A free-floating ("rogue") planet has no host star, so it can't be found by transits or radial velocity. A young one, however, is still glowing in infrared from the heat of its own formation — this is Kelvin–Helmholtz contraction: gravitational potential energy released as the planet slowly shrinks is radiated away as heat, and that glow fades as the planet ages and cools.
log10(L / L_sun) ≈ -3.6 + 1.6·log10(M / M_Jup) - 0.9·log10(t / 10 Myr)
Stefan–Boltzmann: T_eff = ( L / (4π R² σ) )^(1/4), R ≈ 1 R_Jup
Absolute bolometric mag: M_bol = 4.83 - 2.5·log10(L / L_sun)
Apparent magnitude: m = M_bol + 5·log10(d / 10 pc)
This cooling-track scaling is a simplified fit in the spirit of gravitational-contraction models (Burrows et al. 1997) for isolated giant planets and brown dwarfs — real evolutionary tracks depend on atmosphere and initial entropy too, but the trend is the same: heavier and younger objects are brighter, and luminosity collapses fast with age.
- Mass — sets how much gravitational energy is available to radiate; more massive objects stay hot longer.
- Age — luminosity falls roughly as a power law in time as the planet contracts and cools.
- Distance — apparent brightness dims with the inverse-square law, so the same planet becomes undetectable if placed far enough away.
- Survey telescope — each real infrared survey has a limiting magnitude; if the planet's apparent magnitude is fainter (numerically larger) than that limit, it cannot be detected.
This is one of two real ways astronomers find rogue planets — the other is gravitational microlensing, covered in this site's separate microlensing light-curve simulator. Direct infrared imaging only works for young, still-warm free-floating planets; older ones (like a hypothetical rogue Jupiter at billions of years) are far too cold and faint for any current telescope to image directly, and must instead be found by microlensing.