| "Helping women track their ovulation in a smarter way" | 
Key takeaways
- An app prediction and a test strip are doing two different things: one estimates from your history, the other measures a hormone now.
- Neither is a more truthful version of the other. They answer different questions.
- Predictions built from cycle history assume cycle lengths are reasonably stable, which is often not the case in polyendocrine metabolic ovarian syndrome (PMOS, formerly PCOS).
- Disagreement is expected when cycles vary, not a sign something has gone wrong.
- When prediction and measurement conflict, a signal that looks back after the fact, like a progesterone marker, is the clearer guide.
An app prediction estimates a fertile window from your previous cycle lengths. A test strip measures luteinizing hormone right now. They answer different questions, so they can disagree without either being faulty. They disagree more often when cycle lengths vary, which is common in PMOS.
Why do they disagree at all?
Because they are not measuring the same thing.
| App prediction | Test strip |
|---|---|
| Inference from the past | Observation in the present |
| An app prediction is an estimate built from history. It takes the cycle lengths you have logged and projects forward. It knows nothing about your hormones today. | A test strip is a measurement taken now. It measures luteinizing hormone in urine and reports whether the amount is above a set level.1 It knows nothing about your history. |
One is inference from the past. The other is observation in the present. When they agree, that is two independent methods pointing the same way. When they disagree, it usually means the assumption behind the prediction did not hold this month.
What assumption does a prediction rely on?
That your cycles are reasonably consistent.
Any method that estimates a fertile window from previous cycle lengths assumes those lengths are stable enough to project from. The more they vary, the less the projection carries.
That matters here specifically, because the variable part of the cycle is the part that sets the timing. The second half, from ovulation to a period, is comparatively consistent. The first half lasts as long as it takes for a follicle to be selected and mature,2 and in PMOS that selection step is often delayed or does not resolve at all.
So the part a prediction most needs to be stable is the part that is least stable.
This is true of any prediction built from cycle history, including ours. It is not a limitation of one app versus another, and switching apps does not resolve it. It is a property of the method.
Which one should I trust?
Neither, exclusively. The question is what you are trying to find out.
| If you want to know | The more useful signal |
|---|---|
| Roughly when a fertile window might fall next month | The prediction, held loosely |
| Whether LH is raised today | The test strip |
| Whether ovulation actually happened | A progesterone marker after the fact1 |
| Whether there is a pattern at all | Several months of logged cycle lengths |
Notice that the third row is not answered by either of the first two. Neither a prediction nor an LH test can show that ovulation likely happened. That takes a separate measurement afterwards.1
Why does this get harder in PMOS?
Two reasons stack.
- Cycle lengths vary more, so predictions have less to work with.
- LH is harder to read, because it is often raised between surges rather than only at mid-cycle, which can make tests read positive across many consecutive days.
So the two signals that would normally cross-check each other both become less informative in the same population. That is not a reason to abandon either. It is a reason to lean on the signal that looks backward, after the fact, rather than the two that look forward.
What should I do when they conflict?
Nothing dramatic. A conflict is information, not an error to correct.
Three things that are reasonable:
- Log both, including the disagreement. A pattern of conflicts across several months is more informative than any single month, and it is something concrete to bring to a clinician.
- Don’t switch methods mid-cycle hoping one will settle it. They are answering different questions, so neither will.
- Look at after-the-fact signals rather than predictions when cycles are irregular.
If you are trying to conceive and the conflict is persistent, that is a reasonable thing to raise with a clinician rather than to keep resolving on your own.
What this means for tracking with Premom
Premom is an ovulation tracking application. It is not intended to diagnose, treat, cure, or prevent any disease, including PMOS. The information provided is for educational purposes and should not replace consultation with a healthcare provider.
Being plain about it: our predictions are built from cycle history, so they inherit the limitation described above. When your cycles vary substantially, a predicted window is a rougher estimate than it would be for someone with consistent cycles, and a test strip disagreeing with it is expected rather than a fault in either.
What logging genuinely does well here is hold both signals in one place over months, so a pattern becomes visible and you have something specific to show a clinician. That record is an input to a conversation. It is not an assessment, and it does not resolve which signal was right.
When to talk to a clinician
The information here is general and is not a basis for self-diagnosis. Speak with a clinician if:
- Your predictions and tests conflict month after month
- You are seeing surges without a progesterone rise afterwards across several cycles
- Your cycles are consistently long or vary widely from month to month
- You are trying to conceive and are uncertain whether you are ovulating
Frequently asked questions
No. A prediction estimates a fertile window from your previous cycle lengths; a test strip measures luteinizing hormone right now. They answer different questions, so they can disagree without either being faulty.
Neither, as a general answer, because they are not measuring the same thing. The more useful question is what you want to know. If it’s whether ovulation likely occurred, the more useful signal is a sustained rise in progesterone or urinary PdG after the expected ovulation date.
Switching does not resolve it. Any prediction built from cycle history assumes those lengths are reasonably stable, and where they vary substantially that assumption does not hold. This is true of any such prediction, including ours.
More testing does not adjudicate between two methods answering different questions. Logging both, including the disagreement, over several months is more informative, and it gives you something concrete to bring to a clinician.
More recorded cycles give a projection more to work with. Where cycle lengths vary substantially from month to month, though, the underlying limitation stays, because the variable part of the cycle is the part that sets the timing.
About PMOS
Polycystic ovary syndrome was renamed polyendocrine metabolic ovarian syndrome (PMOS) by international consensus published in The Lancet on 12 May 2026.3 A three-year transition runs to 2028, and both names remain in clinical use. Diagnostic criteria did not change, and an existing PCOS diagnosis remains valid.
Reported prevalence varies with the criteria applied; the 2023 International Evidence-Based Guideline reports 10–13% (DOI: 10.1210/clinem/dgad463).4
References
- Su HW, Yi YC, Wei TY, Chang TC, Cheng CM. Detection of ovulation, a review of currently available methods. Bioeng Transl Med. 2017;2(3):238–246. DOI 10.1002/btm2.10058 · PMID 29313033 · PMCID PMC5689497. CC BY 4.0.
- Strauss JF III, Barbieri RL, eds. Yen & Jaffe’s Reproductive Endocrinology: Physiology, Pathophysiology, and Clinical Management. 8th ed. Elsevier; 2019. ISBN 978-0-323-47912-7.
- Teede HJ, et al. The Lancet, 12 May 2026. DOI 10.1016/S0140-6736(26)00717-8 · PMID 42119588.
- Teede HJ, et al. Recommendations from the 2023 International Evidence-Based Guideline for the Assessment and Management of Polycystic Ovary Syndrome. J Clin Endocrinol Metab. 2023;108(10):2447–2469. DOI 10.1210/clinem/dgad463 · PMID 37580314.
