Naegele's rule, and the two assumptions inside it
The calculation is named for Franz Naegele, a German obstetrician who popularised it in the early nineteenth century. Take the first day of the last menstrual period, add seven days, subtract three months, add a year. That is the same as adding 280 days, and 280 days is 40 weeks. Nothing more sophisticated is happening.
Two assumptions are buried in it. The first is that the cycle is 28 days long. The second is that ovulation happened on day 14 of it. Both are averages, and cycle length in particular varies enormously between people and between months for the same person. A 35-day cycle usually means ovulation around day 21, which pushes the real conception date a week later than the rule assumes, and this calculator shifts the working last-period date by the difference so the estimate moves with it. That correction is itself an assumption — it presumes the luteal phase stayed at 14 days while the follicular phase stretched, which is the usual pattern but not a rule.
Why the count starts before conception
Gestational age is measured from the last menstrual period, not from conception, so week zero begins roughly two weeks before there is anything to be pregnant with. This is why a positive test taken when a period is a week late reads as five weeks pregnant rather than three. It is not a mistake in the counting; the last period is a date people can usually name and the moment of conception is not, so obstetrics standardised on the observable one and accepted the two-week offset. Every prenatal screening window, every growth chart and every reference to weeks in a clinical setting uses this convention, so it is worth getting used to rather than mentally correcting.
When ultrasound takes over
Between about 8 and 13 weeks, embryos grow at close to the same rate regardless of how big they will eventually be, which makes crown-rump length an unusually good clock. Measured in that window it estimates gestational age within about five days. Later scans lose that precision quickly, because by the second half of pregnancy fetal size reflects a great deal besides age.
Standard practice is therefore to compare the period-based date with a first-trimester scan and adopt the scan date if the gap exceeds a threshold that tightens as the pregnancy advances — a handful of days early on, more later. Once a due date is set this way it generally is not revised again, even if a later scan measures big or small. If you have been given a due date at a visit, enter it on the pregnancy week calculator in due-date mode rather than recalculating from a period here.
Reading the spread rather than the point
The table in the output exists because the single date does not carry the information people take from it. Births cluster around 40 weeks with a wide shoulder on either side: first pregnancies tend to run slightly longer, subsequent ones slightly shorter, and the individual variation swamps both effects. A pregnancy that reaches 41 weeks has not gone wrong, and one that delivers at 38 has not gone early in any meaningful sense — both sit inside the range that obstetric practice calls term.
What a date is genuinely useful for is scheduling: screening windows open and close on gestational age, and knowing which week you are in tells you which of them are live. That is the work this page is doing, and it is the only work it can do. Everything about how a particular pregnancy is progressing is visible only to the people looking after it.
Questions people ask
The scan gave a different due date. Which one is right?
The scan, if it was done in the first trimester. Crown-rump length between roughly 8 and 13 weeks dates a pregnancy to within about five days, better than any calculation from a remembered period date, and standard practice is to redate to the scan when the two differ by more than a small threshold. Once redated, the pregnancy is managed on that date from then on and later scans do not change it again. The number on this page is useful before a scan has happened, or as a rough cross-check afterwards, and nothing more.
My cycles are irregular. Does the cycle-length adjustment fix that?
Only partially, and it may not help at all. The adjustment assumes your ovulation is displaced by a consistent amount that tracks your average cycle length. With genuinely irregular cycles there is no consistent displacement to correct for, and the last period may not even have been followed by ovulation on any predictable schedule. If cycles have been irregular, treat any period-based estimate as provisional and wait for a first-trimester scan, which does not depend on any of this.
How do I use this after IVF?
Select the ovulation mode and enter the date that corresponds to conception, then let the calculator subtract fourteen days from it. For a fresh cycle that is the egg retrieval date. For a frozen transfer, subtract the age of the embryo at transfer: three days for a day-3 embryo, five or six for a blastocyst, so a day-5 transfer on the 20th is entered as the 15th. IVF dating is more accurate than period-based dating because the timing is known rather than inferred, and clinics normally issue the due date directly.
What are the chances of giving birth on the estimated date?
About one in twenty-five. Roughly half of births fall within a week of the estimate and the large majority somewhere between 37 and 42 weeks, which is a span of over a month. The date is the centre of that distribution rather than a target within it, and the practical consequence is that plans built tightly around the exact day — travel, work handovers, visitors — are being built on a number that is right about four percent of the time.