The rule of thumb everyone repeats. We measured 116 real oppositions with our own engine and the gaps run from 766 to 804 days — a spread of more than five weeks.
You will read it everywhere, including from people who should know better: Earth and Mars line up every 26 months, so launch windows come round on a two-year-and-a-bit clock.
It is a useful approximation and it is not a fact.
We ran our own engine across 1800 to 2050 — the full range JPL's planetary element set is stated to be valid for — and found every opposition in it: the moments when the Sun, Earth and Mars line up. That is 116 oppositions.
The gaps between them run from 766 days to 804 days.
That is a spread of 38 days. More than five weeks of drift between the shortest cycle and the longest. The mean is 780 days; "26 months" works out to about 791, which lands inside the range and is only occasionally right.
Mars has an eccentric orbit — noticeably egg-shaped compared to Earth's. The two planets are not running on fixed gears, so the interval between alignments depends on where in its orbit Mars happens to be when the catch-up happens. Near perihelion it moves faster and the gap shortens. Near aphelion it dawdles and the gap stretches.
The consequence is practical: a launch calendar built on a fixed 26-month rule drifts. Not enough to notice across one cycle. Very much enough across four.
We could have hard-coded a table of window dates. It would have been less work and nobody would have known.
The reason we did not is the same reason the deck computes everything else rather than fetching it: a number you cannot check is a number you have to trust, and we would rather you did not have to. Open the deck, move the clock, and watch the gaps come out uneven yourself.
Mars Deck computes every position on this page from orbital mechanics, live in your browser. Free, no ads, no account.
Open the deck →