No black boxes. Every calculator on this site shows you what it's doing and how confident it is. If a number isn't trustworthy, we say so. If we don't have enough data, we tell you that too.
A polygenic model focused on OCA2 and HERC2 — the genes responsible for roughly 74% of eye-color variation in people of European descent.
We model the inheritance of B (brown), b (blue), and g (green) alleles across two loci, with brown dominant over green dominant over blue. Adding grandparents lets the model see recessive alleles that are hidden in heterozygous parents — which is why two brown-eyed parents can have a blue-eyed child about 6% of the time. Outside European ancestry, accuracy drops; we're working on extending the model.
Mid-parental height with sex adjustment, the same formula pediatricians use. We also factor in current growth-chart percentile if you provide it.
Boys: (mom + dad + 5)/2 inches. Girls: (mom + dad − 5)/2 inches. The ±2 inch range is real — about 80% of kids land within 4 inches of this prediction. Genetics explains roughly 80% of height; the rest is nutrition, sleep, illness, and chance.
Naegele's rule — last menstrual period plus 280 days, with cycle-length adjustment. The same number your OB uses.
Only about 4% of babies are actually born on their predicted due date. About 80% arrive within two weeks of it, which is what 'due date' really means — the middle of a window, not a deadline. First-time pregnancies tend to run longer; subsequent ones tend to be earlier.
A regression model using mom's pre-pregnancy BMI, weight gain, gestational age, baby's sex, and parity (first vs. subsequent pregnancy).
Birth weight prediction is more art than science. We capture the biggest factors but real outcomes vary by ±1 lb routinely. Things we can't see from a form — gestational diabetes, parental height, smoking, IUGR — add more spread. We're honest about that.
A two-axis model: a darkness score (from cumulative variants on TYR/OCA2/MC1R) and a separate red-hair axis (MC1R recessive).
Hair is polygenic — at least a dozen genes contribute. We can't see any of them from a form, so we infer the family's darkening tendency by comparing each parent's childhood color to their current color. The trajectory we show is the population average for your inputs; individual kids can shift differently.
Currently a demo dataset of popular names. Full SSA integration (every U.S. name given to ≥5 babies per year, 1880 to present) is on the roadmap.
The historical curve is exact — it's just a database lookup. The five-year forecast is a simple trend extrapolation; names rarely change direction sharply, but they can. The 'is this name due for a comeback?' label uses 100-year-cycle pattern matching against names like Eleanor, Oliver, and Theodore that re-emerged after a long dip.
We'd rather under-promise and let the science speak. The internet is full of baby calculators that present coin flips as certainty. Ours don't — and they're more useful for it.