--- title: "How do period trackers predict periods?" description: "How period apps turn your past dates into estimates, why they can miss, and how to read a predicted window without treating it as a promise." url: "https://myruva.co/learn/how-period-trackers-predict-your-period" language: "en-GB" topic: "Getting started" published: "2026-10-18" updated: "2026-10-18" author: "Ruva team" reading_minutes: 6 --- # How do period trackers predict periods? How period apps turn your past dates into estimates, why they can miss, and how to read a predicted window without treating it as a promise. *Written and reviewed by the Ruva team · Updated 2026-10-18 · Getting started* A period tracker uses the period start dates you've entered to estimate when your next period might come. One simple method takes the average of your past cycle lengths. It then adds that number of days to your latest start date. A peer-reviewed [comparison of prediction models](https://pmc.ncbi.nlm.nih.gov/articles/PMC8714275/) looks at simple averages. It also looks at more complex methods that adjust as new information comes in and allow for missed entries. All these methods work from records. They can't measure what your body will do next, or know that the next cycle will match the last one. ## A forecast, step by step Cycle length is the time from one period start to the next. It includes the days when you aren't bleeding. The mean is the usual average of your past cycle lengths. The median is the middle value when you put those lengths in order. The [prediction-model study by Li and colleagues](https://pmc.ncbi.nlm.nih.gov/articles/PMC8714275/) used both for comparison. More complex models can combine your history with patterns from a larger dataset. Some also update a prediction as time passes. The dates below are made up to show how an average works. They are not a clinical example, and they don't describe Ruva's algorithm. | Period start you logged | Next start you logged | Cycle length | | --- | --- | --- | | 1 April | 29 April | 28 days | | 29 April | 29 May | 30 days | | 29 May | 30 June | 32 days | The three cycles add up to 90 days. Divided by three, that gives an average of 30 days. Add 30 days to the latest start, 30 June, and you get an estimated next start of 30 July. The maths is still right even if your next period doesn't start on 30 July. The uncertain part is whether the next cycle will match the average. Four start dates give three complete cycles, because the latest cycle has only just begun. Many apps keep their algorithms private. The study's authors couldn't compare their model directly with those private algorithms. So this example can't tell you how any one app works out its dates. Check the app's own explanation of its method. For help keeping track of the dates a calculation uses, see [getting started with cycle tracking](https://myruva.co/learn/getting-started-with-cycle-tracking). ## From a single date to a range In the example, you could repeat the shortest and longest past cycles. That gives a window from 28 July to 1 August. This shows a simple range. It doesn't promise that your next period will fall inside it, and it isn't a scientifically tested confidence interval. A single date and a range are two ways to show a prediction. A range can make the uncertainty easier to see. A spread based on past dates works differently from a model that estimates how likely each outcome is. The [Li study's probabilistic model](https://pmc.ncbi.nlm.nih.gov/articles/PMC8714275/) also allows for the chance that a period wasn't logged. Read how the app chooses its window. If it doesn't say, you can still use the range to plan everyday things, like packing period products. A wider coloured strip doesn't always come with a known probability. ## What temperature and test results add Some methods look at signs linked to ovulation as well as period dates. A [positive LH urine test suggests ovulation may be coming soon](https://medlineplus.gov/ency/article/007062.htm). It doesn't confirm that an egg was released. [Basal body temperature rises after ovulation](https://my.clevelandclinic.org/health/articles/21065-basal-body-temperature). Looking back, it can help show when ovulation probably happened. But the rise can't warn you ahead of time in that cycle. A fever or too little sleep can also affect the readings. Logging a test or a temperature doesn't mean the app uses it in its forecast. Check the app's explanation before you take a new entry as a sign that its prediction is now more certain. ## How researchers tested accuracy In a peer-reviewed [evaluation of period-tracker predictions](https://pmc.ncbi.nlm.nih.gov/articles/PMC8504278/), Worsfold and colleagues entered made-up cycle histories into apps. The apps did well with a cycle that was set to the same length every time. Their period predictions were less reliable when the cycles in the profiles varied. Different apps also gave conflicting results from the same histories. The profiles were built from existing data. The researchers didn't follow new participants, and they tested the app versions available at the time. So the results can't give an accuracy percentage for every tracker today, or for you. Also check what “accurate” means each time. Predicting when a period starts and predicting ovulation are different tasks. A result for one doesn't carry over to the other. Neither study says anything about how Ruva's estimates perform. ## Why your next period may fall outside the forecast Cycles vary naturally, and that makes them harder to predict. In the [Li study](https://pmc.ncbi.nlm.nih.gov/articles/PMC8714275/), cycle histories that varied more were generally harder to predict. Seeing a pattern in past cycles can't make the next cycle follow it. Missing entries matter too. If you have a period but don't log it, the gap between logged starts can look like one very long cycle. Li and colleagues found it helped to allow for possible missed entries. But real changes in your cycle also happen, so a long gap doesn't have to mean a missing entry. Don't add a period you can't remember just to make the chart look regular. An older pattern may be less useful after a life change. [Stress, weight changes, breastfeeding and perimenopause](https://www.nhs.uk/symptoms/missed-or-late-periods/) are some of the reasons a period may be late or may not come. Some hormonal contraception also changes bleeding. A forecast can't tell which reason applies to you. ## Reading the result in your app If a forecast surprises you, check the entries behind it. Mark dates you're unsure about as unsure. Keep logged start dates separate from estimates. A blank entry doesn't prove that nothing happened. [Recorded patterns versus predictions](https://myruva.co/learn/recorded-patterns-vs-predictions) explains how to read the two. A predicted period date or fertile window is not contraception. It can't show you guaranteed “safe days”. [Fertility awareness methods need to be taught by a trained professional](https://www.nhs.uk/contraception/methods-of-contraception/natural-family-planning/). The NHS does not officially recommend fertility apps or monitoring devices. Ruva's [estimates](https://myruva.co/app#estimates) use your own logged history. They stay labelled as estimates, separate from logged events and recorded patterns. Basal temperature logging can help mark when ovulation probably happened, looking back. The app does not detect or confirm ovulation. ## If your period hasn't arrived If you could be pregnant, take a pregnancy test. An app's forecast can't answer that question. [Test from the first day of a missed period. If you don't know when your period is due, test at least 21 days after you last had unprotected sex](https://www.nhs.uk/pregnancy/trying-for-a-baby/doing-a-pregnancy-test/). If the result is negative but you still think you could be pregnant, wait a few days and test again. If a second test is negative and your period still hasn't come, speak to a GP. [See a GP if you've missed three periods in a row, or if your periods have become irregular](https://www.nhs.uk/symptoms/missed-or-late-periods/). You can get help sooner if you're worried. [Pelvic pain when you could be pregnant needs urgent advice](https://www.nhs.uk/symptoms/pelvic-pain/). In the UK, use NHS 111. Elsewhere, contact your local urgent medical service. Severe or worsening pelvic pain needs emergency help, whatever date an app predicts. So does pain with faintness or heavy vaginal bleeding. In the UK, call 999 or go to A&E. Elsewhere, call your local emergency number or go to an emergency department. ## Quick answers ### Why is my period app wrong? A prediction uses the information it has, but your next cycle may differ from past ones. Missing period entries, life changes and hormonal contraception can also make past dates less useful for predicting. ### How accurate are period tracking apps? There isn't one accuracy figure for every app or every person. Published studies have found errors and differences between predictions. But the results depend on which apps, cycle histories and definition of accuracy were tested. ### Can I use my period app to avoid pregnancy? A predicted period date or fertile window is not contraception. It can't give you guaranteed safe days. If you want to use fertility awareness to avoid pregnancy, learn it from a trained professional. Don't rely on an app forecast. ## Sources 1. [A predictive model for next cycle start date that accounts for adherence in menstrual self-tracking](https://pmc.ncbi.nlm.nih.gov/articles/PMC8714275/), Journal of the American Medical Informatics Association 2. [Period tracker applications: What menstrual cycle information are they giving women?](https://pmc.ncbi.nlm.nih.gov/articles/PMC8504278/), Women's Health 3. [Basal Body Temperature: Family Planning Method](https://my.clevelandclinic.org/health/articles/21065-basal-body-temperature), Cleveland Clinic 4. [Ovulation home test](https://medlineplus.gov/ency/article/007062.htm), MedlinePlus 5. [Missed or late periods](https://www.nhs.uk/symptoms/missed-or-late-periods/), NHS 6. [Natural family planning](https://www.nhs.uk/contraception/methods-of-contraception/natural-family-planning/), NHS 7. [Doing a pregnancy test](https://www.nhs.uk/pregnancy/trying-for-a-baby/doing-a-pregnancy-test/), NHS 8. [Pelvic pain](https://www.nhs.uk/symptoms/pelvic-pain/), NHS --- This guide is general information to help you keep and use your own records. It isn’t medical advice. If you’re worried about a symptom, or something changes suddenly, talk to a doctor, nurse or pharmacist.