“When did I conceive?” sounds like a question with a one-day answer, and every conception date calculator — mine included — will happily hand you a single date. I am going to give you that date, and then I am going to be honest about what it really is: the centre of a window, usually about ten days wide, inside which conception almost certainly happened. Whether you are working backwards with a reverse due date calculator, forwards from your last period, building a conception calendar, or asking “when was I conceived?” about your own birthday, the method is the same — and so is the uncertainty. Here is how the calculation works, and how far to trust it.

First, the honest biology: why conception has a window, not a day

Conception means fertilisation — a sperm meeting an egg — and fertilisation is surprisingly hard to pin to a calendar day, for two reasons. First, sperm wait: in fertile cervical mucus, sperm can survive for up to five days, so intercourse on a Tuesday can lead to fertilisation on the Friday. Second, the egg hurries: once released at ovulation, an egg can be fertilised for only about 12 to 24 hours. Put those together and the research picture (Wilcox and colleagues, New England Journal of Medicine, 1995) is that conception can only result from intercourse during a six-day fertile window ending on the day of ovulation — but fertilisation itself happens on ovulation day, give or take a day, whenever that was.

And ovulation day is the thing we usually do not know exactly. It is inferred from a period that may be misremembered, in a cycle that may not be 28 days. So every “when did i conceive” answer carries two layers of estimation: when ovulation happened, and which day’s sperm were waiting. A responsible date of conception calculator therefore gives you a best estimate plus a plausible range — typically the estimate ± 5 days. Anyone who gives you a single day with total confidence is selling precision that biology does not offer.

Working backwards from your due date (the reverse calculation)

If you already have a due date — from a scan or your provider — this is the most reliable direction, because the due date is the firmest anchor you own. The two subtractions are:

Every date computed day by day. Note how the two checks agree: February 15 + 14 days (the standard ovulation assumption) is March 1 — the same conception estimate the 266-day subtraction produced.

So if you are asking “when did i get pregnant” with a due date of November 22, 2026 in hand, the answer I would give you is: most likely around March 1, 2026, plausibly any time from late February to the first days of March, and the act that led to it most likely happened between February 24 and March 1. That is as precise as honest arithmetic gets — and it is precise enough for every practical purpose, from working out a timeline to settling a friendly argument.

One refinement is worth adding: inside that six-day fertile window, the odds are not flat. In the same 1995 study, the probability of conception climbed across the five days before ovulation and peaked on ovulation day itself; the day after ovulation, it falls off a cliff, because the egg survives barely a day. So if you know there was one encounter near the end of the window, the centre of the estimate is more likely to be right than its edges — the window is honest, but its middle carries most of the weight.

Working forwards from your last period

No due date yet? Then start from the first day of your last menstrual period (LMP) and assume ovulation on day 14 of a 28-day cycle: conception estimate = LMP + 14 days, window = LMP + 9 to LMP + 19. Using the same worked example as our pregnancy calculator’s guide — an LMP of January 5, 2026 — conception most likely happened around January 19, 2026, plausibly between January 14 and January 24, and the due date that follows is October 12, 2026.

The catch is cycle length. Ovulation is better predicted by counting backwards — roughly 14 days before the next period — than forwards. If your cycles run longer or shorter than 28 days, the conception estimate moves with them:

Same period, four different conception estimates — an 11-day spread created by cycle length alone. This is the main reason first-trimester ultrasound, which measures the baby directly, overrules period arithmetic when the two disagree.

A third route: working back from an early scan

If you have had a first-trimester ultrasound, you own a better anchor than either of the two above, because the scan measures the embryo itself. The method: conception happened (gestational age at the scan − 14 days) before the scan date, since gestational age counts from two weeks before conception. Say a scan on March 16, 2026 measured you at exactly 10 weeks 0 days. Ten weeks is 70 days; 70 − 14 = 56 days before the scan puts conception at January 19, 2026, and the due date is the scan date plus the 210 days remaining of the 280 — October 12, 2026. It is satisfying that this lands on the same dates as the period-based example earlier: January 5 last period, January 19 conception, October 12 due date. When all three routes agree, you can be unusually confident in the calendar.

How much does the scan narrow the window? A crown–rump length measurement in the first trimester dates a pregnancy to within about 5 to 7 days, which is why ACOG’s redating margins (Committee Opinion No. 700) sit in that range. So the scan route shrinks your conception window from the period method’s ±5 days around a possibly-wrong ovulation guess to roughly ±5 days around a measurement — better, but still a window. Only IVF removes the guesswork entirely.

Building your conception calendar

A conception calendar lays the whole sequence on real dates, so instead of one estimated day you can see the machinery. Here is the full calendar for the reverse example above — due date November 22, 2026, equivalent LMP February 15, 2026:

One woman, one cycle, every key date computed from the two anchors. Your own calendar shifts as a block if your ovulation was earlier or later than day 14.

“When was I conceived?” — running it on a birthday

The same machine runs on any birth date, with one extra layer of uncertainty: you do not know how long that pregnancy actually lasted. If we assume the person was born on their due date, conception = birthday − 266 days. Someone born on June 15, 1990, for instance, was most likely conceived around September 22, 1989. But because babies arrive any time in a several-week span around the due date, the honest range is wider than the pregnancy version — roughly September 8 to October 6, 1989 covers births from two weeks early to two weeks late. Fun for family history; not evidence of anything more specific, and I would not carve it into a cake.

The exceptions that break the arithmetic

Three situations deserve a mention. IVF pregnancies are the great exception in the other direction: conception is not estimated at all, because fertilisation happened in the lab on a recorded date, and dating runs from the transfer instead — this article’s ±5-day window simply does not apply. I work through that transfer math in my IVF due date calculator guide. Very irregular cycles widen the window beyond ±5 days, and an early ultrasound becomes the only reliable anchor. And a conception date can never identify a specific encounter with certainty when encounters happened on several days inside the fertile window; the calendar constrains possibilities, it does not run a paternity test. Questions of that kind belong to a clinician and a DNA test, not a calculator.

Sources: fertile-window timing from Wilcox AJ et al., “Timing of sexual intercourse in relation to ovulation,” New England Journal of Medicine (1995); dating conventions (266 days from conception, 280 from LMP) and ultrasound redating margins from ACOG, Committee Opinion No. 700, “Methods for Estimating the Due Date” (2017, reaffirmed since). All dates computed day by day from those anchors. This article provides general information only and is not medical advice.