How Bloom calculates conception probability
Bloom uses a Bayesian probability model based on the Wilcox et al. 1995 fertility study to calculate daily conception probability. The base probability peaks at 33% on ovulation day and tapers to 10% at 5 days before ovulation. This base score is then adjusted by fertility indicators the user logs:
- Egg-white cervical mucus: +30% boost
- Positive OPK (LH surge): +20-25% boost
- BBT thermal shift detected: -50% (ovulation already passed)
- Irregular cycle history: wider fertile window applied
Ovulation is confirmed using CUSUM (Cumulative Sum) change-point detection on basal body temperature data. The algorithm looks for a sustained thermal shift of 0.15-0.6°F, scoring candidates by shift magnitude, post-shift consistency, and pre-shift variability.
At the end of each cycle, Bloom auto-generates a conception report that classifies each intercourse entry as optimal (days -2 to 0 from ovulation), good (days -5 to -3), or low-probability, giving users clear feedback on their timing.