Built from Gaia DR3 using seven stellar parameters, the Torlakcik Catalog removes 55% of the sample and ships a public pipeline any survey can rerun on its own target list.
The story SETI tells itself is a sensitivity story. Bigger dishes, faster digitizers, smarter signal processing, and the answer will fall out of the noise. A new paper accepted to the Publications of the Astronomical Society of the Pacific argues the binding constraint is actually triage. With up to 400 billion stars in the Milky Way and only so many hours on the world's largest radio telescopes, the harder problem is choosing where to point.
The Torlakcik Catalog, outlined in a preprint on arXiv by Sahin Torlakcik, doesn't crown a shortlist of likely homes for alien civilizations. It does the opposite. Apply a rule-based filter with seven stellar parameters and 55% of the input falls away. The retained 777,835 high-priority targets are dominated by K dwarfs, the orange dwarfs that burn steadily for tens of billions of years, and quiet M dwarfs, the smallest, dimmest main-sequence stars. These are the stars that have been around long enough, are not too massive, and are not flaring their planets into sterility.
Two parameters do most of the work. The model removes roughly 29% of the sample on age (stars under 3 Gyr haven't had evolutionary time for complex life) and 29% on metallicity (sub-solar metal content hampers rocky planet formation). Mass cuts add more: stars above 1.5 solar masses burn out before biology has a chance. Binarity matters because destabilizing binary companions shred planetary orbits. Flaring and chromospheric activity irradiate planets. Photometric variability and a Gaia-derived astrometric noise proxy called RUWE catch the noisier cases.
The paper's most striking methodological move is small but consequential. Gaia, the European Space Agency's astrometry mission, returns age estimates with large uncertainties. A naive filter using point estimates would throw out 355,086 stars whose upper-bound ages do meet the 3 Gyr floor. Using upper bounds instead rescues that population, per the arXiv preprint. The aggregate result is robust: about 55% of the sample is excluded either way. The individual target assignments are not. The same paper notes that RUWE flags 2.7 times as many binaries as Gaia's own non-single-star catalog, a reminder that switching proxies reshuffles individual targets even when the aggregate rate holds.
The catalog also stress-tests an existing survey. Cross-matched against the Breakthrough Listen primary target list, a survey that points the Green Bank Telescope and Parkes at nearby stars in search of narrowband radio technosignatures, 56.5% of those targets would be flagged for exclusion under these habitability criteria, mostly because of low metallicity. The astronomy news site Universe Today frames this as a "detectability versus plausibility" contrast: the radio survey picks what its instruments can see, the new filter picks what could plausibly host a biosphere. The two lists are not the same.
The deliverable is the catalog, the full pipeline, and a generalized community tool, all publicly released, so any future survey can apply the same filter to its own target list. The seven parameters are necessary, not sufficient. The paper calls it a pruning step. The biographical framing in news writeups, which describes the author as a high school student from Ankara, comes from the Universe Today piece, not the arXiv abstract, and should be read as a claim about the author, not a fact about the science. The peer-review status rests on the PASP acceptance noted on the preprint page; the paper itself is a v1 preprint, not yet typeset.
The next test is whether the public pipeline gets adopted. If a major radio survey reruns its target list through the Torlakcik filter and publishes the difference, the catalog graduates from a peer-reviewed argument to a working constraint. Until then, the paper's claim is methodological: the search for ET is now a subtraction problem with a rulebook.