is that the offset degeneracy is ubiquitous for two-fold degenerate planetary events in the selected sample (Zhang et al., 2021).

The mathematical treatment further shows that the superposition principle permits a generalization to $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$2-body microlenses with $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$3 planetary components and $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$4, producing a $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$5-fold degeneracy. In this picture, each planet contributes an independent two-branch offset choice, and the total perturbation is approximately the sum of the single-planet perturbations. This establishes the generalized inner–outer degeneracy as a scalable perturbative framework rather than a phenomenon restricted to a single planet (Zhang et al., 2022).

6. Limitations, observational breakdowns, and terminological boundaries

The offset degeneracy is strongest in the planetary, moderate-separation regime. Its clean independence from $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$6 holds for $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$7, with deviations appearing only for $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$8, and for $$q=\frac{M_p}{M_\star},\qquad q\ll1,\qquad s=\frac{d_\perp}{\theta_E},$$9 the mapping may involve changes in both $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$0 and $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$1. For extreme separations, $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$2, deviations of the null position from the simple analytic expression can exceed $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$3, although those regimes contribute negligible planetary perturbations. Caustic-crossing events often require near-vertical trajectories for strong degeneracy, and oblique trajectories that probe regions far from the axis can reveal detectable differences. Higher-order effects such as parallax, xallarap, orbital motion, and finite-source effects can further break or complicate the degeneracy (Zhang et al., 2021).

A common misconception is that the phrase “generalized inner–outer” should be transferred directly across unrelated fields. In noncommutative Hardy algebras, “inner–outer” refers to a factorization theory for vectors, commutant operators, and Hardy-algebra elements associated with $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$4-correspondences, where the “offset” is a change of representation and the residual ambiguity is uniqueness up to a unitary intertwiner rather than a microlensing null on the lens axis (Helmer, 2015). In algebraic geometry and CAGD, “generic offset” refers to Euclidean offsets of algebraic surfaces and to a generic offset polynomial $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$5, with bad distances marking degeneracies of offset surfaces; this is again unrelated to two-body microlensing magnification matching (Segundo et al., 2010).

Within microlensing proper, the generalized inner–outer, or offset, degeneracy is therefore best understood as a magnification-level symmetry statement: the perturbation induced by a low-mass companion is organized by the combination $z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$6, and two configurations are degenerate when the source crosses the common lens-axis null determined by

$z_s = z - \frac{1}{\bar z} - \frac{q}{\bar z - s}.$7

This formulation replaces purely caustic-based language with a trajectory-conditioned null-matching rule, while preserving the close–wide and inner–outer limits as special cases of a single continuous degeneracy family (Zhang et al., 2022).