How accurate are period tracker predictions?
Period apps predict your next period reasonably when your cycles are steady and poorly when they are not, and calendar-only fertile-window predictions are unreliable. A 2016 study found only a small minority of apps and websites identified the fertile window correctly. In Splash’s own tests on made-up data, the real start fell inside the range 82% of the time for regular cycles.
On this page 7 sections
What does research say about period app accuracy?
Bull and colleagues analysed 612,613 cycles from 124,648 users of a fertility app and found a mean cycle length of 29.3 days, which is why a fixed 28-day assumption misses so often. The study was funded by Natural Cycles.
Setton and colleagues (2016) tested 20 websites and 33 apps against an estimated fertile window: one website and three apps predicted it correctly, as reported by Medical Xpress. Apps that count calendar days alone are guessing at ovulation, because ovulation timing varies from cycle to cycle.
Why do predictions miss?
- Cycles vary by several days even in people with regular periods, and stress, illness, travel, sleep and some medicines shift them.
- The time from ovulation to your next period (the luteal phase) is steadier than the first half, around 14 days for many people but 10 to 16 is common.
- Conditions such as PCOS and perimenopause make cycle length far less predictable.
- An app cannot know about things you have not logged.
How accurate is Splash?
We do not call Splash’s predictions accurate. We test them on synthetic cycle histories where the true dates are known, show the results, and show predictions as ranges. The table compares a simple average of the last six cycles with Splash’s current method.
| Cycle type | Average error, simple average → Splash (days) | Real start inside the range, simple average → Splash | Range width, simple average → Splash (days) |
|---|---|---|---|
| Regular | 1.98 → 1.76 | 59% → 82% | 3.3 → 5.8 |
| Irregular | 4.12 → 4.02 | 45% → 75% | 6.5 → 12.4 |
| PCOS-like | 11.35 → 11.44 | 30% → 77% | 8.8 → 32.6 |
| Perimenopause-like | 9.03 → 9.41 | 32% → 74% | 6.2 → 22.3 |
Splash’s own tests on synthetic (made-up) cycle histories, re-run with each app build. Real cycles are messier. Full method on the how-predictions-work page. Even with learnt luteal length, the estimated fertile window missed the true ovulation day in a third or more of cycles in the short-luteal group.
How can I get better estimates?
- Log every period start, because everything else is built on those dates.
- Read the whole range, and treat the most-likely day as a guide.
- Add LH tests, basal temperature or wearable temperature if you want ovulation estimates. Splash labels each as Estimated, Likely or Confirmed.
- Mark cycles you want left out, such as after stopping hormonal contraception or during pregnancy or breastfeeding.
- Check the insight that replays Splash’s method on your own past cycles once you have four or more logged.
Can I use predictions to avoid pregnancy?
No. Splash is not contraception and not a medical device, and calendar estimates miss the true ovulation day often enough that they are not a method of birth control. If you are avoiding pregnancy, use a method designed for it and ask a pharmacist, GP or sexual health clinic.
Questions people ask
Are period tracking apps accurate?
They are reasonably good at the next period if your cycles are steady, and weaker at the fertile window, especially when they only count calendar days. Accuracy depends on how regular your cycles are and how much you log. Ranges are a more honest way to show a prediction than a single date.
Why is my period app always wrong?
Usually because your cycles vary more than the app assumes, or it assumes 28 days. Stress, illness and travel shift periods, and conditions such as PCOS make cycles unpredictable. Look for an app that shows a range, learns from recent cycles and tells you how sure it is.
Can a period tracker tell when I ovulate?
Only roughly from dates. Tests and body signs do better: LH tests, cervical mucus and a sustained temperature rise after ovulation. Splash combines them and labels the result Estimated, Likely or Confirmed, with a note on how it worked it out. A confirmed estimate is still not contraception.
Do wearables make period predictions more accurate?
Temperature from a watch or ring can show a rise after ovulation, which helps confirm it after the fact. Splash uses it that way and does not use it to predict fertile days ahead. Skin temperature is affected by sleep, alcohol, illness and room temperature, so it is a clue rather than proof.
Sources
- Bull JR et al. Real-world menstrual cycle characteristics of more than 600,000 menstrual cycles. npj Digital Medicine, 2019 accessed 2026-10-28 (Figures read through summaries of the open-access paper; Natural Cycles funded the study.)
- Medical Xpress: Most websites, apps not accurate for predicting fertile window (Setton et al., Obstetrics & Gynecology, 2016) accessed 2026-10-28 (Secondary report of the 2016 paper.)
Splash makes one of the apps mentioned here, so every claim about another app links to its own page or an independent source. Spotted something wrong? Email us and we will fix it and note the change.
Keep reading
- How Splash predicts, with the tests
- How accurate are period apps? The full read
- Why your predictions shift
- Ovulation calculator
- Predictions in Splash
- Is my period tracker data private?
- What is the best period tracker for irregular periods in the UK?
- Is there a period tracker that works without an account?
- Which period tracker is best for perimenopause?
General information, not medical advice. Splash is not contraception, not a medical device and cannot diagnose, and it is not affiliated with the NHS. If you are worried, talk to a GP, pharmacist or sexual health clinic. How we write and check content.
Predictions with the uncertainty showing
Ranges, a confidence label and no account.