Two independent detection methods run side by side here. The transit method (top graph) watches the star's brightness: whenever the planet's disc crosses in front of the star along our line of sight, it blocks a small, precisely periodic fraction of the starlight, producing the classic dip-and-recover light curve. The radial-velocity method (bottom graph) doesn't need a transit at all — the planet's gravity tugs the star into a small orbit of its own around the system's common centre of mass, and the star's velocity toward and away from us oscillates in a Doppler sine wave with the same period as the planet.
transit depth ≈ (R_planet / R_star)²
RV amplitude K ∝ M_planet · sin(i) / (M_star^(2/3) · P^(1/3))
- Planet radius — sets the transit depth. A bigger planet blocks more starlight; depth scales with the square of the radius ratio.
- Planet mass — sets how hard the star gets tugged; heavier planets produce a bigger radial-velocity wobble (the star's motion is exaggerated here so it's visible — in reality a star moves only metres per second).
- Orbital distance — farther orbits take longer to complete (Kepler's third law) and produce a weaker RV pull, but the transit depth is unaffected since it depends only on the size ratio.
- Inclination — at 90° the orbit is edge-on and we see transits; dial it down toward 0° (face-on) and the planet's path swings out of our line of sight, so the transit dip vanishes even though the RV wobble keeps going (radial velocity still works at lower inclination, just with a smaller line-of-sight component).
Real detections combine both: a transit alone gives the planet's radius, an RV curve alone gives a minimum mass, and having both together — as for thousands of confirmed exoplanets — pins down the planet's true mass, radius, and therefore its density and likely composition.