Pregnancy weeks and estimated due date: how it is calculated

Gestational age from the LMP

In obstetrics, pregnancy is counted from the first day of the last menstrual period (LMP), not from the day of conception. It sounds like a contradiction, on that first day the person was not even pregnant, but there is a practical reason: the LMP is a date almost everyone can recall, whereas ovulation and fertilization are rarely observed. Obstetrics needed a starting point that exists in every pregnancy, and the only universal candidate is the beginning of the last cycle. The gestational weeks calculator does exactly that: it counts the days since the LMP and divides by 7.

The consequence of that choice is a built-in offset. Because ovulation in a typical cycle happens around day 14, fertilization occurs roughly two weeks after the LMP. Gestational age therefore sits about two weeks “ahead” of the embryo’s real age, what textbooks call conceptional age. When someone says they are 8 weeks along, the embryo is about 6 weeks old. This is not a mistake: it is the convention, and it applies to every milestone of pregnancy.

Naegele’s rule and the cycle correction

The due-date calculator applies Naegele’s rule, in use since the early 19th century: add 280 days (40 weeks) to the LMP. An equivalent shortcut, handy to do in your head, is to add 1 year, subtract 3 months and add 7 days. Franz Karl Naegele, a German obstetrician (1778–1851), formalized the calculation from the observation that pregnancy lasted about ten lunar months after the last period, and the name stuck.

The detail that almost never makes the label: the 280 days only hold for a 28-day cycle with ovulation on day 14. The rule implicitly dates conception at LMP + 14 and, from there, counts 266 days of true gestation. If your cycle is longer, you ovulate later, conceive later and should therefore deliver later, plain Naegele pulls your due date too early. If it is shorter, it pushes the date too late. The error is systematic, not random, and it grows the further the cycle drifts from 28 days.

Example 1 (cycle correction). An LMP on 1 January 2026. With a 28-day cycle, due date = 1 Jan + 280 days = 8 October 2026, the same result as the shortcut (1 Jan 2026 + 1 year = 1 Jan 2027; − 3 months = 1 Oct 2026; + 7 days = 8 Oct 2026). Now the same LMP but with a 35-day cycle: due date = 1 Jan + 280 + (35 − 28) = 1 Jan + 287 days = 15 October 2026. Seven days of difference from cycle length alone. The due-date calculator assumes a 28-day cycle, so if yours is consistently longer or shorter, apply the correction yourself before trusting the date.

There is a second reason to treat the 280 days loosely: they are an average, not a law. Measuring from ovulation (not the LMP), the study by Jukic and colleagues followed natural pregnancies and found a median of 268 days between ovulation and birth, about 38 weeks and 2 days, with a natural spread of 37 days among uncomplicated pregnancies. Note that 268 days from ovulation works out to roughly 282 days from the LMP, a couple of days beyond the 266 that Naegele’s 280 implies. In other words, real gestation tends to run a little longer than the rule predicts, and it varies by about five weeks from one person to the next.

Weeks, days and trimesters

Gestational age is always expressed as “X weeks and Y days”. Just count the days since the LMP: integer division by 7 gives the completed weeks and the remainder gives the days, the same logic as any unit conversion, only from days to weeks. Example 2 (weeks). LMP on 1 January 2026, today 1 April 2026: that is 90 days. Since 90 ÷ 7 = 12 remainder 6, the pregnancy is at 12 weeks and 6 days, i.e. in the second trimester; about 190 days remain until the 8 October due date.

By convention, a full-term pregnancy lasts around 40 weeks counted from the LMP. Milestones like the end of the first trimester (12 weeks), viability and the anatomy scan are all referenced on this same count. That is why it matters that everyone, from the couple to the obstetrician, uses the same starting date: a one-week slip in the LMP shifts the entire schedule of exams at once.

When ultrasound replaces the LMP

Naegele’s rule assumes a regular 28-day cycle; longer, shorter or irregular cycles shift ovulation and, with it, the real date. This is where the early ultrasound becomes irreplaceable: in the first trimester, the crown-rump length (CRL) measures the embryo with little variation between pregnancies. Per ACOG, CRL dating up to 13 weeks and 6 days is accurate to ± 5 to 7 days, which makes it the most accurate method to establish or confirm gestational age, more reliable than the LMP itself. The ultrasound gestational-age calculator uses the INTERGROWTH-21st consortium formula for the CRL: a CRL of 60 mm, for instance, yields ≈ 12 weeks and 3 days.

But the ultrasound does not override the date every time it disagrees with the LMP. ACOG (Committee Opinion 700) set a tolerance: the due date is only changed when the gap between ultrasound dating and LMP dating exceeds a threshold that widens with gestational age. The logic is direct, the earlier the scan, the more precise it is, so a smaller gap is enough to beat the LMP; the later the scan, the more fetal size varies, so only a large gap justifies moving the date.

Example 3 (redate or not). The LMP points to 9 weeks and 2 days (65 days) on the day of the scan. The ultrasound, by CRL, measures 8 weeks and 3 days (59 days). The gap is 6 days. The fetus measures 8w3d, within the “up to 8w6d” band, where the tolerance is 5 days, and 6 days exceeds it. So ACOG supports switching the due date to the ultrasound’s, which falls about 6 days later (the fetus is a little younger than the LMP suggested, so delivery tends to come later). It is, however, a borderline case: the gap is small and the scan is nearly at the crossover into the next band, where the tolerance rises to 7 days. That is why the final call belongs to the obstetrician, who weighs how reliable the LMP is. Note that the calculator estimates the due date from the scan but does not apply this switching rule on its own, comparing the two dates is up to you and the care team.

Notice how the tolerance widens: 5 days at the very start, 21 days in the third trimester. It is a portrait of accuracy falling as pregnancy advances, the first-trimester CRL is the most precise, while late measurements can be off by weeks. That is why dating is “locked in” early: once set by an early ultrasound, the due date is usually not changed again, even if later scans show a bigger or smaller fetus. In the third trimester, in fact, a discrepancy may signal growth restriction rather than a dating error, another reason not to redate late.

A short history of pregnancy dating

The way we date a pregnancy moved from a fixed calendar rule to a direct measurement of the fetus, and then to criteria that say which of the two to trust. Each step solved a limitation of the one before it.

1. Early 19th centuryNaegele’s rule

   Franz Karl Naegele popularizes the 280-day count from the LMP. Simple and memorable, it becomes the worldwide standard, and remains useful even now that its limitations are understood.

2. 1958Obstetric ultrasound

   In Glasgow, Ian Donald, John MacVicar and engineer Tom Brown publish the use of pulsed ultrasound on the abdomen in The Lancet, opening the way to see the fetus instead of merely counting days.

3. 1973The CRL as a ruler

   Hugh Robinson shows, in the BMJ, that crown-rump length estimates maturity in the first trimester. The curve he produced still underpins the early dating used today.

4. 2013Measuring the real spread

   Jukic and colleagues measure, from ovulation, a median of 268 days and a natural spread of 37 days across pregnancies, the proof that the due date is a range, not a point.

5. 2017The redating rules

   ACOG consolidates, in Committee Opinion 700, when and how ultrasound should replace the LMP, the step that turned “LMP or ultrasound?” into an objective criterion.

The due date is an estimate, not a deadline

The estimated due date is a reference point, in practice, the center of a distribution, not a scheduled date. Only a small minority of babies are born exactly on the estimated day; most arrive within a window of a few weeks around it. The American College of Obstetricians and Gynecologists (ACOG) classifies the timing of birth as follows:

The chart tells the whole story: about 55% of births happen in weeks 39 and 40, and postterm is rare, largely because induction is usually offered by around 42 weeks. Arriving a few days before or after the due date is expected, not an exception. The date is used to schedule exams, track growth and guide decisions, always with a margin, and always in the context of prenatal care.

Like any health calculator, the due date is an educational estimate, the guide how to use health calculators safely explains this logic well, and the same caution applies to any population value placed on a specific body, from heart-rate zones to the due date. An average number guides; it does not decide on its own.