Two black holes circle each other, and their dance isn't always a smooth waltz. If the holes spin on their own axes, their orbit can start to tumble, like a top that's been pushed and wobbles on its side while still spinning. This tumbling changes the waves rippling through space, and from those changes you can learn a lot: how these holes were born, how the universe is expanding, and whether general relativity holds up.
But finding such a tumbling signal in detector data is no easy task. The signal is weak, and the noise is like a market at rush hour. Scientists build 'templates'—precomputed patterns of what signals from different hole pairs might look like. If a real signal matches one of the templates, it gets noticed. The problem is that for tumbling orbits, you need a huge number of templates, and going through them all is like searching for a needle in a haystack where every straw is moving.
The researchers applied this method to data from the LIGO detectors from the third observing run, but found no new events. However, the method itself performed excellently: it catches signals 28% fainter than before and uses five times less computing power. When detectors become more sensitive, this search will help hear even the quietest tumble of a distant black hole pair.
The tumbling of a binary black hole's orbit is like a top that's been given a slight push: it doesn't fall over but starts tracing circles with its axis.