5x

Deviation 60 vs. 70 —
just 10 points apart.
But the "rarity" differs 5-fold.

A deviation value isn't a "smartness score."
It's actually your position on a normal distribution.

StatPlay Columns What Is a Deviation Value?

"High deviation = smart" — does that sound about right?
But surprisingly few people can explain what a deviation value actually means.
In this article, you'll grasp the truth behind deviation values in 3 minutes.
Start by playing with this graph.

01

// Hands-on Simulator

Drag the sliders freely, or pick a preset from the table on the right. Everything stays in sync.

Deviation
Top %
1 in ... people
▶ Adjust with sliders
70
55
15
▶ Rarity presets
DevTop %Rarity
750.62%1-2 per grade
702.28%#1 in class of 40
656.68%Top 2-3 in class
6015.87%Top 1/6
5530.85%Slightly above avg
5050.00%Dead average
4569.15%Slightly below avg
4084.13%Bottom ~16%
02

// What a Deviation Value Really Is

The "deviation" you just played with — its formula is this simple:

Deviation = 50 + 10 × (Your Score − Mean) ÷ Std Dev

It's a straightforward scale transform: re-center the mean to 50 and set the standard deviation (spread) to 10.
In other words, "a number that lets you compare your position across any test, using the same ruler" — that's a deviation value.

03

// 3-Question Quiz — Check Your Understanding

Q1. Does a deviation of 80 mean you scored 80 on the test?

A deviation value is not the raw score — it represents your "position within the group." Even a test with a mean of 30 can produce a deviation of 80.

Q2. Does a deviation of 50 equal the average score?

By definition, someone who scores exactly the mean gets a deviation of 50. This is baked into the formula.

Q3. You got a deviation of 60 on two different tests. Same ability level?

If the test populations (who took the test) differ, a deviation of 60 means completely different things. An elite-school mock exam and a nationwide exam aren't comparable.
04

// The Pitfall: When Deviation Values Lie

Deviation values are calculated under the assumption that scores follow a normal distribution.
But in reality, test results rarely form a perfect bell curve.
Drag the skewness slider to the right and feel how much the "top %" diverges for the same deviation value.

65
1.0
Deviation says top % %
vs
Actual top % %
Gap: points

// KEY TAKEAWAY

FAQ

// Frequently Asked Questions

Deviation = 50 + 10 × (Your Score − Mean) ÷ Std Dev

First, compute the mean and standard deviation from all scores. Then plug your score in. At the mean you get 50; one standard deviation above the mean gives 60.

A deviation of 50 means "exactly the group average." No matter what the raw test score is, if it matches the mean, the deviation is 50.

A deviation of 60 is roughly the top 15.87% (about 1 in 6). At 65 it's top 6.68%, at 70 it's top 2.28%. Each 10-point increase causes rarity to jump dramatically.

No. Deviation is a relative position assuming a normal distribution; rank is an absolute ordering from actual data. When scores aren't normally distributed, the rank predicted by deviation can differ significantly from the real rank. Try the "Pitfall" section above to see this in action.

Mathematically, yes. A deviation of 100 means "5 standard deviations above the mean" — extremely rare, but possible with small samples or outliers.

Generally, no. Deviation measures your position within the specific group that took that test. If the test populations differ, a deviation of 60 on one test may represent a completely different ability level than 60 on another.

Behind every deviation value is a "normal distribution."

Want to understand the bell curve you just saw on a deeper level?
StatPlay's interactive materials let you learn statistics by touching.