When to take a pregnancy test
Enter your ovulation date, or your last period and cycle length, plus the sensitivity printed on your test. This gives the earliest day a positive is plausible, the day a negative finally becomes reliable, and the odds on every day between. Testing early is the main cause of avoidable heartbreak, so the numbers here are deliberately honest rather than encouraging.
When to take a pregnancy test
No sign-up · PrivateThe day of your positive LH test, temperature shift or trigger shot. This is the accurate way to use this tool — every number below is counted from this day.
The first day of proper bleeding, not spotting.
First day of one period to the first day of the next. Ovulation is estimated as cycle length minus 14 days — the luteal phase is the steady part of the cycle, but it still varies by a few days.
Printed on the leaflet, usually as "sensitivity 25 mIU/mL". If the box says nothing, assume 25 — that is the standard supermarket test.
The earliest a positive is worth trying for
12 days past ovulation — about 17% of pregnancies show up by then on a 25 mIU/mL test, so a negative on this day tells you almost nothing
A negative stops meaning "not pregnant" from: 19 July 2026 — 18 days past ovulation, 4 days after your period is due
Better-than-even odds from: 15 July 2026 (14 DPO, 62%)
Period expected: 15 July 2026
Ovulation used: 1 July 2026 — the date you gave
Nothing you type leaves your device. The whole calculation runs in your browser.
How this is calculated
FormulaStep 1 — find ovulation (day 0). If you tracked it with LH strips, temperatures or a trigger shot, that date is used directly. From a last period instead, ovulation = LMP + (cycle length − 14), because the luteal phase varies far less than the follicular phase. Expected period = ovulation + 14 days.
Step 2 — when implantation happens. hCG production starts at implantation, not conception. Distribution of implantation day from Wilcox et al., NEJM 1999 (221 women, daily urine hCG): 6 DPO 0.5%, 7 DPO 5%, 8 DPO 22%, 9 DPO 36%, 10 DPO 26%, 11 DPO 8%, 12 DPO 2.5% — so 84% on days 8 to 10, and cumulative 0.5 / 5.5 / 27.5 / 63.5 / 89.5 / 97.5 / 100%. That cumulative column is a hard ceiling: no test detects a pregnancy that has not implanted. Wilcox et al., JAMA 2001 reached the same ceiling independently — implantation had occurred by the first day of the missed period in 90% of pregnancies (95% CI 84–94%), which they describe as the highest possible screening sensitivity on that day.
Step 3 — how fast hCG rises. Median urine hCG is modelled as 2 × 2(days since implantation ÷ 1.2) mIU/mL: about 2 mIU/mL on the day implantation completes, doubling roughly every 29 hours through the first week. That reaches 10 mIU/mL about three days after implantation and 25 mIU/mL about four and a half days after, consistent with the standard clinical statement that hCG is detectable in urine 12 to 14 days after conception, and with Gnoth and Johnson (2014), who report 25 mIU/mL as the 2nd centile of hCG on the day of the expected period. Between-woman spread is modelled as lognormal with a standard deviation of one doubling, reproducing the roughly ten-fold spread seen at a fixed gestational age.
Step 4 — combine. P(positive on day D) = Σi P(implant on day i) × P(hCG at D − i ≥ test sensitivity). “Earliest worth trying” is the first day the result reaches 10%; “a negative becomes reliable” is the first day it reaches 99%. Validation: for a 10 mIU/mL test at 14 DPO the model returns 89%, against Wilcox’s independently derived 90% ceiling. The model was not fitted to that figure.
What this deliberately does not do. It does not restate manufacturers’ “6 days early” claims as fact. Clearblue publishes 77% five days before the expected period, which on a 14-day luteal phase would be 9 DPO — a day on which under two thirds of pregnancies have implanted at all. Those figures are dated from a predicted period rather than a known ovulation and cannot be moved into a days-past-ovulation frame without contradicting the implantation research, so they are shown in their own column, labelled as manufacturer claims.
A test cannot find a hormone that is not being made
Home pregnancy tests detect human chorionic gonadotrophin. Your body does not produce hCG at conception; it produces it once the embryo has implanted in the wall of the uterus. Everything about early testing follows from that one fact, and almost every distressing early negative traces back to ignoring it.
Wilcox and colleagues measured daily urine hCG in 221 women trying to conceive and published the result in the New England Journal of Medicine in 1999. Among pregnancies that lasted six weeks or more, 84% implanted on day 8, 9 or 10 after ovulation. None implanted before day 6, and none after day 12.
| Days past ovulation | Implanted that day | Implanted by then |
|---|---|---|
| 6 | 0.5% | 0.5% |
| 7 | 5% | 5.5% |
| 8 | 22% | 27.5% |
| 9 | 36% | 63.5% |
| 10 | 26% | 89.5% |
| 11 | 8% | 97.5% |
| 12 | 2.5% | 100% |
Read the right-hand column carefully, because it is a ceiling nothing can beat. At 9 days past ovulation, barely two thirds of pregnancies have implanted at all. A negative that day is not evidence of anything. The same research team reached the identical conclusion from the other direction in JAMA in 2001: implantation had occurred by the first day of the missed period in 90% of pregnancies (95% CI 84–94%), so 90% is, in their words, the highest possible screening sensitivity for an hCG-based pregnancy test on that day. A test that claimed better than 90% on the day your period is due would be claiming to detect pregnancies that have not started producing hormone yet.
What the numbers actually look like
Once implantation happens, hCG climbs fast — roughly a doubling every 29 hours through the first week, from about 2 mIU/mL on the day implantation completes. Combining that rise with the implantation distribution above gives a probability of a positive for every day and every common test sensitivity.
| DPO | vs. period due | Implanted | 10 mIU/mL | 20 mIU/mL | 25 mIU/mL | Maker claim |
|---|---|---|---|---|---|---|
| 9 | 5 days before | 64% | 3% | 1% | under 1% | 77% |
| 10 | 4 days before | 90% | 12% | 3% | 1% | 93% |
| 11 | 3 days before | 98% | 28% | 9% | 6% | 98% |
| 12 | 2 days before | 100% | 52% | 24% | 17% | over 99% |
| 13 | 1 day before | 100% | 74% | 47% | 38% | over 99% |
| 14 | the day it is due | 100% | 89% | 70% | 62% | over 99% |
| 15 | 1 day late | 100% | 97% | 87% | 82% | over 99% |
| 16 | 2 days late | 100% | 99% | 96% | 93% | over 99% |
| 17 | 3 days late | 100% | over 99% | 99% | 98% | over 99% |
| 18 | 4 days late | 100% | over 99% | over 99% | over 99% | over 99% |
Three things stand out. First, the standard 25 mIU/mL supermarket test is genuinely slower than an early-detection strip: it reaches about 62% on the day your period is due, against 89% for a 10 mIU/mL test. Second, none of them is a trustworthy negative until several days after that — roughly two days late for a 10 mIU/mL test and four days late for a 25. Third, the gap between the modelled figures and the numbers printed on the box is large, and worth explaining.
Why we do not simply repeat the packaging
Clearblue publishes 77% five days before the expected period for its Early Detection test. On a 14-day luteal phase that is 9 days past ovulation — a day on which under two thirds of pregnancies have implanted at all. The figures are not dishonest; they are counted from a predicted period date rather than a known ovulation, and predicted period dates are wrong in both directions. They simply cannot be translated into a days-past-ovulation frame without contradicting the implantation literature. This tool reports both, in separate columns, and lets you see the disagreement rather than picking a side quietly. Gnoth and Johnson made a related point in 2014: about half the tests they examined did not match their own package-insert claims, and they called for enforceable international standards.
Getting the most out of the test you have
Use first morning urine, when hCG is most concentrated — drinking a lot of water beforehand dilutes the sample and is a real cause of false negatives this early. Read the result inside the window printed on the leaflet, usually three to ten minutes. A faint line inside that window is a positive. A line that surfaces an hour later is an evaporation line, not a result.
If you did not track ovulation, the calculator estimates it as cycle length minus 14 days. The luteal phase is the steadier half of the cycle, which is why that assumption is standard, but ovulation still moves by several days with illness, stress or travel, and every date shifts with it. If you have LH strips or temperature data, use them; the answer tightens considerably.
The case for waiting
More sensitive tests find more very early pregnancies, including ones that would have ended before a period was ever missed. Somewhere between one in ten and one in four conceptions end this way, usually without anyone knowing. Testing at 9 or 10 days past ovulation means learning about losses you would otherwise never have had to grieve. Some people would rather know. It is worth deciding which you are before you buy the box, rather than at six in the morning holding a stick.
The NHS advises testing from the first day of a missed period, or at least 21 days after unprotected sex if you are unsure when that was. If the result is negative and your period still has not arrived, test again a few days later, and speak to your GP if a second test is also negative. Positives are far more reliable than negatives: a positive is almost always real, whereas a negative may simply be early.
Sources
- Time of implantation of the conceptus and loss of pregnancy. N Engl J Med 1999;340(23):1796-9 — New England Journal of Medicine (via PubMed), accessed
- Natural limits of pregnancy testing in relation to the expected menstrual period. JAMA 2001;286(14):1759-61 — JAMA, accessed
- Strips of Hope: Accuracy of Home Pregnancy Tests and New Developments. Geburtshilfe Frauenheilkd 2014;74(7):661-9 — PubMed Central (NCBI), accessed
- Doing a pregnancy test — NHS, accessed
- Early Detection Pregnancy Test: published detection rates by day before expected period — Clearblue (SPD Swiss Precision Diagnostics), accessed