The formula behind the result
A menstrual cycle is counted from the first day of one period to the first day of the next. Predicting ahead is repeated addition of that length.
Start of period n = L + n × C
Last day of period n = start + D − 1
L is the first day of the last period, C is the cycle length in days, n is 1 for the next period, 2 for the one after, and so on. D is the number of days a period usually lasts; one is subtracted because the first day counts as day 1. This is plain calendar counting, the same approach paper period diaries have always used, and it is not tied to any single published source.
Average cycle = (sum of the gaps between start dates) ÷ (number of gaps)
Standard deviation = √( Σ (gap − average)² ÷ (number of gaps − 1) )
When you enter earlier start dates, the calculator sorts them, measures the gap in days between each neighbouring pair, and averages the gaps. It also reports the shortest and longest gap and, when there are at least two gaps, the sample standard deviation, a standard statistical measure of spread.
Worked example
The last period started on 3 March 2026. The cycle is 29 days and periods last 5 days.
- Next start. 3 March + 29 days. March has 31 days, so 28 days reach 31 March and one more reaches 1 April 2026.
- Last day. 1 April + 5 − 1 = 5 April 2026. The predicted days are 1, 2, 3, 4 and 5 April.
- The one after. 1 April + 29 days = 30 April 2026.
- The third. 30 April + 29 days = 29 May 2026.
Now the average. Suppose four periods started on 5 January, 2 February, 5 March and 1 April 2026.
- Gaps. 5 January to 2 February is 28 days, 2 February to 5 March is 31 days, and 5 March to 1 April is 27 days.
- Average. (28 + 31 + 27) ÷ 3 = 86 ÷ 3 = 28.7 days.
- Range. The shortest gap is 27 days and the longest is 31, a spread of 4 days.
- Standard deviation. The squared differences from the average are 0.44, 5.44 and 2.78, which sum to 8.67. Dividing by 2 gives 4.33, and the square root is 2.1 days.
Predictions use whole days, so the 28.7-day average is rounded to 29.
How to read the result
The first predicted date is the most dependable and each later one is less so. If your cycles vary by a few days, that uncertainty stacks up: after six cycles, the real date could plausibly sit a week or more from the prediction. The shortest and longest gaps in your own history are the best practical guide to how far a date might move.
An average needs data. Two start dates give one gap, which is a single observation and not much of an average. Five or six dates begin to show a pattern. A missed entry looks like one very long cycle, so if a gap is roughly double the others, check whether a period was left out. The calculator sets aside a history that contains a gap shorter than 10 days or longer than 90, on the assumption that a date was mistyped or missed.
Cycle length commonly shifts with stress, illness, travel, changes in weight or exercise, hormonal contraception, breastfeeding, and at the beginning and end of the reproductive years. A calendar cannot foresee any of these. It also cannot tell you when ovulation happens with any certainty, which is why it must not be relied on to avoid pregnancy.
If the last period you entered was more than one cycle ago, the calculator keeps adding cycles until it reaches a date that is today or later. It does not know whether a period was missed in between.