Printable Graph Paper

Probability Graph Paper

A vertical scale stretched so the cumulative normal distribution is a straight line, making non-normal data visible by eye.

210 × 297 mm
Download PDF
Download PDF

How to print Probability Graph Paper

  1. Pick the paper

    Choose Probability Graph Paper from the list. It opens with the usual size already set.

    Probability Graph Paper selected in the list of paper types
  2. Set it up

    Pick the paper size, the grid and the orientation. Customize holds colours, margins and more.

    Settings for Probability Graph Paper: paper size, grid and orientation
  3. Download or print

    Download the ready PDF, or print straight from the page at 100% scale.

    Probability Graph Paper preview with the Download PDF and Print buttons

A normality test you can see

The vertical axis is spaced by the inverse normal — the probit — so the S-shaped cumulative normal curve is pulled straight. Plot cumulative percentages of a normal sample and the points fall on a line.

What makes it useful is how the failures look. Points curving upward at both ends mean heavy tails; a bend in the middle means skew; a distinct kink often means two populations mixed together. A formal test gives you a p-value; this gives you a picture of what is wrong.

Reading mean and spread off it

Once the points are on a line, the distribution's parameters are just positions on it.

  • Checking normality — the fastest visual test there is.
  • Estimating the mean — read the value at the 50% line.
  • Estimating the standard deviation — half the distance between 16% and 84%.
  • Quality control — process capability studies against specification limits.
  • Grain size and particle analysis — sieve results plotted cumulatively.

Every sheet here is free, adjustable and prints at exact scale.

Questions

How do I read the standard deviation from the plot?

Read the values at the 16% and 84% lines. Those are one standard deviation either side of the mean, so half the distance between them is the standard deviation.

What does a curved plot mean?

That the data is not normal. Curving up at both ends means heavier tails than normal; a systematic bend one way means skew; a step or kink often means two populations mixed together.

Is this the same as a Q-Q plot?

The idea is the same. A Q-Q plot puts the normal quantiles on the axis directly; probability paper labels the axis in percentages instead, which is easier to plot by hand.