The planets that only count if they line up
Three in four known exoplanets were found by watching a star dim. But an Earth twin dims its star, as seen from here, only one time in 215. The catalogue of 6,451 worlds is mostly a list of the ones that happened to line up.
In 1995 nobody knew of a single planet orbiting another Sun-like star. On the first of November 2026, NASA's Exoplanet Archive listed 6,451 confirmed ones. The list passed 5,000 in March 2022 and 6,000 in September 2025, and it is filling faster every year. A new survey telescope, Roman, launched at the end of August, and Europe's PLATO is due early next year.
Almost none of those planets has been seen. They are known by what they do to their star. Of the 6,451, the archive credits 4,785 to the transit method: a planet crosses the face of its star and the starlight dips. Another 1,204 came from the wobble, the small back-and-forth motion a planet's gravity gives its star. Everything else, from microlensing to direct photographs, adds up to 462.
So roughly three-quarters of the known planets transit. It is natural to read that as a fact about planets. It is actually a fact about telescopes.
The geometry of a lucky angle
A transit only happens if we are looking at the orbit almost exactly edge-on. Tilt the orbit by a sliver and the planet passes above or below the star's disc, and nothing dims at all. For orbits pointed in random directions, the chance of a transit works out to about the star's radius divided by the radius of the orbit.
For the Sun and Earth, that is 695,700 kilometres over 149.6 million. The answer is 0.465 percent: one chance in 215. For every Earth twin a transit survey could possibly catch, roughly 214 others exist that never cross their star from our point of view, however long we watch.
Close-in planets are luckier. A giant planet orbiting at a twentieth of Earth's distance transits about one time in eleven, twenty times more often. It also blocks a full one percent of the light, against Earth's 84 parts per million, and it repeats every few days rather than once a year. That is why the first decade of discoveries was crowded with “hot Jupiters” that no one had predicted.
Calculator
Will it line up?
Set the star and the planet. The page works out the odds that the orbit is lined up for a transit, how deep the dip is, and how long it lasts.
Sun = 1 · TRAPPIST-1 = 0.1192
Sun = 1 · TRAPPIST-1 = 0.0898
Earth = 1 · Jupiter = 10.97
Earth = 1 · hot Jupiter = 0.05 · TRAPPIST-1e = 0.02925
Try TRAPPIST-1, a red dwarf about the size of Jupiter. Its planet e sits in the zone where sunlight is Earth-like, yet its odds of lining up are one in 53, its dip is sixty times deeper than an Earth twin's, and its year lasts six days. Small, close, frequent and deep is exactly the profile a transit survey rewards. That is why all seven of TRAPPIST-1's known planets were found by transit.
The wobble has the same bias
The other main method does not escape it. Jupiter tugs the Sun around at about twelve and a half metres per second, faster than an Olympic sprinter's average over 100 metres. Earth manages nine centimetres per second. In the star's light, that shows up as a colour shift of roughly 1/670,000 of the width of an atom. Massive planets close to their stars are easy. Earth twins sit at the very edge of what the best spectrographs can do.
The result is a catalogue that over-represents large planets on tight orbits around small stars. None of it is wrong. Every planet in it is real. It is simply a census taken by instruments that see some kinds of planet far more easily than others. Statisticians call this a selection effect.
A floor, not a count
The honest way to use the catalogue is to correct for what it cannot see. NASA's Kepler telescope stared at roughly 150,000 stars for four years. Its list contains 23 candidates that are about Earth's size and receive roughly Earth-like sunlight.
Make the simplest possible assumption: every one of those stars is Sun-like, and every candidate orbits at Earth's distance. If Kepler had caught every twin that lined up, then 23 caught twins stand for about 23 × 215, nearly 5,000 twins among 150,000 stars. That is about 3.3 percent. But Kepler did not catch every one. An 84-parts-per-million dip that lasts thirteen hours and returns once a year is easy to miss in noisy data. So 3.3 percent is a floor. The real fraction is larger, and astronomers still disagree about how much larger. PLATO was designed to close exactly that gap.
That is the useful habit to take from a list of 6,451 planets. Ask what the instrument could see before asking what the universe contains.
For the classroom
The Planets We Can See
24 exoplanet task cards for grades 10–12 and the first year of college, run as four 15-minute stations in a 90-minute block.
- Students derive the one-in-215 number themselves on card B3, after voting on it blind.
- Archive counts dated November 2026, transit depths, wobble speeds, Doppler shifts and TRAPPIST-1's habitable zone.
- A capstone that turns Kepler's 23 candidates into a floor, not a count, with a four-criterion rubric.
- Teacher guide, minute-by-minute pacing, two-period split, recording sheet and a fully worked answer key.
Left: the classroom packet on Teachers Pay Teachers. Right: Elizabeth Tasker's The Planet Factory: Exoplanets and the Search for a Second Earth, a book-length account of the hunt for the very planet this article shows is hardest to catch — an Amazon affiliate link. As an Amazon Associate this site earns from qualifying purchases, at no additional cost to you.
Sources
NASA Exoplanet Archive, “Exoplanet and Candidate Statistics,” accessed 1 November 2026: the 6,451 total, the counts by discovery method, and the 23 Kepler candidates (1.25 Earth radii or smaller, receiving 0.32–1.78 times Earth's sunlight). NASA/IPAC milestone announcements: 5,000 planets (March 2022) and 6,000 (September 2025). M. Mayor & D. Queloz, Nature 378, 355 (1995). E. Agol et al., Planetary Science Journal 2, 1 (2021), for TRAPPIST-1. NASA planetary fact sheets for radii, masses and orbital speeds. NASA's Roman Space Telescope launched on 30 August 2026; ESA lists PLATO for launch in early 2027. Modelled, not measured: the transit chance assumes circular orbits pointed in random directions; the transit length assumes the planet crosses the centre of the star; the 3.3 percent floor assumes every Kepler star is Sun-like and every candidate orbits at Earth's distance. All arithmetic was checked in code.