Degree to which repeated measurements or computations under unchanged conditions yield the same or similar results; commonly associated with low random error or small dispersion of outcomes.

Principle

Precision is characterized by the spread (variance, standard deviation) of repeated outcomes around their mean; high precision means low spread regardless of proximity to the true value.

Demonstration

Running the same simulation multiple times with the same deterministic code and inputs on a stable platform yields negligible variation in output metrics, indicating high computational precision; in experimental work, repeated instrument readings with narrow scatter show high precision.

Misapplication

Treating high precision as evidence of correctness; a measurement system can be highly precise yet biased, producing repeatable but systematically wrong results if calibration or model bias is neglected.

Consequence

High precision enables reproducibility, confident statistical estimation of averages and uncertainty reduction by repeated sampling; it allows distinguishing small systematic effects from random noise.

Reversal

Imprecision: large scatter among repeated observations that undermines reproducibility and makes it harder to detect systematic trends or validate models without increasing sample size.

Boundary

Concerns repeatability and dispersion under nominally identical conditions; does not imply closeness to truth (accuracy), and may be affected by uncontrolled changing conditions or hidden systematic influences.

Semantic Tension

Precision often competes conceptually with accuracy: one can have high precision and low accuracy (consistent bias) or low precision and high accuracy on average; both properties must be considered together.

Synthesis

Precision quantifies reproducibility via the spread of repeated results; it is necessary for reliable estimation and detecting systematic effects, but must be paired with accuracy and bias control to yield correct conclusions.