18 min read

Accuracy vs Precision

Accuracy is closeness to the true value. Precision is how tightly repeated results cluster — they are not the same idea.

Two Different Ideas

Imagine throwing darts at a bullseye. Accuracy asks: are the darts near the centre (the true target)? Precision asks: are the darts tightly clustered together, even if that cluster is in the wrong place? A measurement can be precise but inaccurate (tight cluster, wrong spot), accurate but imprecise (around the centre but scattered), both, or neither.

Dartboard picture

$Accuracy$	$Precision$
$Near the bullseye (true value)$	$Shots close to each other$
$About correctness$	$About consistency / repeatability$
$One lucky good shot can look accurate$	$Needs several results to see the cluster$

Memorise these four target patterns: 1 High accuracy, high precision — tight cluster on the bullseye. 2 Low accuracy, high precision — tight cluster away from the centre (systematic bias). 3 High accuracy, low precision — results average near the centre but are widely scattered. 4 Low accuracy, low precision — scattered and not centred on the bullseye. Exam questions often show four dartboard pictures and ask which description matches which image. Name accuracy and precision separately for each picture before you pick an option.

Example 1 — Example — high precision, low accuracy

A poorly zeroed scale always reads about 0.5 kg too heavy. Repeated readings of the same bag: 3.51 kg, 3.50 kg, 3.51 kg. The readings agree \to high precision They are all wrong by ~ 0.5 kg \to low accuracy Fixing the zero error improves accuracy; the tight clustering was already there.

Example 2 — Example — high accuracy, low precision

Four people estimate a room length that is truly 4.0 m : 3.6 m, 4.3 m, 3.8 m, 4.2 m The average is near 4.0 m (roughly accurate as a group), but the estimates are scattered \to low precision .

Practice

Swipe or use the arrows to move between questions.

1 / 2

Shots are tightly clustered together, but far from the bullseye. Which description fits?

Which pattern is both high accuracy and high precision?