Loaf weights from Brightwater Bakery's line cluster around their average and thin out evenly on both sides: close to a normal distribution. Once you estimate its mean and SD from the data, scipy.stats.norm answers questions about it:
stats.norm.cdf(x, m, s): the probability of a value below x.
stats.norm.sf(x, m, s): the probability above x (1 minus the cdf).
stats.norm.ppf(q, m, s): the value with a share q below it, the inverse of the cdf.
stats.norm.pdf(x, m, s): the curve's height at x. It is not a probability.
The probability of a range is a difference of two cdfs. Without m and s, scipy uses the standard normal (mean 0, SD 1), which works with z-scores.
A model is only as good as its fit: compare what it predicts with what the data shows.
A probability below 790 g, and the weight with 2 percent below it.