How accurate are
these times?
Every station and cazimi on this site is computed from NASA JPL's DE421 ephemeris and rounded to the minute in UTC. On November 7, 2026 we checked five of those moments against JPL Horizons, NASA's own ephemeris service, minute by minute. All five agreed.
Cazimi against JPL Horizons
Horizons was asked for Mercury's apparent geocentric ecliptic longitude at one-minute steps around each moment (observer table, quantity 31, centre 500@399). A station is the minute where the longitude stops changing direction; because Mercury moves only a few ten-thousandths of a degree per minute near a station, Horizons sometimes shows a three-minute plateau, and we take its centre. The cazimi is the minute Mercury's longitude equals the Sun's.
| Moment | Cazimi (DE421) | JPL Horizons | Difference, minutes |
|---|---|---|---|
| Station retrograde, October 2026 | 07:12–07:14 plateau, centre 07:13 | 0 | |
| Station direct, November 2026 | 15:54, single minimum | 0 | |
| Station retrograde, February 2027 | 17:35–17:37 plateau, centre 17:36 | 0 | |
| Cazimi, February 2027 | Mercury passes the Sun's longitude at 16:39 | 0 | |
| Station direct, March 2027 | 12:32, single minimum | 0 |
Horizons prints longitudes to 10⁻⁷ degree, so a plateau of equal values spans the minute of the station and its neighbours; the centre of the plateau is the station to within ±1 minute, which is also the precision we claim.
The same retrograde on other sites
For the February 2027 retrograde (station retrograde February 9, 2027 at 17:36 UTC, station direct March 3, 2027 at 12:32 UTC) this is what other widely read pages printed on November 7, 2026. None of them is wrong for its own time zone; the differences show why a date without a time and a zone can be a day out for a reader elsewhere.
| Site | What it prints | Precision | Against Horizons |
|---|---|---|---|
| The Old Farmer's Almanac | February 9 – March 3, 2027 | Dates only, Eastern Time | Same dates; the 17:36 UTC station is 12:36 EST, so the day agrees in both zones. |
| Farmers' Almanac | February 9 – March 3, 2027 | Dates only, Eastern Time | Same dates. |
| Cafe Astrology | February 9, 2027, 12:36 pm EST at 5° Pisces 59′ | Time for the retrograde station only | 12:36 EST is 17:36 UTC, the same minute and the same degree. |
| Natal Echo | February 9, 17:35 UTC; March 3, 12:34 UTC | Times in UTC | 1 minute early and 2 minutes late against Horizons; a different ephemeris or rounding. |
| Sterngucker.de | 9 Feb 18:36 MEZ; 3 Mar 13:32 MEZ; also 10 Jun 20:15, 4 Jul 21:39, 7 Oct 16:37, 28 Oct 16:11 MESZ | Times in Central European Time | All six 2027 stations agree with ours to the minute once converted to UTC. |
| Star Walk | February 9, 2027 | Date only, no time zone | Agrees on the day. |
| Zodiac Astrology Horoscopes | 10 February – 3 March 2027 | Dates only, no time zone | Starts a day late for every zone west of UTC+6:24; correct only east of it. |
| Jessica Adams | February 1–18 in Pisces, then Aquarius; "allow ±24 hours" | Sign periods with a one-day tolerance | The tolerance is what a minute-precise time removes. |
| MercurioRetrogrado.com | Pre-shadow from 26 January; post-shadow until 17 March | Shadow dates, no times | Our pre-shadow begins 25 January 18:58 UTC and post-shadow ends 23 March 14:58 UTC; the shadow definition differs (see below). |
Why dates differ by a day
A station is an instant, and the calendar date of an instant depends on where you are. The February 9, 2027 station at 17:36 UTC is 12:36 in New York, 18:36 in Berlin and 04:36 on February 10 in Sydney. Sites that print a date in Eastern Time are right for North America and a day early for Australia; sites that print the Sydney date are a day late for everyone in the Americas. This is why every table on Cazimi gives the UTC minute and converts it to your zone in the browser.
Why shadow dates differ by a week
The two stations are unambiguous. The shadow boundaries are a convention: pre-shadow begins when Mercury first reaches the degree where it will later station direct, and post-shadow ends when Mercury returns to the degree where it stationed retrograde. Some sites use whole degrees, some use a fixed number of days, some use the sign boundaries. Cazimi computes both boundaries from the exact station longitudes, to the minute, which is why our post-shadow for the February 2027 cycle ends on March 23 and not mid-March. What the shadow period is.
Where the live position comes from
The "Mercury right now" block on the guides and the today page uses astronomy-engine, a VSOP87-based library accurate to about one arcminute for the planets, recomputed every day at 00:05 UTC. The retrograde tables themselves never come from it: they are DE421 values computed once with Skyfield and published as open JSON with a DOI. The two agree on Mercury's retrograde flag and on its position to the arcminute.
Check it yourself
Open JPL Horizons, choose Mercury (199), observer location Geocentric (500), a one-minute step across the station, and quantity 31 (observer ecliptic longitude and latitude). The minute where the longitude reverses is the station. Our values for every cycle from 1900 to 2050 are at /v1/mercury/cycles-1900-2050.json.
Questions
How accurate are Cazimi's Mercury retrograde times?
To the minute. Five station and cazimi moments between October 2026 and March 2027 were checked against NASA JPL Horizons on November 7, 2026 and all five agreed, including the February 9, 2027 station at 17:36 UTC.
Why do different websites give different Mercury retrograde dates?
Most sites print dates without times, in their own time zone. The station on February 9, 2027 at 17:36 UTC is still February 9 in New York but already February 10 in Sydney, and shadow periods have no single agreed definition. Cazimi prints the minute in UTC and shows it in your zone.
What ephemeris does Cazimi use?
NASA JPL's DE421 planetary ephemeris, read with Skyfield; positions are apparent geocentric ecliptic longitudes of date. The live "Mercury right now" block uses astronomy-engine (VSOP87, about 1′) and is checked against the same DE421 tables; on November 7, 2026 both put Mercury at 8°37′ Scorpio at 12:00 UTC.