Try it: Set the population parameters, click "Generate" to draw a sample, then "Fit" to estimate the model. Try small N (e.g., 20) vs large N (e.g., 500) — notice how the fitted curve (orange) tracks the population curve (teal dashed) more closely with more data. In the linear, quadratic, and cubic models, increase σ to see how noise obscures the signal; the logistic model has no σ, because the Bernoulli outcome carries its own variance.