The plausibility of an unseen planet in the outer solar system, and the expected orbit and mass of such a planet, have long been a topic of inquiry and debate. We calculate the long-term orbital stability of distant trans-Neptunian objects (TNOs), which allows us to expand the sample of objects that would carry dynamical information about a hypothetical unseen planet in the outer solar system. Using this expanded sample, we find statistically significant clustering at the ∼3σ level for TNOs with semimajor axes >170\;AU, in longitude of perihelion (ϖ), but not in inclination (i), argument of perihelion (ω) or longitude of node (Ω). Since a natural explanation for clustering in ϖ is an unseen planet, we run 300 n-body simulations with the giant planets, a disk of test particles representing Kuiper belt objects, and an additional planet with varied initial conditions for its mass, semimajor axis, eccentricity, and inclination. Based on the distribution of test particles after 1--2 Gyr, we compute relative likelihoods given the actual distribution of ϖ as a function of semimajor axis for distant TNOs on stable orbits using a significantly larger sample (N=51) than previous work (N=11). We find that the unseen planet parameters that best fit the data are a mass of mp=4.4±1.1M⊕, a semimajor axis of ap=290±30AU, an eccentricity of ep=0.29±0.13, and an inclination of ip=6.8±5.0∘ (all error bars are 1σ). Only 0.06% of the Brown & Batygin (2021) reference population produce probabilities within 1σ of the maximum within our quadrivariate model, indicating that our work identifies a distinct preferred region of parameter space for an unseen planet in the solar system. If such an unseen planet exists, it is likely to be discovered by LSST.