Does the orbital velocity of the pair predict the ejected body? A test against the mass ratios

We test whether the comparison between the orbital velocity of each pair and the arrival velocity of the third body predicts which body will be ejected in the disruption of three-body systems, and whether this information adds anything to the masses. In 4,200 regularized planar integrations, with fixed mass groups and with masses drawn at random between 1 and 10, we compare the lightest body predictor, two predictors based on the velocity ratio, and a conditional logit model that combines masses and velocities, evaluated by cross-validation. With equal masses, the velocity ratio carries a weak signal, with 34% to 47% success against 33% for chance. With unequal masses, velocity alone loses to the masses, and the combined model only outperforms them when two candidates have the same mass, in which case velocity breaks the tie. With random masses there is no gain. The fit recovers an ejection probability approximately proportional to the mass raised to the power −2.4, close to the known law with exponent −3.