CDF

(Evaluated via standard normal integration)

Deep Breakdown

What it tells me (Situation Awareness): $\Phi(Z)$ is the CDF of the standard normal distribution $N(0,1)$ — it has no closed form but is tabulated and available in all statistical software. For any Normal($\mu, \sigma^2$), probabilities are computed by standardizing: $P(X \leq x) = \Phi\left(\frac{x-\mu}{\sigma}\right)$. This is the most-used single function in classical statistics.

Flowchart

Real-world use

Where it is used

Computing z-test p-values, confidence interval bounds, probability of exceeding a threshold, and any Normal probability query. Foundation of hypothesis testing and quality control (Six Sigma).

Problem it solves

Converts a Normal probability query into a table lookup or scipy call — the integral $\Phi(z) = \int_{-\infty}^z \phi(t) dt$ has no closed form but is well-approximated.

How it solves the problem

Uses precomputed tables or approximations like the error function: $\Phi(z) = \frac{1}{2}\left[1 + \text{erf}\left(\frac{z}{\sqrt{2}}\right)\right]$.

Example in a business

Z-test for A/B conversion rates: compute z-statistic, then $p = 2(1 - \Phi(|z|))$ for two-tailed test. If $p < 0.05$, reject null.

The formula, piece by piece

$\Phi(Z) = P(Z \leq z)$ where $Z \sim N(0,1)$

• No algebraic closed form.
• In terms of error function: $\Phi(z) = \frac{1}{2}\left[1 + \text{erf}\left(\frac{z}{\sqrt{2}}\right)\right]$.
• Quantile (inverse): $\Phi^{-1}(p)$ = norm.ppf(p) in Python.

The formula in words

"$\Phi(Z)$ is the probability that a standard normal random variable is at or below $Z$." Key values: $\Phi(1.96) = 0.975$, $\Phi(-1.96) = 0.025$, $\Phi(0) = 0.5$.

Critical Concepts

• $\Phi(-z) = 1 - \Phi(z)$ (symmetry of standard normal).
• 95% CI uses $\Phi^{-1}(0.975) = 1.96$.
• Python: from scipy.stats import norm; norm.cdf(z).
• R: pnorm(z).

Step-by-step calculation (tiny example)

Step 1: Revenue ~N(10000, 2000²), query P(Revenue < 13000)

Step 2: Standardize $Z = (13000 - 10000)/2000 = 1.5$

Step 3: Lookup $P(X < 13000) = \Phi(1.5) \approx 0.9332$

Step 4: Interpret 93.3% of days have revenue below $13,000.

Code examples

Beginner example (Python)

from scipy.stats import norm
mu, sigma = 10_000, 2_000
x_threshold = 13_000
z = (x_threshold - mu) / sigma
p_below = norm.cdf(z)
print(f"Z = {z:.2f}")
print(f"P(revenue < {x_threshold:,}) = {p_below:.4f} ({p_below*100:.1f}%)")

Beginner example (R)

mu <- 10000; sigma <- 2000; x_threshold <- 13000
z <- (x_threshold - mu) / sigma
p_below <- pnorm(z)
cat(sprintf("Z = %.2f | P(X < %d) = %.4f\n", z, x_threshold, p_below))

Professional example (Python)

import numpy as np
from scipy.stats import norm

def z_test_two_proportion( conversions_a: int, n_a: int, conversions_b: int, n_b: int, alpha: float = 0.05 ) -> dict[str, float | bool]:
    """Two-proportion z-test for A/B conversion rate comparison. Uses Φ(Z) for p-value computation. Args: conversions_a: Number of conversions in control. n_a: Total users in control. conversions_b: Number of conversions in treatment. n_b: Total users in treatment. alpha: Significance level. Returns: Dict with z-statistic, p-value, and decision. """
    p_a = conversions_a / n_a
    p_b = conversions_b / n_b
    p_pooled = (conversions_a + conversions_b) / (n_a + n_b)
    se_pooled = np.sqrt(p_pooled * (1 - p_pooled) * (1/n_a + 1/n_b))
    z_stat = (p_b - p_a) / se_pooled
    p_value = 2 * (1 - norm.cdf(abs(z_stat)))
    return { "p_a": p_a, "p_b": p_b, "z_stat": z_stat, "p_value": p_value, "significant": p_value < alpha, "relative_lift": (p_b - p_a) / p_a, }
result = z_test_two_proportion(100, 2000, 130, 2000)
print("A/B Test Z-test result:")
for key, val in result.items():
    print(f" {key}: {round(val, 4) if isinstance(val, float) else val}")

Source: domain knowledge

#probability-statistics#statistical-operations-and-random-variables#common-continuous-distributions