It often feels like we planet hunters earn our chops by how many planets we successfully discover. Why, then, did I spend the last year working with observations that we already knew hadn’t detect any planets? Such “null results” may be harder for a parent to brag about, but it can often be just as valuable to question what isn’t observed as what is. Don’t get me wrong: getting successful observations is exhilarating. But even when your observations return a null result, you can learn a lot by carefully characterizing the reasons why you aren’t seeing what you hoped for.

Let’s take a moment to consider everyone’s favorite null result: Olbers’ paradox. Yes, yes, since the beginning of humanity astronomers have been gazing up at the stars in wonder and all that jazz, but in 1823, Heinrich Wilhelm Olbers gazed up and wondered not about the stars, but about the darkness between them, asking why the sky is dark at night. If we assume a universe that is infinite, uniform and unchanging, then our line of sight should land on a star no matter where we look. For instance, imagine you are in a forest that stretches around you with no end. Then, in every direction you turn, you will eventually see a tree. Like trees in a never-ending forest, we should similarly be able to see stars in every direction, lighting up the night sky as bright as if were day. The fact that we don’t indicates that the universe either is not infinite, is not uniform, or is somehow changing.

100 years later, Edwin Hubble finally acquired observational evidence to prove that the Universe was not static but instead expanding. More distant stars are constantly moving away from us, meaning their light takes more time to reach us—and for the most distant, that’s more than the amount of time the Universe has been in existence. Olbers was able to deduce this simply by asking why it gets dark at night—a question every three-year-old asks. By noticing that we aren’t seeing as many stars as we think we should and asking the right question, he was able to probe fundamental qualities of the entire Universe.

With Olbers as our muse, Debra Fischer, John M. Brewer, and I (at Yale University) along with Bárbara Rojas-Ayala (of the Universidad Andrés Bello, in Chile) turned our sights to the Alpha Centauri star system. This system consists of three stars—Alpha Centauri A, B and C. They are, in essence, the stellar version of “the boy/girl next door”; they are close, familiar and charming. Alpha Centauri A and B orbit each other at a scant 4.37 light years from Earth. They are both orbited by Alpha Centauri C, or Proxima Centauri, which is the closest star to us at 4.25 light years away due to where it is in its 540,000 year orbit.

Alpha Centauri A is just slightly bigger than our beloved Sun, Alpha Centauri B is slightly smaller, and Proxima Centauri is about one tenth the size of the Sun. This highly charismatic system has been the target of several studies, yet only two planets have ever been discovered orbiting around any of these three stars. The first was disproved shortly thereafter. The other planet, discovered to be orbiting around Proxima Centauri, was only announced in 2016.

Though the observations have returned few real planets, they are still rich with information. Many of the observations are radial-velocity measurements, which can characterize a star’s movement toward and away from us along our line of sight. The star will wobble back and forth in a predictable, periodic way thanks to the gravity of an orbiting planet. Even though our radial-velocity measurements didn’t detect any planets, we can still use this null result to better characterize what types of planets are still possible around each star.

In order to constrain what types of planets might still exist, we imagined a range of planets from a tenth the mass of Earth to a thousand times the mass of Earth, and which orbit at distances closer to their host star than Mercury to twice as far as Earth. For simplicity, we considered only the case of one planet in the system with a perfectly circular orbit. We took over a decade of observations and “injected” the signal a planet with a given mass and separation would make if it existed. This meant simulating the observations we would expect to see from such a planet over the decade-long time scale of the existing observations. The strength of this injected signal depends on the pretend planet’s mass and its distance from the host star.

We wanted the resulting simulated observations to be as similar to the original observations as possible. To that end, we kept the same timing and frequency of the original observations. We also added a representative degree of error to each simulated point, such as what you might get when actually observing a star either from changes in the star as you’re observing it or changes in the instrument as it is doing the observing.

With the most realistic fake observations in hand, we could then run these observations through a discovery pipeline and see if the planetary signal we injected could be teased out again with greater than 99 percent confidence. In some cases, the injected signal was too weak and so was lost because observations were too infrequent or the signal became buried beneath the added error. However, if we could tease out the signal, then we could conclude that if such a planet really did exist, then we would have discovered it by now. The fact that we have found nothing allows us to definitively conclude that such a planet must not exist.

Despite finding no new planets, this analysis presented a myriad of other results. We showed that for sure the field remains wide open to habitable planets ranging from Earth-mass to a few Earth-masses, the most statistically likely type of planet according to results from the Kepler satellite. We also exclude the possibility of larger, Jupiter-mass planets, which could potentially endanger rocky planets by sending asteroids hurtling towards them or affecting the stability of their orbits. Even better, we were able to characterize the composition of these three stars, which hints at the amount of iron, magnesium or silicon available to form planets around them. These compositions suggest that if Earth-mass planets do exist around alpha Centauri A or B, then these planets are very likely to be made of the same wholesome stuff we have here on Earth.

Though we have yet to detect these possible planetary systems, a series of next-generation, radial-velocity instruments are coming online within the year that will be capable of the extreme precision needed to detect Earth analogs around alpha Centauri A, B and Proxima Centauri. Just like the boy next door right before the climax of any nineties rom-com, we might soon view these neighborhood stars in a whole new light.