How the Guess Plausibility Score Works

What is the GPS?

The Guess Plausibility Score (GPS) shows how statistically realistic each of your birth guesses was, based on real-world birth data from the CDC and ACOG. Instead of just saying "you were wrong," GPS tells you how plausible your guess was compared to what actually happens in births every year.

The GPS Buckets

Label	GPS Range	Meaning
Very Plausible	>= 0.75	Well within typical birth ranges
Plausible	0.50-0.74	Within normal variation
Unlikely	0.25-0.49	On the edges of typical ranges
Wild Guess	< 0.25	Well outside typical birth ranges

How GPS Is Calculated

BabyGuessr matches each attribute to a model that reflects its real-world shape. Weight uses a log-normal model and percentile. Length and due-date offset use percentiles from skewed distributions. Time of day uses an hourly birth-probability distribution. Accuracy is always the primary driver of your score, and plausibility fine-tunes the top of each category rather than overriding a close guess.

Percentile-Based GPS

percentile = cumulative probability at the guess

GPS = 1 - (|percentile - 0.50| / 0.50)

Accuracy comes first: how close your guess is to the actual result drives most of your points
Plausibility adjusts only the top portion of each category, so a spot-on guess stays richly rewarded even for an unusual birth
A guess at the distribution's median earns the full plausibility bonus for length and due-date offset
Time of day compares the guessed hour's birth probability with the most common hour

Real-World Data Behind Each Attribute

Birth Weight

Source: CDC National Vital Statistics System (NVSS), "Births: Final Data for 2023"
Model: Log-normal newborn birth-weight distribution
Median: ~3,300 g (7 lb 4 oz)
Distribution: Right-skewed; converted to a percentile
GPS basis: Distance of that percentile from the 50th percentile

Birth Length

Model: Smoothed newborn LMS (Lambda-Mu-Sigma) approximation
Median: 19.8 in
Distribution: Right-skewed; converted to a percentile
GPS basis: Distance of that percentile from the 50th percentile

Due-Date Offset (Days Early/Late)

Model: Smoothed CDC/NCHS-inspired gestational-age distribution
Pattern: More births before the due date, with a longer post-term tail
Peak: The days shortly before the due date
GPS basis: Cumulative-distribution percentile distance from the median

Time of Day

Source: CDC NVSS natality microdata (time-of-birth distribution)
Pattern: Daytime peak due to scheduled deliveries
Model: Smoothed hourly probability distribution
GPS basis: Guessed hour probability compared with the peak hour
Note: Overnight births remain plausible and are not treated as impossible

Why These Distributions?

Birth outcomes do not all follow the same bell curve. Birth weight has a natural lower bound and a right skew. Newborn length is also skewed, so BabyGuessr uses an LMS approximation before finding its percentile. Birth timing is asymmetric: births cluster around 39 to 40 weeks, often before the due date, then taper into a longer post-term tail. Hourly births also follow a real delivery pattern rather than clustering around a made-up central hour. These models are smoothed approximations of published CDC/NCHS and clinical reference patterns.

Why GPS Makes Scoring Fair

GPS means that guessing something impossible (for example, a birth weight of 15 lbs) does not just cost you points - it earns a Wild Guess label that reflects how far outside real-world norms the guess was. Conversely, if a baby genuinely arrives at an unusual time or weight, GPS acknowledges that and your guess is still evaluated against real distributions, not penalized for the baby's individual outcome.