Pregnancy
Pregnancy risk by cycle day: what the evidence actually supports
The six-day fertile window, the two day-specific probabilities that are genuinely published, why the days in between are deliberately left blank, and the trap of confusing an annual failure rate with the risk from one encounter.
Free online tool
Pregnancy Risk
The fertile window is shorter than most people think
The classic study on this is Wilcox and colleagues in the New England Journal of Medicine in 1995: 221 women, 625 cycles, with ovulation dated from daily hormone measurements rather than guesswork. The finding was clear-cut. Conception occurred only when intercourse fell within a six-day window ending on the day of ovulation. Not a single pregnancy came from intercourse after that day.
The window is asymmetric, and that asymmetry is biological. Sperm can survive several days in the reproductive tract waiting for an egg; the egg itself is viable for less than a day. That is why the fertile days sit before ovulation and not after it.
The two probabilities that are actually published
Day −5
10%
Five days before ovulation: the start of the window.
Day 0
33%
The day of ovulation itself: the highest published value.
Why the days in between are left blank
Wilcox published point estimates for day −5 and day 0. The intermediate days (−4 to −1) are not given as figures in any source we were able to fetch. Drawing a smooth curve between the two endpoints would look convincing and would be invented data, so those days are shown as a band instead of a false percentage.
Later re-analyses of the same dataset place the peak one to two days before ovulation rather than on the day itself. We could not reach the full text, so that shapes how the result is worded and never produces a number.
Why cycle day is an estimate, not a measurement
Everything above depends on knowing when ovulation happened. Outside a study, nobody does. The usual approach works backwards from the luteal phase, the stretch between ovulation and the next period, which is far more stable than the first half of the cycle.
- The luteal phase lasts about 14 days in most people, and varies little between cycles.
- The follicular phase — from the period to ovulation — is the part that stretches and shrinks. That is what makes cycles 'irregular'.
- So ovulation is estimated as roughly 14 days before the NEXT period, not 14 days after the last one.
- Illness, travel, stress, and sleep disruption can all shift ovulation in a given cycle, which moves the whole window with it.
The trap: annual failure rates are not per-encounter risk
Contraceptive effectiveness is published as a first-year, typical-use failure rate: out of 100 people using the method for a year, how many become pregnant. It already contains every real-world slip — the forgotten pill, the condom put on late. It is not, and cannot be converted into, the chance that one particular encounter results in pregnancy.
Typical-use failure in the first year
Implant and IUD
< 1%
Nothing to remember, so typical use and perfect use nearly coincide.
Injectable
4%
Depends on returning for the next dose on time.
Pill, patch, ring
7%
The gap versus perfect use is almost entirely missed doses.
Male condom
13%
Sensitive to consistent and correct use on every occasion.
Withdrawal
14–27%
Reported ranges vary widely between studies.
Fertility awareness
2–34%
The spread reflects very different methods and levels of training.
Never multiply the two together
Combining a day-specific probability with an annual failure rate produces a number with no meaning. They measure different things over different time spans. Read them side by side, not as a single figure.
What this does not cover
- It says nothing about emergency contraception, which is time-sensitive: if that is the question, the answer is a pharmacist or a doctor today, not a calculator.
- It does not account for individual fertility. Age, medical conditions, and previous surgery all shift the real probability and none of them are inputs here.
- It offers no protection against sexually transmitted infections, which no cycle calculation addresses.
- It cannot confirm or rule out an existing pregnancy. Only a test can do that, and only after enough time has passed.
An estimate, not a contraceptive method
Using cycle timing to avoid pregnancy is a recognised approach, but it requires training and consistent daily tracking to reach the low end of its range. A one-off estimate is not that, and should not be relied on as contraception.
Sources
- Wilcox AJ, Weinberg CR, Baird DD (1995). Timing of sexual intercourse in relation to ovulation. N Engl J Med 333(23):1517–1521.
- Wilcox AJ, Dunson D, Baird DD (2000). The timing of the "fertile window" in the menstrual cycle: day specific estimates from a prospective study. BMJ 321(7271):1259–1262.
- Guttmacher Institute (2020). Contraceptive Effectiveness in the United States, from Hatcher RA et al., Contraceptive Technology, 21st ed.
