A larger payoff in the model
A set of simulations suggests that changing pulsar-timing plans could improve the hunt for a single, individually resolvable supermassive black-hole binary. Under the baseline plan, the modeled detection fraction was 10% for the Light population, 16% for the Intermediate population and 27% for the Heavy population. With both timing precision and cadence improved, those figures rose to 12%, 21% and 33%, respectively.
Those percentages are outcomes across simulated cases, not observations of confirmed individual binaries. The study asks what the first individually resolvable source might look like, and how pulsar timing could be arranged to find it and support electromagnetic identification.
The biggest modeled sensitivity gain came from sampling more often. Forecasts estimated a factor of about 1.5 when cadence was increased fourfold, compared with about 1.2 with improved timing precision; combining the two changes gave a factor of about 2.1 relative to the current strategy.
That result aligns with the simulated distribution of information across the array: about 70% of the continuous-wave signal typically came from the ten most sensitive pulsars. In practical terms, the forecast points toward concentrating extra observations on those best-timed targets.
How the forecast was built
To build the forecast, the researchers used population synthesis constrained by the simulated stochastic-background signal and examined the loudest binary in each realization. They represented each of three cosmological scenarios—Light, Intermediate and Heavy—with 500 simulated populations.
The populations were matched to a stochastic-background signal-to-noise ratio between 3.0 and 3.4. Each simulation was extended with 4.5 years of injected observations to create a synthetic nine-year dataset. A continuous-wave signal-to-noise ratio of 5.94 served as the conservative detection threshold.
That setup is why the results should be read as detection fractions inside a model: a case counted when the loudest simulated source crossed the chosen threshold. It is not a count of binaries already seen in real data.
The strategy tests produced a clear but uneven pattern. Doubling precision yielded simulated fractions of 11%, 17% and 28% for Light, Intermediate and Heavy populations. Increasing cadence yielded 12%, 18% and 30%, while applying both changes yielded 12%, 21% and 33%. A combined plan that conserved telescope time was instead lower than the fiducial plan by as much as 10%.
What kind of source might stand out
The loudest simulated binaries were most often found around 15–20 nanohertz, although the possible range was broad. The reported 95% credible intervals ran from 2.3 to 77 nHz for Light, 2.6 to 74 nHz for Intermediate and 2.8 to 73 nHz for Heavy populations.
The authors’ overall profile for a first resolved source is a massive, relatively close system with redshift below 2 and a gravitational-wave frequency between 2 and 75 nHz. This is a population-level description rather than a prediction of one object with fixed properties.
Another result was exploratory: the simulated strain distributions were bimodal, with two distinct groupings rather than one. All of the highest-signal-to-noise sources in the subgroup discussed came from the lower-strain B mode. The origin of that split was left for future study.
The harder half: finding a counterpart
The paper also considers whether a gravitational-wave candidate could be checked against optical variability. In its electromagnetic case study, a useful signal had to complete at least five observed cycles in a ten-year dataset, so the analysis focused on frequencies of at least 32 nHz, corresponding to changes on timescales of no more than two years. About 15% of resolvable systems met that criterion; 85% fell below it.
That leaves a relatively small high-frequency subset for electromagnetic follow-up. The study’s broader recommendation is to favor observing strategies that strengthen continuous-wave sensitivity while keeping the sources most suitable for cross-checks in view.
The document is a draft arXiv preprint, and its numbers are forecasts based on selected population models and synthetic observing plans. They do not establish that an individual binary has been detected, nor do they show that the same strategy will deliver the same result in real observations. The study is instead a guide to where future timing effort might yield the most information under the assumptions tested.
Paper data and sources
Original title: The properties of the first continuous gravitational waves: Optimizing pulsar timing observations for supermassive black hole binary detection
Authors: Blaze Houlden, Eric Thrane, Katie Auchettl
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text