Retrograde duration depends on the planet and the individual cycle. In our 100-year window, Mercury’s median complete retrograde lasts 22.48 days, Venus’s 42.14 days and Mars’s 76.47 days. Slower outer bodies remain retrograde for several months.
The table measures elapsed time from the retrograde station to the direct station. It excludes pre-shadow and post-shadow. For a particular year, use the retrograde calendar; a median duration should not be used to predict an end date.
Minimum, median and maximum durations
Analysis window: 1951-01-01 00:00 UTC inclusive to 2051-01-01 00:00 UTC exclusive. We retain only cycles whose start and end both fall within the window. A retrograde crossing an outer boundary is excluded rather than shortened. The dates below identify the starts of the shortest and longest retained cycles.
| Body | Cycles | Shortest days | Median days | Longest days | Shortest starts (UTC) | Longest starts (UTC) |
|---|---|---|---|---|---|---|
| Mercury | 315 | 19.76 | 22.48 | 24.22 | 1958-11-30 07:15 UTC | 1993-07-01 15:29 UTC |
| Venus | 63 | 40.89 | 42.14 | 43.13 | 2045-12-12 01:32 UTC | 1956-05-31 18:04 UTC |
| Mars | 47 | 60.01 | 76.47 | 81.12 | 2003-07-29 07:38 UTC | 2012-01-24 00:56 UTC |
| Jupiter | 91 | 117.51 | 120.61 | 123.57 | 1952-09-09 19:37 UTC | 1994-02-28 13:47 UTC |
| Saturn | 97 | 133.65 | 138.38 | 141.40 | 1973-10-17 05:39 UTC | 1959-04-16 15:28 UTC |
| Uranus | 98 | 149.04 | 151.06 | 154.32 | 2039-11-08 20:57 UTC | 1999-05-21 22:23 UTC |
| Neptune | 99 | 156.91 | 159.47 | 160.34 | 2049-08-29 03:08 UTC | 1985-04-05 01:39 UTC |
| Pluto | 99 | 156.15 | 158.95 | 164.60 | 1981-01-26 12:51 UTC | 2049-06-08 19:10 UTC |
The JSON includes the ending station for every listed extreme. The median is the central sorted duration, or the average of the two central values for an even count. Each complete cycle receives equal weight; this is not a time-weighted statistic or a percentage of people born retrograde.
Why different planets have different durations
Retrograde is apparent backward movement in geocentric longitude. It comes from the changing viewing geometry between Earth and a planet, rather than the planet reversing its orbit. Orbital speeds and geometry differ by body. Eccentric orbits and changing geometry also make cycles of the same planet unequal.
Consequently, “Mercury lasts three weeks” or “Mars lasts ten weeks” is a rough description. The distribution in this table gives a more useful range for this defined window. These observed minima and maxima are not all-time physical limits.
Dates touched are different from elapsed days
A cycle starting late on one UTC date and ending early on another can touch more calendar dates than its elapsed duration suggests. We subtract timestamps and divide by 86,400,000 milliseconds. We do not count inclusive date labels, round each cycle to whole days first, or include days in shadow.
Local calendars may show a station on a different date from UTC. That changes the displayed date, not the duration between the two instants. Check the time-zone label before comparing calendar entries.
Sources, scope and reproducibility
The source is the site’s saved station table, generated 2026-10-25 using Astronomy Engine and numerical speed searches. The comparison script recomputes elapsed durations from timestamps rather than using the lookup module’s rounded day count. Eight astrological bodies are included from Mercury through Pluto; the Sun, Moon, nodes, Lilith and Chiron are excluded.
The window spans 100 calendar years, but retained cycle counts vary because orbital cycles have different lengths and boundary-crossing cycles are omitted. We make no claim that every source station has been independently validated. Reporting to minutes or hundredths of a day is a formatting choice, not a proven accuracy bound.
For another use of the underlying intervals, see how many planets can be retrograde at once. For Venus recurrence rather than duration, see the eight-year comparison. None of these motion statistics establishes effects on people.
Download the unrounded comparison data (JSON). To reproduce it, run node scripts/build-comparison-research.ts from the project repository. The script uses the same calculation modules and saved station table as our tools. This is reproducibility within one implementation, not an independent accuracy check.
Cite this comparison
Natal Chart Lab (2026). How Long Do Planets Stay Retrograde? 100 Years Compared. Article and worked examples; supporting data. Calculated comparison output is available under CC BY 4.0: credit Natal Chart Lab, link to this article and the license, and indicate changes.
Prepared with AI assistance and checked against calculation code and saved examples. Astronomical calculations do not establish astrological claims about personality, relationships or future events.