Visual Tools
Calculators
Tables
Mathematical Keyboard
Converters
Other Tools

How to Find a Normal Probability

Every normal distribution is a shifted and stretched copy of the standard normal. Convert each boundary to a z-score, which says how many standard deviations it sits from the mean, then read the area from the standard normal CDF Φ\Phi.

Recognize the task

A measurement is said to be normally distributed with a given mean and standard deviation, and the question asks for the chance of a range: "heights with μ=175\mu = 175 cm, σ=7\sigma = 7 cm", "scores N(70,102)N(70, 10^2), P(score>85)P(\text{score} > 85)". Test: the model is stated, or the data are bell-shaped and symmetric about the mean. Then X∼N(μ,σ2)X \sim N(\mu, \sigma^2).
The model is defined in the normal distribution lesson.

Two ways to do it

Standardize and use Φstandard
P(X≤x)=Φ ⁣(x−μσ)P(X \le x) = \Phi\!\left(\frac{x - \mu}{\sigma}\right)
The standardization Z=(X−μ)/σZ = (X - \mu)/\sigma turns any normal into N(0,1)N(0, 1); read Φ(z)\Phi(z) from a table or software.
The 68-95-99.7 rulewhole σ only
P(∣X−μ∣<kσ)≈0.68, 0.95, 0.997(k=1,2,3)P(|X - \mu| < k\sigma) \approx 0.68,\ 0.95,\ 0.997 \quad (k = 1, 2, 3)
A quick estimate when the boundaries sit exactly 11, 22 or 33 standard deviations from the mean.

Steps

StepOn heights N(175,72)N(175, 7^2), P(X<182)P(X < 182)
1Name μ\mu and σ\sigma. If the variance σ2\sigma^2 is given, take its square root first.μ=175\mu = 175, σ=7\sigma = 7
2Standardize each boundary. z=(x−μ)/σz = (x - \mu)/\sigma. A negative zz lies below the mean.z=182−1757=1z = \frac{182 - 175}{7} = 1
3Write the event as left-tail areas. Left tail: Φ(z)\Phi(z). Right tail: 1−Φ(z)1 - \Phi(z). Between: Φ(z2)−Φ(z1)\Phi(z_2) - \Phi(z_1). For negative zz use Φ(−z)=1−Φ(z)\Phi(-z) = 1 - \Phi(z) if the table only lists positive values.P(X<182)=Φ(1)P(X < 182) = \Phi(1)
4Look up Φ\Phi and combine. Use a table or software; round the final answer to 44 places.Φ(1)≈0.8413\Phi(1) \approx 0.8413

Worked examples

Example 1 · a right tail

1 − Φ
scores N(70,102)N(70, 10^2), P(X>85)P(X > 85)
σ=10\sigma = 10
z=85−7010=1.5z = \frac{85 - 70}{10} = 1.5
standardize
1−Φ(1.5)≈1−0.933191 - \Phi(1.5) \approx 1 - 0.93319
right tail
P(X>85)≈0.0668P(X > 85) \approx 0.0668

Example 2 · between two values

Φ − Φ
heights N(175,72)N(175, 7^2), P(168<X<189)P(168 < X < 189)
z1=168−1757=−1z_1 = \frac{168 - 175}{7} = -1, z2=189−1757=2z_2 = \frac{189 - 175}{7} = 2
two boundaries
Φ(2)−Φ(−1)≈0.97725−0.15866\Phi(2) - \Phi(-1) \approx 0.97725 - 0.15866
Φ(−1)=1−Φ(1)\Phi(-1) = 1 - \Phi(1)
P(168<X<189)≈0.8186P(168 < X < 189) \approx 0.8186

Example 3 · variance given

σ² → σ
X∼N(50,16)X \sim N(50, 16), P(X<44)P(X < 44)
σ2=16\sigma^2 = 16, so σ=4\sigma = 4
z=44−504=−1.5z = \frac{44 - 50}{4} = -1.5
below the mean
Φ(−1.5)=1−Φ(1.5)≈1−0.93319\Phi(-1.5) = 1 - \Phi(1.5) \approx 1 - 0.93319
symmetry
P(X<44)≈0.0668P(X < 44) \approx 0.0668
mirror image of the first example
Check your areas in the continuous distributions calculator.

See it happen

Shifting a normal curve by μ\mu and stretching it by σ\sigma does not change the share of area that lies within a given number of standard deviations of the center. So one table of the standard normal, indexed by zz, serves every normal distribution. Shade an interval and watch the area in the normal distribution visualizer.

Common mistakes

Dividing by the variance

✗ N(70,100)N(70, 100): z=85−70100=0.15z = \frac{85 - 70}{100} = 0.15
✓ σ=100=10\sigma = \sqrt{100} = 10: z=1.5z = 1.5

Using Φ for a right tail

✗ P(X>85)=Φ(1.5)≈0.9332P(X > 85) = \Phi(1.5) \approx 0.9332
✓ Φ\Phi is the area to the left: 1−Φ(1.5)≈0.06681 - \Phi(1.5) \approx 0.0668

Sign slip with negative z

✗ Φ(−1)=−0.8413\Phi(-1) = -0.8413
✓ areas are never negative: Φ(−1)=1−Φ(1)≈0.1587\Phi(-1) = 1 - \Phi(1) \approx 0.1587

Treating strict and non-strict ends differently

✗ P(X≤182)=P(X<182)+P(X=182)P(X \le 182) = P(X < 182) + P(X = 182) with P(X=182)>0P(X = 182) > 0
✓ for a continuous variable P(X=182)=0P(X = 182) = 0, so both are Φ(1)\Phi(1)