The habitable band trails each sunset across the planet. Nothing here is a loop; the sun, the heat, and the band all come from the real orbit, computed live.
Running the planet through four orbits so every strip of ground carries an honest thermal history before you arrive. You land twelve days short of perihelion, on the one meridian where the next sunset takes two tries: the sun will half-set, hang on the horizon for days, climb back out, and only then commit.
Mercury sits deep in the sun's gravity well, which produces the least intuitive fact in planetary travel: one of the closest planets is the hardest to visit. Earth circles the sun at 29.8 km/s and a spacecraft inherits all of it. To spiral inward it must throw that speed away, and as it falls the sun pays it back with interest, so it arrives at Mercury's distance moving far too fast for the little planet to catch. Stopping there by brute force costs more total velocity change than leaving the Solar System entirely. Measured in delta-v, the currency that actually matters, Mercury is farther away than Pluto.
Rockets pay for velocity exponentially: every extra kilometer per second multiplies the fuel required, and that fuel must itself be accelerated. A direct Mercury orbiter would be almost entirely propellant with a camera taped to the top. So nobody flies direct. Drag the slider and watch the teal sliver of everything-that-is-not-fuel disappear.
The workaround is to let planets do the braking. Sweep past Venus or Mercury on the right side and the encounter trades a whisper of the planet's orbital momentum for a useful shove on the spacecraft. Mariner 10 did it first in 1974, using Venus to reach Mercury, following an insight from the Italian scientist Giuseppe Colombo: shape the flyby so the new orbit is in resonance with Mercury's and the spacecraft keeps meeting the planet again and again. Colombo is the same man who first explained the 3:2 spin lock this entire site runs on, and the mission arriving now carries his nickname, BepiColombo.
MESSENGER launched in 2004 and needed 6.6 years, fifteen loops around the sun, and six flybys (Earth once, Venus twice, Mercury three times) before it could finally stop in March 2011 as the first spacecraft ever to orbit Mercury. BepiColombo, launched in 2018, raised the count to nine flybys and added continuous ion-engine braking, a thrust about the weight of a few coins applied for months at a time; after six Mercury flybys it arrives for orbit insertion this November. For scale, New Horizons flew to Pluto, more than eighty times Mercury's distance from us at its farthest, in a comparable flight time, because flying past something is cheap and stopping is what costs.
Near Mercury, sunlight runs up to eleven times Earth intensity, and the dayside ground below glows at up to 700 K, so an orbiter is roasted from both directions at once. MESSENGER hid behind a ceramic-cloth sunshade whose front face ran near 350 °C while the instruments a few centimeters behind it sat at room temperature, and it flew a deliberately lopsided 12-hour orbit, dashing low over the surface briefly and then retreating thousands of kilometers to radiate the heat away. Even the orbit itself is under attack: this deep in the well the sun's gravity constantly warps any path around Mercury, and when MESSENGER's correction fuel ran out in 2015, solar tides drove it into the ground it had been mapping.
Mercury has no atmosphere worth mentioning, so there are no parachutes and no aerobraking. Every meter per second must be killed by rocket, Apollo-style, at the end of the most expensive journey in the inner system, and that is why the number of landers in history is zero. The concept NASA has studied most seriously solves the heat problem with scheduling instead of shielding: arrive at high latitude and touch down just after local sunset, into the mild trailing band, then work through the long night before the dawn ends the mission. The plan for the first Mercury landing, in other words, is to aim for exactly the strip of ground this site lets you stand on. The Goldilocks band is not a curiosity. It is the runway.
The orbit is a real Kepler solution: semi-major axis 0.387 AU, eccentricity 0.2056, period 87.969 days, solved each frame for true anomaly and distance. The planet spins prograde with a 58.646 day sidereal period, the locked 3:2 resonance, and obliquity is taken as zero. The subsolar point, terminator, retrograde perihelion loop, and 176 day solar day all fall out of those numbers; none of the motion is scripted.
Every cell of the surface (a 192 by 96 grid) integrates a one-slab energy balance: absorbed sunlight in (albedo 0.09, falling off with the true square of solar distance), thermal radiation out (emissivity 0.95), a small interior term, and an effective heat capacity of about 1.2 MJ per square meter per kelvin chosen to match Mercury's real regolith cooling. The model reproduces the published landmarks: roughly 700 K noon at the hot poles at perihelion, about 570 K at the warm poles, deep-night ground near 110 K and still falling toward the real 95 K floor, and a comfortable band a few hours of walking wide that drifts west at 2πR·cos(latitude) per 175.94 days, which is 3.6 km/h at the equator. One detail the model taught us while being built: the perihelion loop swings the subsolar point only about half a degree of longitude each way, so the pure double sunset lives on a razor-edge meridian 90° from a hot pole. What rescues the spectacle is the sun itself, 1.7° wide at perihelion: during the two-week hover its upper limb never fully drops below the horizon.
Simplified away: topography and crater shadows (so no polar cold traps here), subsurface conduction layering, thermal re-radiation between slopes, and Mercury's 2 arcminute obliquity. The sky panorama draws the sun at six times its true angular size by default because at true scale (1.1° to 1.7° across) it is a few pixels; the toggle shows honest scale. Comfort text assumes the counterpressure-suit scenario from our conversation: skin-temperature radiative balance is real physics, the breathing is borrowed.
The stars are the real sky: 1,627 stars to magnitude 5 from the Yale/HIPPARCOS catalog (via the d3-celestial dataset), rotated into Mercury's orbital frame using the planet's J2000 elements (i 7.005°, Ω 48.331°, ϖ 77.458°) and wheeling at the true 58.65 day sidereal rate. Stellar parallax between Mercury and Earth is far below one pixel, so these are exactly your constellations; only the sun's slow crawl through them is different. The surface is the actual MESSENGER MDIS global photomosaic (NASA / JHU-APL / Carnegie, via the CC-BY Solar System Scope texture set), verified in place by matching the ray craters Kuiper and Xiao Zhao to their published coordinates, which also puts Caloris beside the 180° hot pole where it belongs.
Chase mode holds your longitude offset from the subsolar point, so the sun freezes in your sky while the planet turns underneath. The walker's cadence comes from the physically required ground speed (about 0.72 m per stride) and runs on the wall clock, never the time-warp: crank the simulation to 8 days per second and the figure still strolls at human pace, because that is the pace the real walk would take. Near perihelion the required velocity passes through zero and briefly reverses; the walker stops, turns around, and shuffles east for a few days while the sky does the work.
The pace is a latitude tax, not a constant: holding any sun-fixed spot costs 2πR·cos(latitude) per solar day, so the equator's 3.6 km/h average (peaking near 6 km/h at aphelion) falls to 0.63 km/h at 80°, 0.32 at 85°, 0.19 at 87°. The catch is that the band thins toward the pole: at aphelion the noon sun at 85.9° just barely warms the ground to 295 K, so that is the highest latitude where the perch works year-round. The hold-temperature chase mode tracks the 295 K isotherm itself instead of the sun, which at high latitude means shuffling around the pole once per 176 day solar day under a sun that never leaves the horizon.
One honest wrinkle the model surfaced while being tested: at the equator the thermostat cannot be perfect. Around each perihelion the hovering sun flattens the thermal gradient behind the stalled terminator, and the 295 K isotherm briefly sprints at 10 to 30 km/h, faster than anyone walks, so a rate-limited walker rides a swing of roughly minus 20 to plus 60 °C through that window. At the 85° perch the required pace never exceeds about 1.7 km/h and the ground holds 18 to 29 °C orbit after orbit. High latitude is not just lazier; it is the only place where holding the perfect temperature is literally possible on foot.
Landmark values are consistent with MESSENGER-era thermal modeling of Mercury. Everything else is computed in front of you.
Credits and licenses: surface photomosaic NASA / JHU-APL / Carnegie Institution (MESSENGER MDIS), via the Solar System Scope texture set, CC BY 4.0. Star positions from the d3-celestial dataset (BSD-3), built on the Yale Bright Star Catalogue and HIPPARCOS. Rendering by three.js (MIT). Physics, design, and code built in conversation.