Concept Architecture
Concept
Theoretically, Finite Precision is the limitation of representing numerical values using a finite number of digits or bits within a computer system. It is a fundamental concept in numerical analysis and computer arithmetic because digital computers cannot represent most real numbers exactly. Finite precision gives rise to approximation errors during numerical computation and underpins the behaviour of floating-point arithmetic, rounding error and numerical stability.
Mathematically, finite precision constrains numerical representation to a discrete set of representable values determined by the number of bits allocated to the sign, exponent and significand (mantissa). Under standards such as IEEE 754 floating-point arithmetic, numbers are approximated by the nearest representable value, introducing quantisation and rounding errors whose magnitude depends on machine precision.
In practice, Finite Precision affects all computational methods used in health economics, including decision-analytic models, simulation, optimisation, regression analysis and probabilistic sensitivity analysis. Understanding finite precision is essential for interpreting numerical accuracy, selecting stable algorithms and ensuring reliable implementation of health economic models.
Purpose
Used to characterise the numerical limitations of digital computation, evaluate computational accuracy, understand approximation error and support reliable implementation of numerical algorithms in health economic modelling.
Mathematical Formulae
Primary Formula
Floating-Point Representation
x = (?1)? ? m ? 2?
where:
- s = sign bit
- m = significand (mantissa)
- e = exponent
Supporting Formulae
Machine Epsilon
� = smallest positive number such that
1 + � > 1
Relative Representation Error
Relative Error = |x ? fl(x)| � |x|
where:
- x = exact value
- fl(x) = finite-precision representation
Related Mathematical Methods
- Floating-point arithmetic
- Binary representation
- Machine precision
- Rounding error
- Truncation error
- Numerical stability
Example
The decimal number
0.1
cannot be represented exactly in binary floating-point arithmetic.
Instead, it is stored as the nearest representable binary approximation.
Although the stored value differs from 0.1 by only a very small amount, repeated calculations involving this approximation may accumulate numerical error. This illustrates the practical consequences of Finite Precision in computational modelling.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| ABS | =ABS((0.1+0.2)-0.3) | Demonstrate finite-precision representation error. |
| ROUND | =ROUND(A2,12) | Reduce visible effects of finite precision for reporting. |
| EXP | =EXP(A2) | Evaluate functions affected by floating-point precision. |
| LN | =LN(A2) | Assess numerical stability in logarithmic calculations. |
| IF | =IF(ABS(A2-B2)<1E-12,"Equivalent","Precision Difference") | Compare values within finite-precision tolerance. |
VBA (Optional)
Automate detection of floating-point precision issues, evaluate numerical tolerances and validate health economic models against finite-precision computational limits.
Sources
- IEEE. IEEE Standard for Floating-Point Arithmetic (IEEE 754).
- Goldberg D. What Every Computer Scientist Should Know About Floating-Point Arithmetic. ACM Computing Surveys.
- Higham NJ. Accuracy and Stability of Numerical Algorithms.
- Wilkinson JH. Rounding Errors in Algebraic Processes.
- Press WH, Teukolsky SA, Vetterling WT, Flannery BP. Numerical Recipes: The Art of Scientific Computing.
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation.
Related Concepts (7)
Library
Publications
1
Accuracy and Stability of Numerical Algorithms — Nicholas J. Higham, 2nd Edition ed., 2002 (Society for Industrial and Applied Mathematics (SIAM))
The standard reference on floating-point error, conditioning, stability and the numerical reliability of matrix, polynomial and linear-algebra algorithms.
BookView source →
Frequently Asked Questions (6)
What is finite precision?
The limitation of computer arithmetic arising from the use of a fixed number of digits to represent numerical values.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
What causes finite precision in computer arithmetic?
Finite precision is caused by the use of a fixed number of digits to represent numerical values, which limits how exactly a computer can store and calculate with numbers. Because only so many digits are available, values that need more must be approximated. This limitation arising from a fixed digit budget is what finite precision describes Because only so many digits are ever available, a computer necessarily approximates most values, which is why the small errors of finite precision are a permanent feature of numerical work.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
Why can a computer not represent every number exactly under finite precision?
A computer cannot represent every number exactly under finite precision because it uses a fixed number of digits, and many values, including some simple fractions and irrational numbers, would require more digits than are available. Those values must be rounded to fit, introducing small differences. This inability to hold arbitrary values exactly follows directly from the fixed number of digits that defines finite precision Because only so many digits are ever available, a computer necessarily approximates most values, which is why the small errors of finite precision are a permanent feature of numerical work.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
What limitation does finite precision impose?
Finite precision imposes the limitation that computer arithmetic can only represent numerical values to a fixed number of digits, so results carry small approximations rather than being exact. Calculations that would need more digits than are available are rounded to fit. This constraint on the exactness of stored and computed values is the limitation finite precision names Because only so many digits are ever available, a computer necessarily approximates most values, which is why the small errors of finite precision are a permanent feature of numerical work.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
Is finite precision a property of the computer or the problem?
Finite precision is a property of the computer's arithmetic, since it arises from the use of a fixed number of digits to represent numerical values rather than from the mathematical problem being solved. Any computation carried out with such fixed-digit arithmetic is subject to it. This makes finite precision a feature of how the machine represents numbers Because only so many digits are ever available, a computer necessarily approximates most values, which is why the small errors of finite precision are a permanent feature of numerical work.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
How does finite precision relate to floating-point arithmetic?
Finite precision is the limitation arising from representing numerical values with a fixed number of digits, while floating-point arithmetic is the method computers use to represent and calculate with real numbers under that limitation. Floating-point arithmetic is the scheme that operates within finite precision. The two are connected, since floating-point arithmetic is how a computer works with real numbers given the finite precision available to it Because only so many digits are ever available, a computer necessarily approximates most values, which is why the small errors of finite precision are a permanent feature of numerical work.
Source: Higham NJ. Accuracy and Stability of Numerical Algorithms. 2nd ed. SIAM; 2002. doi:10.1137/1.9780898718027.
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 2 Apr 2026
Content version: 1.0.0
Canonical Identity
- Persistent URI
- https://healtheconomics.wiki/concept/finite-precision
- Term code
- CS-NA-FPA-004
Stable URI · Machine-readable · Resolvable · CC BY 4.0