Warm Jupiters — giant planets orbiting between 0.1 and 1 AU — show a pattern that cold Jupiters don't: the more massive ones have more eccentric orbits. Dong, Lee, Kokubo, Murray-Clay, and Gupta explain this correlation through a single mechanism: planet-planet scattering, where the outcome depends on mass.
When two low-mass giant planets in a compact system destabilize and encounter each other, they collide. The merged planet inherits a relatively circular orbit. When two high-mass giants scatter, they don't merge — one ejects the other from the system entirely. The survivor retains a highly eccentric orbit, carrying the memory of the violent encounter in its orbital shape.
The transition from collision-dominated to ejection-dominated outcomes is what creates the mass-eccentricity correlation. Below a threshold mass, scattering ends in merger and circularization. Above it, scattering ends in ejection and eccentricity. The relationship between mass and orbital shape isn't set during formation — it's carved by the dynamics of destruction.
The same mechanism explains several other observations simultaneously. Stellar obliquities — the misalignment between the star's spin axis and the planet's orbit — increase with eccentricity, as expected if violent scattering events tilt orbits relative to the original disk plane. The model also predicts that warm Jupiters should have unseen companion planets at wider orbits, detectable through radial velocity surveys, because scattering from a compact multi-planet system doesn't always eject all the neighbors.
The mass doesn't cause the eccentricity. The mass determines whether the collision sticks or the competitor flies away. The orbit remembers which happened.
Note (2026-11-11): CORRECTION. Checked against the full text of the paper (Dong, Lee, Kokubo, Murray-Clay and Gupta, arXiv 2603.22426). The mass-eccentricity correlation among warm Jupiters, its interpretation as a transition from collision-dominated to ejection-dominated planet-planet scattering, and the prediction of additional companions detectable by radial velocity are as in the paper. One statement is the reverse of the paper. The essay says that stellar obliquities 'increase with eccentricity, as expected if violent scattering events tilt orbits relative to the original disk plane'. The paper says warm Jupiters are frequently observed to have low stellar obliquities, that spin-orbit alignment is common 'regardless of their eccentricities', and that this is explained by the scattering picture: at these small orbital distances inclination excitation by scattering is limited, so eccentricity can be excited without large tilts. It adds that large obliquities would suggest a migration history beyond local dynamical interactions. Two further details: the sample is 85 planets in 84 systems with periods of 10 to 365 days and masses of 0.3 to 15 Jupiter masses, and the paper presents its account as an interpretation ('we interpret', 'we argue') that fits several observations, with the possibility that warm and cold Jupiters share an origin as a further suggestion.