Where each method stands
Every figure in the table and in the paragraphs below is generated from
certification.json (node scripts/readme.mjs),
except inside the blocks marked Research note: those are a dated record
of the research, kept so nothing is lost, and their counts are not part of
the certification.
The gate: 100 % or not at all
A method may be used only once it reproduces every time of its
authority’s timetable on the printed minute — not one minute off, anywhere.
npm run certify checks each method day by day and writes
certification.json; timesFor throws UncertifiedMethodError for any
method not certified there, and says status: 'method-only' for a date
beyond the span it was checked over. candidateTimes runs a method
regardless, for research and for the check itself — not for showing anyone
their prayer times.
Days the timetable itself gets wrong (another town’s times under a name, a previous year’s figures for the date, a figure that differs from the authority’s own) are set aside and listed in the record, so each can be judged.
The goal is 100 % on the printed minute, for every time of a full year, for each method.
| Authority | Tables matched | Places | Times checked | Off the printed minute | By the method alone | Correction entries | Checked over |
|---|---|---|---|---|---|---|---|
| Sweden | Islamiska Förbundet’s tables | 107 | 65,442 | 0 | 99.969 % | 270 | 11 Oct 2026 → 21 Dec 2027 |
| Diyanet’s own tables | 159 | 1,393,794 | 0 | 99.981 % | 1 Jan 2024 → 31 Dec 2027 | ||
| Diyanet | its tables worldwide | 8,318 | 71,630,078 | 0 | 99.987 % | 9,101 | 1 Jan 2024 → 31 Dec 2027 |
| Morocco | Habous | 164 | 94,350 | 0 | 99.885 % | 108 | 27 Aug 2026 → 30 Nov 2026 |
| Algeria | Ministry of Religious Affairs | 68 | 101,184 | 0 | 99.365 % | 642 | 1 Oct 2026 → 5 Jun 2027 |
| Indonesia | Kemenag | 517 | 362,934 | 0 | 99.935 % | 234 | 7 Oct 2026 → 31 Jan 2027 |
| Türkiye Takvimi | its cities’ years | 821 | 1,812,768 | 0 | 99.952 % | 860 | 31 Dec 2025 → 2 Jan 2027 |
| Kazakhstan | the Spiritual Administration’s times | 84 | 551,880 | 0 | 100.000 % | 0 | 1 Jan 2025 → 31 Dec 2027 |
| Egypt | the Survey Authority’s times | 77 | 168,630 | 0 | 99.373 % | 1,056 | 1 Jan 2026 → 31 Dec 2026 |
| Brunei | the Ministry’s times | 4 | 17,064 | 0 | 99.671 % | 56 | 9 Jan 2025 → 31 Dec 2026 |
| Qatar | the Calendar House’s times | 1 | 1,194 | 0 | 95.393 % | 55 | 16 Jun 2026 → 31 Dec 2026 |
| Kuwait | the Ministry’s table | 46 | 201,480 | 0 | 95.205 % | 9,660 | 1 Jan 2026 → 31 Dec 2027 |
| Jordan | the Ministry’s times | 17 | 37,230 | 0 | 98.380 % | 603 | 1 Jan 2026 → 31 Dec 2026 |
| Tatarstan | the Administration’s timetable | 44 | 96,360 | 0 | 98.071 % | 1,858 | 1 Jan 2026 → 31 Dec 2026 |
| Russia (Central Muftiate) | the Central Muftiate’s timetables | 18 | 39,342 | 0 | 99.877 % | 48 | 1 Jan 2026 → 31 Dec 2026 |
| Albania | the Community’s calendar | 22 | 48,180 | 0 | 98.082 % | 924 | 1 Jan 2026 → 31 Dec 2026 |
| Bulgaria | the Grand Mufti’s Office’s timetable | 48 | 105,120 | 0 | 97.671 % | 2,448 | 1 Jan 2026 → 31 Dec 2026 |
| Pakistan (Banuri Town) | Jamia Banuri Town’s timetable | 100 | 219,000 | 0 | 99.996 % | 8 | 1 Jan 2026 → 31 Dec 2026 |
| United Arab Emirates | the Authority’s times | 41 | 179,580 | 0 | 98.221 % | 3,036 | 1 Jan 2026 → 31 Dec 2027 |
| Iran (Tehran Institute of Geophysics) | the Institute of Geophysics’ sheets | 31 | 90,520 | 0 | 99.002 % | 903 | 21 Mar 2025 → 20 Mar 2027 |
| United Kingdom (Wifaqul Ulama) | Wifaqul Ulama’s timetables | 81 | 350,400 | 0 | 98.625 % | 4,641 | 1 Jan 2026 → 31 Dec 2027 |
| Tunisia | the Institute’s times | 26 | 94,855 | 0 | 97.727 % | 2,156 | 1 Jan 2025 → 31 Dec 2026 |
21 of 21 methods certified; 0 of 77,661,385 times off the printed minute. Last checked 17 November 2026. Generated from certification.json by scripts/readme.mjs.
Sweden
Certified over 1 Jan 2024 → 31 Dec 2027 against Diyanet’s own tables for Sweden (159 places), and over 11 Oct 2026 → 21 Dec 2027 against Islamiska Förbundet’s published tables (107 of its towns). Set aside from Islamiska Förbundet’s: 35,852 times that differ from Diyanet’s own table for the day (its next year’s rows repeat this year’s figures), and 710 rows copied from another day. On each, the method gives Diyanet’s own table.
Research note, 2026-10-16 (not generated): Kiruna, printed with Umeå’s times, is not compared; Islamiska Förbundet lists 108 towns.
Diyanet
Certified over 1 Jan 2024 → 31 Dec 2027 for 8,318 places in Türkiye, Northern Cyprus and abroad (Sweden’s are under Sweden). Each place is computed at a point of its own, checked to lie in the place’s country.
Research note, 2026-10-16 (not generated). Of the 8 528 places Diyanet lists, 8 291 were included, with Sweden’s 159. Not yet included:
- 147 places that could not be placed — seven of them Diyanet serves no table for (Grozny, Gudermes, Shali and Urus-Martan, Rožaje, Monte Carlo, Punta Arenas), the rest names the gazetteer does not match;
- 75 whose tables the method does not yet follow over all four years — about half of them in Mexico and Australia, whose tables keep a clock of their own, the rest scattered (Kazan, Fez, Ushuaia, Svalbard among them);
- 12 whose tables Diyanet computes at a point elsewhere — in another country, at sea or in the other hemisphere (Tupiza, Ribeirão das Neves, Montería, Hami, Kungälv and Galena among them);
- Syktyvkar, whose 2024 and 2027 tables print a Fajr hours from its 2025 and 2026 ones;
- Ping Tung City and Rilchingen-Hanweiler, each listed again at the same point (as Pingtung and Kleinblittersdorf) with a table a minute apart on two days.
Research note, 2026-11-11, the places not yet included (not generated). The 237 were gone through again, blind: each table’s latitude from its day length, its longitude and clock from its Dhuhr, then the gazetteer by name and zone, then the 0.001° lattice. 186 more are now included, 61 of them on a clock of their table’s own. Still out:
- 7 Diyanet serves no table for, and 22 that could not be placed;
- 9 the method does not follow: Lugano, Mistelbach, Zemplín, Ushuaia, Syktyvkar, Kløfta, Ukhta and Río Gallegos, and Le Pont-de-Claix, whose Fajr is 2–23 minutes from its table’s between 18 May and 20 June 2026;
- 7 whose table is another place’s than the name says (listed as Uliastay, Amur, Boro Burgaasa, Karatau, Awash, Pokachi and Velikiye Luki);
- 4 listed a second time at the same point as a place already included;
- Brazzaville and Pointe-Noire, which Diyanet lists under no country.
Research note, 2026-11-11 (not generated). Every place with two or more corrections was tried one step of the 0.001° lattice away in each direction (2,086 places). Fifteen sit a step from where their table is computed, and moving them takes 256 corrections off the list: Logoysk 55 → 13 times off by the method alone, Alchevsk 35 → 2, Changchun 31 → 1, Nuuk 28 → 1, Bad Dürrheim 34 → 11, Keshan 26 → 10, Okazaki 15 → 0, Playa del Carmen 12 → 0, Sandgate 12 → 1, Gosnells 12 → 1, Al Fayyum 10 → 0, Ayr 9 → 0, Guaymas 7 → 0, Mérida 6 → 1, Ciudad Delicias 5 → 0. The other 2,071 are where they should be. The fifteen are for Diyanet’s next revision: moving a point changes what the method prints there, and a revision raised for fifteen places would ask every reader of Diyanet’s times to update for nothing they would see.
Morocco
Certified over 27 Aug 2026 → 30 Nov 2026 (164 cities): the method gives 99.885 % alone, and a list of 108 corrections the rest. Certification covers only the dates in the table: each newly published month is checked against the method, and its corrections are added in a new list.
Research note, 2026-11-12 (not generated). Revision 2’s points were fitted on the autumn of 2026, and autumn is the season that says least about a city’s latitude. The ministry’s pages of July and August 2022, in the Internet Archive for nearly every city, are the same program in summer. On those days and the rest since May 2022 revision 2 gave the printed minute for 98.95 % (185 of 17,580 times off). Refitting each city’s point on the 2026 tables and half of those days took the misses on the other half from 185 to 147, so revision 3 is the refit on all of them: 112 points moved, none by more than 0.014° of latitude or 0.01° of horizon, and on every day held the misses fell by half. Only Rabat has every season of this program; winter and spring are still seen through one city.
Algeria
Certified over 1 Oct 2026 → 5 Jun 2027 (68 cities): the method gives 99.365 % alone, and a list of 642 corrections the rest. The method is at its limit: the angles, horizon, Maghrib margin, rounding, sun series, iterations and city points are each sharp optima, and the corrections are independent differences of about a second or less (no pattern by prayer, city, date or season; refitting the points to them only fits noise, and is worse on days held out).
Indonesia
Certified over 7 Oct 2026 → 31 Jan 2027 (517 regencies and cities): the method gives 99.935 % alone, and a list of 234 corrections the rest. Since revision 2 nothing is fitted: each regency is computed at the ministry’s own coordinates, with the sun taken without ΔT as the ministry’s program takes it. The corrections are independent differences at a minute’s edge.
Research note, 2026-11-11 (not generated), on revision 1. Its points were fitted to October–December 2026, and on those months the method alone was a minute off once in 3,200 times; on January 2027, which the fit had not seen, once in 300. The fitted points had taken up a fifth of a second that belongs to the sun (ΔT), which no point can carry through the seasons. A figure “by the method alone” over the months a method’s points were fitted to says less than it seems: Morocco’s and Algeria’s points are fitted too (Morocco’s against every season of 2014–2025; refitting them freely to the 2026 autumn made unseen days three times worse, so they stand).
Revision 1 said: Karawang’s Asr, Maghrib and Isha are computed at a second point, about 2 km north-east of its first (its afternoon fits it, and it holds on days held out). The corrections are independent differences at a minute’s edge, and refitting points to them only fits noise.
Research note, 2026-10-16 (not generated), on the list of 84 then: median 0.2 s from the minute’s edge, nine in ten under 1.3 s, the largest 3.9 s — eight of them at Sibolga, whose mornings no point fits. Certification covers only the dates in the tables: each newly published month is checked against the method, and its corrections are added in a new list.
Research note, 2026-11-11, eleven earlier months (not generated). Revision 2 was run, without its list, against the ministry’s tables for November 2025 → September 2026, which it had never seen: 1,036,068 times, never more than a minute off, and month by month off 0.03–0.14 % of the time, as over the certified months. The exception is two days, 21 and 22 March 2026, when a third of the ministry’s own times are a minute later than on the days around them, in every prayer and across the country. A small yearly wave fitted to what is left halves it on the months it was fitted to and does nothing for the months after, so the method was left as it is.
Türkiye Takvimi
Certified over 31 Dec 2025 → 2 Jan 2027 (821 cities — Türkiye’s province centres and districts): the method gives 99.952 % alone, and a list of 860 corrections the rest. Fajr is the calendar’s Sabâh; its İmsâk is not among the six times the library gives. Each city is at the point and with the temkin its year shows: the calendar’s own files give the point only to the arc-minute.
Research note, 2026-11-11, the districts (not generated). For each district Diyanet lists in Türkiye the calendar’s city of the same name was looked for within 20 km of it: 833 of 869 have one. Each was fitted to its year (latitude within 0.02° of the calendar’s file, then the longitude and the temkin) and taken in when the method alone was off at most six times in the year: 740 were, beside the 81 centres. The 16 left out are all on the northern edge, the Black Sea coast of Kastamonu and Sinop and three districts of Edirne and Kırklareli, off 8–39 times a year; Sinop itself needed a point outside its file’s arc-minute, so the calendar does something there that is not read yet. Districts named alike carry their province in the name.
Research note, 2026-11-11 (not generated). The calendar lists about 120,000 cities; the first list was Türkiye’s 81 province centres, where the method alone is a minute off about once in 3,400 times and never more. Each point and temkin was then searched within 0.03° for the fewest times off (71 → 52). İstanbul keeps ten corrections, all on sunrise and Maghrib in December 2026 with one in March, which no point or temkin within 0.12° removes; no other city has more than three. The rule for the nights that do not reach 19° (north of about 47.5°) is in the method as the calendar’s northern cities show it, and no place of the list holds it to a table. A sample of 40 cities elsewhere in Türkiye and 38 in the rest of the world below 60°, fitted the same way, came out at about the same rate.
Kazakhstan
Certified over 1 Jan 2025 → 31 Dec 2027 (84 cities and towns, each at the coordinates the Administration gives it): the method gives 100.000 % alone, and its correction list has 0 entries. Asr is the Hanafi one, the only Asr the Administration prints; the library’s Hanafi switch leaves it as it is, and there is no table to hold a one-shadow Asr to.
Research note, 2026-11-11 (not generated). The method was read off ten cities’ year first and then run against every place: nothing was fitted per place, and no point was moved. The margin steps from three minutes to five between 47.97° and 48.00° N in the places listed (three at 47.967°, five at 48.001°); 48° is taken as the step. Five more places the Administration lists have no year to check against and are left out.
Egypt
Certified over 1 Jan 2026 → 31 Dec 2026 (77 cities and towns): the method gives 99.373 % alone, and a list of 1,056 corrections the rest. The Authority publishes the current year, so the span goes on a year at a time.
Research note, 2026-11-11 (not generated). The times off by the method alone are spread evenly over the six prayers, Dhuhr included (about 0.6 % each), and each is a single minute. No sun tried in its place — NOAA’s series, the precise ephemeris with or without ΔT, the equation of time at a fixed hour — and no rounding by the second first removes them, so they are taken as the Authority’s own program’s and left to the list. Two things moved the fit most: Asr’s noon shadow from the declination of local noon (about twenty times off a year per place down to two), and the equation of time at each time’s own moment rather than noon’s. The horizon was read as 0.845°, between 0.84° and 0.85°. Each place’s point is fitted to its year; five more names the Authority lists carry no table and are left out.
Brunei
Certified over 9 Jan 2025 → 31 Dec 2026 (4 districts): the method gives 99.671 % alone, and a list of 56 corrections the rest. The Ministry prints one table for the country and the districts as minutes on it, so the four districts are one check counted four times.
Research note, 2026-11-11 (not generated). The Ministry’s published list is entered by hand and carries slips: days with a figure that cannot be read as a time or that puts the day’s times out of order, three dates entered twice, and a few days whose figure on one prayer stands more than two minutes from both the day before and the day after. Those days — eleven over the two years — are left out of the check and are not taught to the correction list: the method is held to the days the list prints soundly. What the sound days show: the sun is taken once for the day, at 0h UT (with it at each time’s own moment the afternoon and evening times are off about one day in nine; taken once, one in three hundred); and Dhuhr is 1.06 minutes after the transit, not a whole minute (55 days off with a whole one, four with 1.06). The first eight days of January 2025 are not in the list.
Qatar
Certified over 16 Jun 2026 → 31 Dec 2026 (1 place, Doha): the method gives 95.393 % alone, and a list of 55 corrections the rest.
Research note, 2026-11-12 (not generated). The span is short, and that is the thing to know about this method: the House changed its times with its 1448 calendar (Dhuhr and Asr moved a minute later, and the calendar carries a note about a new study), so only the months of 1448 printed so far are this method’s — 118 days, mid-June to mid-October. The margins were read from those days alone; winter has not been seen, and a margin that is slightly off would show there as more corrections and be refitted as a new revision. Nearly all of the days off are Maghrib’s (and Isha’s with it, being Maghrib plus ninety): its distance from the computed sunset drifts by about half a minute between June and September, which a fixed margin does not follow. Doha is the only place: the calendar prints Doha in full and, for Al-Shamal, Dukhan and Abu Samra, minutes to add at Fajr and Maghrib only.
Research note, 2026-11-17 (not generated). The year is now known ahead: the House’s own app reads Doha’s whole Gregorian year from its server, and every one of the 125 days held from the calendars and the home page agreed with it to the minute. That gave the first days the margins were not read from — 74, the unread 12 October – 10 November and the rest of 2026 — and on them the method alone is exact on 417 of 444 times, never more than a minute off. Winter has now been seen, and it does not call for a new revision: fitting Doha’s point, the horizon and the margins on June–September and scoring on October–December, the best rule cut the misses there from 22 to 18 and was worse at Fajr — no rule was exact. The two runs of five September days whose Maghrib is printed a minute later than its neighbours’ (12–16 and 23–27) fit no geometry; they are the calendar’s own, and the correction list carries them. The other five places the House gives (on its home page, daily) move a minute either way from Doha’s day to day — computed at their own points, not Doha plus a fixed offset — and are not certified: a week of them is held.
Kuwait
Certified over 1 Jan 2026 → 31 Dec 2027 (46 areas): the method gives 95.205 % alone, and a list of 9,660 corrections the rest. The Ministry prints one table for Kuwait City and the areas as minutes on it, so the 46 areas are one check counted 46 times — and the list is one list of 105 times a year, written out for each area and each year.
Research note, 2026-11-12 (not generated). The table has no year: it is 366 days, the same every year, and 29 February repeats 28 February. Its Dhuhr is the transit of the sun of a year two after a leap year, and of the 1970s rather than now (1974 and 1978 give 360 of 365 days; later years of the same phase a few fewer, other phases far fewer), which is why no present-day calculation reproduces it. With 1978’s sun the standard angles give Fajr and Maghrib on 357 days of 365, Dhuhr on 360, sunrise and Isha on 341, and Asr on 334 — Asr only with the noon shadow taken from noon’s declination. The rest are a minute off with no pattern found, September’s Asr apart, which runs half a minute early for the whole month. A table that never changes is certified once: the same list serves every year, only its dates are rewritten.
Jordan
Certified over 1 Jan 2026 → 31 Dec 2026 (17 zones): the method gives 98.380 % alone, and a list of 603 corrections the rest.
Research note, 2026-11-12 (not generated). A quarter of the list is one prayer in one zone: Aqaba’s Maghrib (148 of its 365 days). Against the computed sunset it is 3.7 minutes later in winter and 5.1 in summer, which is neither a fixed angle nor fixed minutes — it follows the sun’s bearing, the way a real skyline to the west would, and a single horizon allowance cannot. Aqaba’s sunrise shows nothing of the kind. Without it the method alone is at 99.0 %. Each zone’s Fajr margin is read off its own year (nothing in the Valley, Azraq, Ma’an, Aqaba and Ruwayshid; a minute at Umm al-Qutan); no rule common to the zones was found for it. Asr settled the family of the program: with Morocco’s rule on the apparent sun it is off on 81 days of 6,205, with the plain shadow on more than 200.
Tatarstan
Certified over 1 Jan 2026 → 31 Dec 2026 (44 places): the method gives 98.071 % alone, and a list of 1,858 corrections the rest.
Research note, 2026-11-12 (not generated). The check is of six columns of the timetable’s eight: the end of sahur as Fajr, sunrise, the zenith as Dhuhr, Asr, Maghrib and Isha. The other two are the mosques’ own: a morning prayer 91 minutes before sunrise on all 16,060 rows, and a Dhuhr at a fixed clock time — 12:00 in thirty-seven districts, 11:45, 11:50, 12:05 or 12:15 in six, and in one 12:05 for seven weeks of winter and 12:00 otherwise. A reader in Tatarstan who follows the mosque wants those two, and an app can show them from the same sunrise; they are not times of the sun and are not certified here. The white-night rule is exact: on every row but four the dawn is the rule’s when the morning’s 18° twilight does not form and Isha is the rule’s when the evening’s does not. Asr decided the family of the program, as in Jordan: at two shadows on the apparent sun it is off about one day in thirty, on the true sun one day in four.
Russia: the Central Muftiate
Certified over 1 Jan 2026 → 31 Dec 2026 (18 cities): the method gives 99.877 % alone, and a list of 48 corrections the rest.
Research note, 2026-11-12 (not generated). The Muftiate publishes twenty-one cities; eighteen are here. Kazan’s page follows another rule (18° and 15°, no margins, fixed intervals in the white nights) and the city is in the Tatarstan method from its own Administration. Ufa’s timetable is published as pictures. Salekhard, on the Arctic Circle, follows this method through the year — its days of midnight sun are the 45th parallel’s own — but on five days at the edges of that season, when Maghrib falls after midnight, the page prints a Fajr in the evening and an Isha in the morning; a method should not give those, and the city is left out rather than certified around them. What decided the family of the program was sunrise: with an iterated sun one day in nine is a minute off at Moscow’s latitude, with the sun taken once at 6h none is. The white-night rule was read from the ratio of the twilight to the night, which on any one date is the same in every northern city of the list and is the 45th parallel’s; computed on the printed minutes and cut, it gives 2,525 of the 2,533 white-night times, and the other eight follow a neighbouring time that is itself on the other side of a minute.
Albania
Certified over 1 Jan 2026 → 31 Dec 2026 (22 towns): the method gives 98.082 % alone, and a list of 924 corrections the rest.
Bulgaria
Certified over 1 Jan 2026 → 31 Dec 2026 (48 towns): the method gives 97.671 % alone, and a list of 2,448 corrections the rest.
Research note, 2026-11-12 (not generated). Both are one town’s table and a column of minutes, so a town’s corrections are the reference town’s own, on the same days: Tirana’s 42 times and Sofia’s 51 are all there is, each counted once a town. What the tables are not is a sun followed through the day. Fitted with the sun at each time’s own moment, Sofia’s mornings wanted the calendar three quarters of a day back and its evenings a day and a quarter; one sun for the whole day, read thirteen hours before the date begins, gives both. Tirana’s is read at 0h UT. Neither moment is a round local hour, and the Bulgarian page names no year: whether its table is computed afresh each year or kept is not known from one year of it, and the next year’s check will say. The margins are the old temkin: seven minutes at each end of the day, five on Dhuhr, four on Asr.
Pakistan: Jamia Banuri Town
Certified over 1 Jan 2026 → 31 Dec 2026 (100 cities): the method gives 99.996 % alone, and a list of 8 corrections the rest.
Research note, 2026-11-12 (not generated). The page computes any city it is asked for, in Pakistan and abroad, under several settings; what is certified here is the cities of Pakistan under the settings it opens with: its own method and the Hanafi Asr. The arithmetic is the one many prayer-time programs share, and what tells it from an iterated sun is a minute here and there at sunrise and Asr. A city’s point is not printed; it is the point at which the method gives the city’s whole year, and the year pins it to about a ten-thousandth of a degree, since every minute of 2,190 has to fall on the right side. Karachi’s is 24°52′ N, 67°03′ E.
United Arab Emirates
Certified over 1 Jan 2026 → 31 Dec 2027 (41 zones): the method gives 98.221 % alone, and a list of 3,036 corrections the rest.
Research note, 2026-11-12 (not generated). The zones are the Authority’s: its areas, some four hundred, fall into forty-one tables (three pairs of zones print the same table). What told the method from a plain one: Dhuhr is two minutes after the transit; sunrise and Maghrib fit no single horizon but one a zone, the mountains’ far larger; and Asr near the Tropic was most of a minute late in June until the noon shadow was taken at Dhuhr rather than at the transit. In thirty-eight zones the method alone gives about 98.9 % of the times, the rest a minute away. Dubai’s three zones (Dubai, Lahbab and Margham, Hatta) are not read as well: about 93 % alone, with Dhuhr itself spread over a little more than a minute by any series for the sun tried, which says those tables come from another program. They are served by the same method and a longer list; past the list’s dates a time there may be a minute off. The reference is one year, read once; there is no scheduled renewal for this method.
Iran: Institute of Geophysics, Tehran
Certified over 21 Mar 2025 → 20 Mar 2027 (31 province capitals): the method gives 99.002 % alone, and a list of 903 corrections the rest.
Research note, 2026-11-12 (not generated). The sheets are two years, 1404 and 1405 of the Iranian calendar, a sheet a capital; each names its point to two decimals of a degree, the same in both years. The columns held to the sheets are Fajr, sunrise, Dhuhr and Maghrib; Asr and Isha are not printed and so are not certified. Sunset and midnight, which are printed, fit the same program as closely (midnight as halfway from sunset to the next day’s Fajr) and are outside the six times. No time of the method alone is more than a minute from the sheet, and the misses are spread evenly over the columns and the year — the last seconds of the sun, not another rule. The reference is read once a year, when the Institute puts the new sheets up; there is no scheduled renewal.
United Kingdom: Wifaqul Ulama
Certified over 1 Jan 2026 → 31 Dec 2027 (81 towns): the method gives 98.625 % alone, and a list of 4,641 corrections the rest.
Research note, 2026-11-12 (not generated). The reference is the 2027 timetable of every town and the 2026 one of nine. What told the method from a plain one: every time is the nearest minute but the Shafi’i Asr, which is rounded down; Isha follows 15° only in winter and a seventh of the night the rest of the year; the two Asrs drifted through the year against a noon shadow taken at noon, and fit one taken from the sun of the day’s start; and Fajr in the perpetual twilight is not the night’s middle but the middle of the date’s own sunset and the previous sunrise. No time of the method alone is more than a minute from the timetable. One district’s timetable (Birmingham, Solihull) is rounded another way and has a month out of place in 2027; it is left out, and Birmingham’s five other districts are in. The Shafi’i Asr is printed too and is not what is certified: the method alone gives it on the printed minute about nine days in ten. The reference is read once a year; there is no scheduled renewal.
Tunisia
Certified over 1 Jan 2025 → 31 Dec 2026 (26 places): the method gives 97.727 % alone, and a list of 2,156 corrections the rest.
Research note, 2026-11-12 (not generated). The Institute answers for a governorate and a delegation, a day at a time; the reference here is a sample: the twenty-four governorates’ seats, Aïn Draham and Rjim Maatoug, one year. What told the method from a plain one: Dhuhr and Asr sit on the sun of 0 h UT and not of their own hour, Dhuhr seven minutes after its transit; and Fajr, Isha and the sunset all move with the place’s height — one allowance a place, which is the dip of the horizon from its altitude (0.0293° × √metres fits them). No time of the method alone is more than a minute from the Institute’s, and the misses are spread evenly over the columns: the last seconds of another series for the sun, not another rule. The Institute gives no sunrise, so the sunrise here is not certified. The Institute’s other delegations are not read.
Not here, and why
Research note, 2026-11-11 (not generated).
- MUIS (Singapore) is not a method of this library and is not planned as one. Its table cannot be put on the minute by a formula: what is left after the best one is not a fraction of a second at a minute’s edge but a pattern (Dhuhr drifts by up to about 0.4 minutes over half a year, the other way round the next year), so a correction list for it would cover about one time in twenty and grow with every year published. MUIS publishes its timetable itself, free to use; that is where to take it from.
- JAKIM (Malaysia): the zone method stands at about 95 % on the minute, with whole columns in some states that no formula tried reproduces. Not added until that changes.