Late Period or Pregnant? Probability Calculator

What this estimates

This tool provides an educational estimate of the probability that a late or expected period is due to pregnancy versus normal cycle variation. It combines (1) a prior probability based on exposure type (no sex, protected only, any unprotected) with (2) the likelihood of being late if pregnant vs not pregnant, and (3) optional test results, using a simple Bayesian update. Bars are complementary and are not a diagnosis.

“Days late” and expected period

You supply days late directly (negative values mean “not due yet”). The expected period date is today minus the “days late” value, which lets the tool align any past tests to how many days before/after the expected period they were taken.

Cycle variation when not pregnant

Cycle length varies within and between people. We approximate day-to-day variation around your average using a normal-like spread with your chosen standard deviation (SD). The likelihood of being late if not pregnant is modeled as the upper-tail probability that a cycle runs at least “mean + days-late.” Large app-dataset studies show substantial real-world variability in cycle length and follicular-phase timing, supporting use of a dispersion parameter rather than a fixed 28-day assumption. Key evidence: large cohort of app users describing cycle-length distributions and within-user variability (Bull et al., 2019, Scientific Reports).

Pregnancy prior by exposure

We set conservative priors: very low for “no sex,” low for “protected only,” and higher for “any unprotected.” These anchors reflect day-specific fecundability curves (probability of conception given intercourse on days around ovulation) and the fact that timing is usually imprecise outside active trying. Classic day-specific data show conception probabilities concentrated in the fertile window with a sharp peak near ovulation (Wilcox et al., 2000, BMJ). Typical-use contraceptive effectiveness supports a small residual risk with protected sex across a cycle (CDC Contraception Effectiveness).

Likelihood of lateness if pregnant

If pregnant, the chance that menstruation has not occurred by the expected date is high and grows with additional late days. This reflects early implantation timing and onset of hCG production, which suppresses menses; most successful implantations occur roughly 8–10 days after ovulation (plausible range ~6–12 days) (Wilcox et al., 1999, NEJM).

Incorporating pregnancy tests

If a positive test is reported, the model sets a very high pregnancy probability. For negative tests, we down-weight the probability by the test’s expected sensitivity on the day it was taken, mapped by days before/after the expected period (for example, lower sensitivity before the missed period, ~90% around the due day, rising to >95% several days after). Bench studies and reviews show that analytical sensitivity and real-world performance vary by brand and timing; accuracy rises as hCG increases after implantation (Johnson et al., 2015, CCLM). The biological basis for early detectability is the implantation-to-hCG timeline (Wilcox et al., 1999).

Computation summary

Posterior = [Prior_preg × P(late | preg)] / ([Prior_preg × P(late | preg)] + [(1 − Prior_preg) × P(late | not-preg)]). Each negative test multiplies the posterior by (1 − sensitivity) for that day. We floor extremely small values (<0.0001%) to avoid showing 0%.

Interpreting the output

The two bars sum to ~100% and represent a relative split between pregnancy vs cycle variation under the model assumptions. “Next best test day” and “retest” suggestions follow the common practice of testing on or after the expected period and repeating 48 hours later if negative, when hCG has typically doubled.

Limitations

Population averages and simplified assumptions cannot capture individual variation in ovulation timing, luteal phase length, test brand/technique, or medical conditions. If you are 7+ days late with negative tests, or have pain, heavy bleeding, or other concerns, seek clinical advice.

Key sources

• Bull JR, et al. “Real-world menstrual cycle characteristics of more than 600,000 app users.” Scientific Reports, 2019. Link.
• Wilcox AJ, Dunson D, Baird DD. “Timing of sexual intercourse in relation to ovulation.” BMJ, 2000. Link.
• Wilcox AJ, et al. “Time of implantation of the conceptus and loss of pregnancy.” NEJM, 1999. Link.
• Johnson S, Cushion M, Bond S, Godbert S. “Strips of Hope: Accuracy of home pregnancy tests and new developments.” Clinical Chemistry and Laboratory Medicine, 2015. Link.
• CDC. “Contraception—Effectiveness of Family Planning Methods.” Link.