Why Your Local Calendar Might Be Wrong: The Triple-Engine Hijri Calendar

If you have ever found yourself asking, "Wait, is today the 1st or the 2nd of the month?" you are not alone. The Islamic (Hijri) calendar is notorious for local variations, often leading to neighbouring countries, or even different mosques in the same city, starting Ramadan or celebrating Eid on different days.

This happens because there is no single, universally agreed mathematical definition for the Islamic calendar. Some regions rely strictly on local naked-eye observation, some follow a neighbouring country, and others rely on a fixed astronomical or arithmetic rule. Each approach answers a slightly different question, and so each can produce a slightly different date.

To make this disagreement transparent rather than confusing, Hilal Vision built the Triple-Engine Hijri Calendar: a system that runs three distinct calendar algorithms side by side. Below we explain how each engine works, why they disagree, and how to decide which one you actually need.

Engine 1: The Astronomical (Conjunction) Calendar

The purely astronomical calendar is based on celestial mechanics. A new lunar month is tied to the moment the Sun, Earth and Moon line up in ecliptic longitude, an event called the conjunction (often loosely called the "new moon"). It is crucial to be precise here: the conjunction is the invisible instant the Moon passes between the Earth and the Sun. It is not the hilal, the first visible sliver, which only appears 15 to 40 or more hours later.

• How it works: If the conjunction occurs before a specified reference time (often midnight or sunset at a chosen reference longitude), the new month begins the next day. Modern engines locate the conjunction to within seconds using high-precision solar and lunar theories.
• The catch: At the moment of conjunction the Moon is entirely invisible, its illuminated face points away from Earth. A calendar that starts the month at conjunction therefore frequently runs one full day ahead of any calendar based on actually seeing the crescent.
• Best for: Scientific precision and long-range planning, where reproducible celestial mechanics matter more than human observation.

Engine 2: The Umm al-Qura Calendar

The Umm al-Qura calendar is the official civil calendar of Saudi Arabia and is widely followed by Islamic communities worldwide for civil and administrative purposes.

• How it works: It uses the coordinates of the Kaaba in Mecca as its reference point, and applies two strictly mathematical conditions on the 29th of the current month. In its current form, adopted from 1423 AH (2002 CE), the rule is: the geocentric conjunction must occur before sunset in Mecca, and the Moon must set after the Sun in Mecca. If both hold, the new month begins the next day; otherwise the current month runs to 30 days.
• The catch: It is highly organised and predictable, but it does not guarantee that the crescent will actually be visible to the naked eye. The geocentric "moonset after sunset" test is far weaker than a true visibility criterion. In some months the Umm al-Qura calendar declares a new month even though detailed modelling places the crescent in Zone E (not visible even with a telescope) or Zone F (below the Danjon limit of roughly 7 degrees of elongation, not visible by any means whatsoever). Zone F does not mean "the Moon is below the horizon"; it means the crescent is too geometrically thin to be perceived even with the most powerful optical instruments.
• Best for: Civil organisation, flight bookings and aligning with Saudi Arabia's official dates. For a complete breakdown of this rule, its history and the controversies it generates, see the Umm al-Qura calendar explained.

Engine 3: The Tabular (Kuwaiti) Calendar

The Tabular Hijri calendar is an arithmetic approximation. Instead of tracking the actual position of the Moon, it relies on a fixed mathematical scheme refined by early Islamic astronomers.

• How it works: It assumes a mean lunar month that closely tracks the true synodic month of 29.530589 days, but the arithmetic cycle of 10631 days over 360 months yields an exact mean of 29.530556 days, fractionally short of the true synodic period. To keep dates whole, it alternates months between 30 and 29 days. To absorb the leftover fraction, it inserts a leap day 11 times in every 30-year cycle. Microsoft Windows and many digital systems use this scheme, commonly called the Kuwaiti algorithm.
• The catch: Because it ignores the Moon's actual elliptical orbit (which speeds the Moon up near perigee and slows it down near apogee), individual months can drift by one, or occasionally two, days relative to either a computed or an observed calendar.
• Best for: Digital systems and software where fast, deterministic dates are required years in advance.

The Three Engines Side by Side

It helps to see the engines compared directly. Note that none of these three computed engines, on its own, certifies that a human could actually see the crescent that evening; that is a separate question handled by visibility models such as Yallop and Odeh.

Why the Calendars Disagree

The three engines disagree because they are answering different questions. The astronomical and Umm al-Qura engines ask where the Moon is; the tabular engine asks what the long-run average says; and an observation-based calendar asks the much harder question, can the crescent be seen tonight from here? That last question is governed by geometry the simpler rules never evaluate.

Calculation is not the same as visibility

A month declared on paper can still leave an invisible crescent in the sky. Whether the young Moon can actually be detected depends on a cluster of geometric quantities, not on the calendar rule:

• ARCV (arc of vision): the altitude difference between the Moon and the Sun at sunset. A larger ARCV lifts the crescent into a darker sky.
• ARCL (arc of light): the Moon to Sun elongation. This drives the crescent's width and underpins the Danjon limit; below roughly 7 degrees of elongation the crescent simply cannot be perceived.
• W: the topocentric crescent width in arcminutes, the physical thickness of the lit sliver.
• DAZ: the difference in azimuth between Sun and Moon, which determines how far the crescent sits from the Sun's glare along the horizon.

A subtle but important effect here is topocentric parallax. The Moon's horizontal parallax of about 57 arcminutes lowers its observed (topocentric) altitude relative to the geocentric value by up to about 1 degree near the horizon. A purely geocentric "moonset after sunset" test, like the one Umm al-Qura uses, can therefore report a positive lag while a real observer on the ground sees the Moon already gone. This is one mechanism by which a mathematically declared month can correspond to an unobservable crescent. The deeper geometry of why the same crescent is visible in one country and invisible in another is explored in why a crescent is visible in some countries and not others and in why altitude and terrain matter.

The Hidden Fourth Question: Which Visibility Criterion Applies?

None of the three engines above answers the hardest question: regardless of what the calendar says, can the crescent actually be seen from here tonight? That question requires a separate layer of calculation entirely. The leading models translate the geometry into a single score. Bernard Yallop's 1997 method produces the dimensionless q-value, classifying the result into six mutually exclusive bands. These are not a cumulative "greater-than" ladder; each is a distinct range:

The q-value is computed as:

q = (ARCV - (11.8371 - 6.3226·W + 0.7319·W² - 0.1018·W³)) / 10

where W is the topocentric crescent width in arcminutes, evaluated at Yallop's "best time", about four ninths of the lag time after sunset (the lag time being the interval between sunset and moonset).

Mohammad Odeh's 2004 criterion refines this using 737 observation records, roughly half of them drawn from ICOP, to produce the V-value:

• V ≥ 5.65: crescent visible to the naked eye.
• 2 ≤ V < 5.65: visible with optical aid, and may then be seen with the naked eye.
• -0.96 ≤ V < 2: visible only with optical aid.
• V < -0.96: not visible even with optical aid.

Odeh's empirical optical-aid limit sits at about 6.4 degrees of elongation. The fuller treatment of these criteria lives in the science behind the crescent and the moon sighting versus calculation debate.

Beware the age myth

It is tempting to think a calendar can simply wait until the Moon is "old enough" to be seen. It cannot. Moon age, the number of hours since conjunction, is a poor predictor of visibility and is not a parameter in either the Yallop or the Odeh model. ARCV, ARCL, W and DAZ are what actually decide the matter. Record young crescents make the point vividly: Thierry Legault imaged a crescent in 2013 at essentially the instant of conjunction (an age near zero hours) at an elongation of about 4.4 degrees, while the record binocular sighting by Mohsen Mirsaeed in 2002 came at about 11 hours and 40 minutes. Two crescents of nearly the same "age" can have completely different visibility depending on their geometry.

The Global Versus Local Problem

Even if everyone agreed to compute visibility rather than rely on the eye alone, a second disagreement would remain: visibility is local. A crescent that is comfortably in Zone A over the Atlantic at dusk may be in Zone E from South-East Asia, where the Sun set hours earlier with the Moon barely separated from it. So whose sunset should define the start of the month?

This is the heart of the modern jurisprudential and astronomical debate, and it produces three broad doctrines:

• Wujud al-Hilal ("existence of the crescent"): the rule behind the Umm al-Qura approach. It is enough that, geometrically, the Moon sets after the Sun somewhere (canonically at Mecca). It does not require that the crescent be perceptible, which is precisely why it can declare a month while the crescent is still below the Danjon limit.
• Imkan al-Ru'ya ("possibility of sighting"): the month begins only when a genuine visibility criterion (such as Yallop's or Odeh's) is satisfied at some reference location, so that the crescent could actually have been seen. This is the position adopted, for example, by Turkey's Diyanet.
• A unified global Hijri calendar: rather than re-deciding the date region by region every month, this approach fixes a single worldwide rule in advance so that the whole Muslim world shares one date. The best-known modern proposal is the Universal Hijri Calendar developed by the Moroccan scholar Jamaluddin Abd al-Raziq, which uses a fixed reference meridian and a calculated visibility condition to pre-compute a single global calendar. Variants of this thinking were the focus of the 2016 Istanbul international congress on the unification of the Hijri calendar, convened under the Diyanet, which recommended a single global calendar based on calculated crescent visibility, although the recommendation has not been universally adopted.

The tension is fundamental. Local sighting (ikhtilaf al-matali') honours the literal geometry of the sky over each place but fragments the calendar. Global unification (ittihad al-matali') delivers one shared date but must accept that, on the night it begins, the crescent was genuinely invisible from large parts of the world. There is no purely astronomical way to dissolve this; it is a choice about which value, local fidelity or global unity, to prioritise. The wider history of how communities have wrestled with this is traced in the history of crescent observation, and the mechanics of the calendar itself in how the Islamic lunar calendar works.

Bringing It All Together

When you look at the Hijri date on the Hilal Vision dashboard, you are not seeing one arbitrary number. The platform computes all three engines at once, so you can switch contexts instantly: align with your local mosque, track the official Umm al-Qura date for travel, or read the strict astronomical reality. You can compare them against an actual visibility prediction for your own coordinates on the global visibility map, inspect the geometry on the moon dashboard, browse confirmed historical sightings in the ICOP archive, and read the full computation rules on our methodology page. To run the numbers for your exact location tonight, open the Hijri calendar on moonsighting.live and let the three engines, plus the Yallop and Odeh criteria, do the work. Pro members additionally get live cloud-cover overlays and an extended ICOP archive for deeper historical research.

Understanding why these engines disagree turns an annual source of confusion into something you can reason about. A calendar is only ever a model of the sky, and knowing which model you are reading, and what question it answers, is the first step to never being caught off guard by the date again.

Clear skies and happy sighting.

References and Further Reading

• Yallop, B.D. (1997). A Method for Predicting the First Sighting of the New Crescent Moon. HM Nautical Almanac Office, NAO Technical Note No. 69.
• Odeh, M.Sh. (2004). "New Criterion for Lunar Crescent Visibility." Experimental Astronomy. (Odeh founded ICOP in 1998; the criterion is derived from 737 observation records.)
• Danjon, A. (1932, 1936). L'Astronomie. (First report and quantification of the Danjon limit.)
• Fatoohi, L.J., Stephenson, F.R. & Al-Dargazelli, S.S. (1998). "The Danjon limit of first visibility of the lunar crescent." The Observatory.
• Bennett, G.G. (1982); Saemundsson (atmospheric refraction); Kasten, F. & Young, A.T. (1989) (air-mass).
• The Islamic Crescents' Observation Project (ICOP) and the International Astronomical Center: https://www.astronomycenter.net.