Skip to content

Error Terms

The Prime Number Theorem states that

number-theory book chatgpt

Measuring the Accuracy of Prime Approximations

The Prime Number Theorem states that

π(x)∼li⁡(x), \pi(x)\sim \operatorname{li}(x),

or equivalently,

π(x)∼xlog⁡x. \pi(x)\sim \frac{x}{\log x}.

These asymptotic formulas describe the main growth of the prime counting function. However, they do not explain how large the difference between the approximation and the true value may be.

To study the precision of prime estimates, one introduces error terms.

Basic Error Functions

The two standard error expressions are

E1(x)=π(x)−xlog⁡x E_1(x)=\pi(x)-\frac{x}{\log x}

and

E2(x)=π(x)−li⁡(x). E_2(x)=\pi(x)-\operatorname{li}(x).

The Prime Number Theorem implies only that

E1(x)=o(xlog⁡x) E_1(x)=o\left(\frac{x}{\log x}\right)

and

E2(x)=o(li⁡(x)). E_2(x)=o(\operatorname{li}(x)).

Thus the error becomes small relative to the main term, but this does not specify its exact magnitude.

Understanding the true size of these errors is one of the deepest problems in analytic number theory.

Error in the Chebyshev Function

It is often technically easier to study the Chebyshev function

ψ(x)=∑pk≤xlog⁡p. \psi(x) = \sum_{p^k\leq x}\log p.

The Prime Number Theorem becomes

ψ(x)∼x. \psi(x)\sim x.

The associated error term is

ψ(x)−x. \psi(x)-x.

This formulation interacts naturally with the zeta function because logarithmic derivatives of Euler products produce sums weighted by the von Mangoldt function.

Complex Zeros and Oscillation

The irregularity of prime distribution is controlled by the zeros of the Riemann zeta function

ζ(s). \zeta(s).

The explicit formulas of analytic number theory express prime-counting functions as sums involving these zeros.

Very roughly, one obtains expansions resembling

ψ(x)=x−∑ρxρρ+⋯ , \psi(x) = x - \sum_\rho \frac{x^\rho}{\rho} +\cdots,

where the sum runs over nontrivial zeros

ρ=β+iγ \rho=\beta+i\gamma

of the zeta function.

Each zero contributes an oscillating term. Thus fluctuations in prime distribution arise from interference among the zeros.

Zero-Free Regions

Hadamard and de la Vallée Poussin proved that

ζ(s)≠0 \zeta(s)\neq0

on the line

Re⁡(s)=1. \operatorname{Re}(s)=1.

This fact alone yields the estimate

ψ(x)=x+O(xe−clog⁡x) \psi(x) = x + O\left( xe^{-c\sqrt{\log x}} \right)

for some positive constant cc.

Consequently,

π(x)=li⁡(x)+O(xe−clog⁡x). \pi(x) = \operatorname{li}(x) + O\left( xe^{-c\sqrt{\log x}} \right).

Although much smaller than xx, this error term still decays relatively slowly.

The Riemann Hypothesis

The most famous conjecture in analytic number theory is the Riemann Hypothesis, which asserts that every nontrivial zero satisfies

Re⁡(ρ)=12. \operatorname{Re}(\rho)=\frac12.

If true, this would imply the much stronger estimate

ψ(x)=x+O(x(log⁡x)2). \psi(x) = x + O\left( \sqrt{x}(\log x)^2 \right).

Equivalently,

π(x)=li⁡(x)+O(xlog⁡x). \pi(x) = \operatorname{li}(x) + O(\sqrt{x}\log x).

Thus the Riemann Hypothesis predicts that the fluctuations in prime distribution are extremely small compared with the main term.

Sign Changes

One might expect

$$ \operatorname{li}(x)

\pi(x) $$

always, since numerical computations support this for enormous ranges of xx.

However, entity[“people”,“John Edensor Littlewood”,“British mathematician”] proved that

π(x)−li⁡(x) \pi(x)-\operatorname{li}(x)

changes sign infinitely many times.

Therefore the error oscillates between positive and negative values.

These oscillations reflect the influence of zeta zeros in the explicit formulas.

Omega Results

Upper bounds alone do not describe the true size of fluctuations. Lower-bound oscillation results are also important.

Statements of the form

E(x)=Ω(g(x)) E(x)=\Omega(g(x))

mean that the error is at least as large as g(x)g(x) infinitely often.

For example, one can show that

ψ(x)−x=Ω±(xlog⁡log⁡log⁡x), \psi(x)-x = \Omega_\pm(\sqrt{x}\log\log\log x),

assuming the Riemann Hypothesis.

Such results demonstrate that prime distribution necessarily contains substantial irregularity.

Error Terms in Short Intervals

Another major question concerns primes in intervals

[x,x+h]. [x,x+h].

The Prime Number Theorem predicts approximately

hlog⁡x \frac{h}{\log x}

primes in such an interval.

Determining how small hh can be while retaining reliable asymptotic estimates is a difficult problem closely related to zero distributions and exponential sum estimates.

Importance

Error terms measure the fine structure of prime distribution. The main term

xlog⁡x \frac{x}{\log x}

describes only average behavior. The error term captures deviations from this average.

Much of analytic number theory focuses on improving these estimates. The subject connects:

  • zeros of the zeta function,
  • oscillatory phenomena,
  • exponential sums,
  • spectral theory,
  • random matrix models,
  • arithmetic statistics.

The study of error terms therefore lies at the heart of modern prime number theory.