Two conventions, and the dates where they part
Two rules for the same calculation are in general use, and both are stated openly by the sources that use them. This page runs them next to each other, counts the dates where they give different figures, and shows the structural reason a disagreement can never be an ordinary one.
The same date, added two ways
The first route adds every digit of the date in one string and reduces the total. The second reduces the day, the month and the year first, adds those three results, and reduces again.
22 September 1980, whole date: 2 + 2 + 9 + 1 + 9 + 8 + 0 = 31, then 3 + 1 = 4. Separately: day 22 → 2 + 2 = 4; month 9; year 1 + 9 + 8 + 0 = 18 → 1 + 8 = 9. Adding those, 4 + 9 + 9 = 22, which is left standing. One route gives 4, the other 22.
12 June 1977 runs the other way. Whole date: 1 + 2 + 6 + 1 + 9 + 7 + 7 = 33, held back. Separately: day 1 + 2 = 3; month 6; year 1 + 9 + 7 + 7 = 24 → 2 + 4 = 6. And 3 + 6 + 6 = 15, then 1 + 5 = 6.
Across the 35,064 calendar dates from 1930 to 2025 the two routes part on 4,167 of them — 11.88 per cent, about one date in eight. On 3,162 of those the whole-date route ends on a held-back total and the separate route does not; on the other 1,005 it is the separate route that ends on one.
Every disagreement has a master number in it
Both routes are nothing but repeated digit sums, and a digit sum leaves a number’s remainder on division by nine unchanged. So whatever else the two routes do, they arrive carrying the same remainder — and two ordinary single digits with the same remainder are the same digit.
That leaves exactly one way for them to differ: one of them stops on 11, 22 or 33 while the other carries on. All 4,167 disagreements counted above are of that kind, without exception. It also gives you a diagnosis for free: two calculators that hand you two different plain digits for one date are not following two conventions. At least one of them has made a mistake.
What the tradition makes of the split
In this tradition each school names its rule and works within it, and several print both figures so the reader can see what was set aside. None of them treats the other route as a corruption; the disagreement is old, visible and largely unremarked, in the way that house style is unremarked.
The consequence is worth stating plainly, because it is usually skipped. The descriptions the tradition attaches to 4 and to 22 are different descriptions. So for 22 September 1980 the tradition holds out two readings and offers nothing that would choose between them — not evidence, not seniority, not a test. Which reading a person meets depends on which website they opened.
That is a fact about the sources, not about anybody born that day. It is also the strongest reason to record the rule beside the number: a figure with its rule attached is a claim somebody else can follow, and a figure without one is an assertion.
The second column, exactly as this page computes it
The right-hand column of the calculator below takes the digit sum of each part once, and then reduces that result. The effect is easy to state and easy to check.
The day: 29 has a digit sum of 11, which is held back, so the 29th is the only day of the month that carries a master number into this column. A day of the 11th has a digit sum of 2 and a day of the 22nd has 4.
The month: never. November’s 11 has a digit sum of 2, December’s 12 has 3, and no other month gets near.
The year: only when its digits add to 11 or 22. Between 1900 and 2025 there are four years of the first kind — 1901, 1910, 2009 and 2018 — and seven of the second: 1939, 1948, 1957, 1966, 1975, 1984 and 1993. No year in that span adds to 33.
Those eleven years and the 29th of any month are the only places a master number can enter this column as an input. It can still finish on one by adding three ordinary parts, the way 4 + 9 + 9 = 22 does above. A page that tells you a day of the 11th supplies a master number here is describing a different implementation from the one it is standing on.
A third variation, and where it shows
Some implementations apply the stop rule to each part as well, so a day of the 11th stays 11 instead of becoming 2, and a day of the 22nd stays 22. It is the same convention with the rule pushed one level down, and it differs from the column on the right on 891 of the 35,064 dates in the window — 2.54 per cent.
11 February 1980 shows all three at once. Whole date: 1 + 1 + 2 + 1 + 9 + 8 + 0 = 22, held back. The separate route as computed here: day 1 + 1 = 2, month 2, year 1 + 9 + 8 + 0 = 18 → 9; then 2 + 2 + 9 = 13, and 1 + 3 = 4. The third variation: day 11 kept, month 2, year 9; then 11 + 2 + 9 = 22, held back.
Two of the three answer 22 and one answers 4, and each is applying its own rule correctly. There is no arithmetic error anywhere in that paragraph, which is the point of it.
Working with two answers
- Which convention does the calculator on this page use?
- Both, side by side. The left-hand figure adds the whole date; the right-hand one reduces the day, the month and the year first. The line underneath says whether your date is one where they agree.
- If the two disagree on my date, which number is mine?
- The method has no answer, and anyone who gives you one has substituted their house style for a reason. What you can say precisely is which rule produced which figure, and that is what belongs beside the number when you write it down.
- Does it matter whether I write the date day-first or month-first?
- Not for the life path. Both routes add the same set of digits and both treat the day and the month the same way, so 3/10 and 10/3 give the same figure on either route. Other date-based numbers do use the day as a value in its own right, and those are a separate calculation.