Binomial approximation

Not to be confused with Binomial distribution § Normal approximation.

The binomial approximation is useful for approximately calculating powers of sums of 1 and a small number x. It states that

$(1+x)^{\alpha }\approx 1+\alpha x.$

It is valid when $|x|<1$ and $|\alpha x|\ll 1$ where $x$ and $\alpha$ may be real or complex numbers.

The benefit of this approximation is that $\alpha$ is converted from an exponent to a multiplicative factor. This can greatly simplify mathematical expressions (as in the example below) and is a common tool in physics.

The approximation can be proven several ways, and is closely related to the binomial theorem. By Bernoulli's inequality, the left-hand side of the approximation is greater than or equal to the right-hand side whenever $x>-1$ and $\alpha \geq 1$.

Derivations

Using linear approximation

The function

$f(x)=(1+x)^{\alpha }$

is a smooth function for x near 0. Thus, standard linear approximation tools from calculus apply: one has

$f'(x)=\alpha (1+x)^{\alpha -1}$

and so

$f'(0)=\alpha .$

Thus

$f(x)\approx f(0)+f'(0)(x-0)=1+\alpha x.$

Using Taylor Series

The function

$f(x)=(1+x)^{\alpha }$

where $x$ and $\alpha$ may be real or complex can be expressed as a Taylor Series about the point zero.

${\begin{aligned}f(x)&=\sum _{n=0}^{\infty }{\frac {f^{(n)}(0)}{n!}}x^{n}\\f(x)&=f(0)+f'(0)x+{\frac {1}{2}}f''(0)x^{2}+{\frac {1}{6}}f'''(0)x^{3}+{\frac {1}{24}}f^{(4)}(0)x^{4}+\cdots \\(1+x)^{\alpha }&=1+\alpha x+{\frac {1}{2}}\alpha (\alpha -1)x^{2}+{\frac {1}{6}}\alpha (\alpha -1)(\alpha -2)x^{3}+{\frac {1}{24}}\alpha (\alpha -1)(\alpha -2)(\alpha -3)x^{4}+\cdots \end{aligned}}$

If $|x|<1$ and $|\alpha x|$ ≪ $1$, then the terms in the series become progressively smaller and it can be truncated to

$(1+x)^{\alpha }\approx 1+\alpha x$.

This result from the binomial approximation can always be improved by keeping additional terms from the Taylor Series above. This is especially important when $|\alpha x|$ starts to approach one, or when evaluating a more complex expression where the first two terms in the Taylor Series cancel (see example).

Sometimes it is wrongly claimed that $|x|$ ≪ $1$ is a sufficient condition for the binomial approximation. A simple counterexample is to let $x=10^{-6}$ and $\alpha =10^{7}$. In this case $(1+x)^{\alpha }>22,000$ but the binomial approximation yields $1+\alpha x=11$. For small $|x|$ but large $|\alpha x|$, a better approximation is:

$(1+x)^{\alpha }\approx e^{\alpha x.}$

Examples

Example simplification

Consider the following expression where $a$ and $b$ are real but $a$ ≫ $b$.

${\frac {1}{\sqrt {a+b}}}-{\frac {1}{\sqrt {a-b}}}$

The mathematical form for the binomial approximation can be recovered by factoring out the large term $a$ and recalling that a square root is the same as a power of one half.

${\begin{aligned}{\frac {1}{\sqrt {a+b}}}-{\frac {1}{\sqrt {a-b}}}&={\frac {1}{\sqrt {a}}}{\Big (}{\big (}1+{\frac {b}{a}}{\big )}^{-1/2}-{\big (}1-{\frac {b}{a}}{\big )}^{-1/2}{\Big )}\\&\approx {\frac {1}{\sqrt {a}}}{\Big (}{\big (}1+{\big (}-{\frac {1}{2}}{\big )}{\frac {b}{a}}{\big )}-{\big (}1-{\big (}-{\frac {1}{2}}{\big )}{\frac {b}{a}}{\big )}{\Big )}\\&\approx {\frac {1}{\sqrt {a}}}{\Big (}1-{\frac {b}{2a}}-1-{\frac {b}{2a}}{\Big )}\\&\approx -{\frac {b}{a{\sqrt {a}}}}\end{aligned}}$

Evidently the expression is linear in $b$ when $a$ ≫ $b$ which is otherwise not obvious from the original expression.

Example keeping the quadratic term

Further information: Binomial series

Consider the expression:

$(1+\epsilon )^{n}-(1-\epsilon )^{-n}$

where $|\epsilon |<1$ and $|n\epsilon |$ ≪ $1$. If only the linear term from the binomial approximation is kept $(1+x)^{\alpha }\approx 1+\alpha x$ then the expression unhelpfully simplifies to zero

${\begin{aligned}(1+\epsilon )^{n}-(1-\epsilon )^{-n}&\approx (1+n\epsilon )-(1-(-n)\epsilon )\\&\approx (1+n\epsilon )-(1+n\epsilon )\\&\approx 0\end{aligned}}$.

While the expression is small, it is not exactly zero. It is possible to extract a nonzero approximate solution by keeping the quadratic term in the Taylor Series, i.e. $(1+x)^{\alpha }\approx 1+\alpha x+{\frac {1}{2}}\alpha (\alpha -1)x^{2}$ so now,

${\begin{aligned}(1+\epsilon )^{n}-(1-\epsilon )^{-n}&\approx (1+n\epsilon +{\frac {1}{2}}n(n-1)\epsilon ^{2})-(1+(-n)(-\epsilon )+{\frac {1}{2}}(-n)(-n-1)(-\epsilon )^{2})\\&\approx (1+n\epsilon +{\frac {1}{2}}n(n-1)\epsilon ^{2})-(1+n\epsilon +{\frac {1}{2}}n(n+1)\epsilon ^{2})\\&\approx {\frac {1}{2}}n(n-1)\epsilon ^{2}-{\frac {1}{2}}n(n+1)\epsilon ^{2}\\&\approx {\frac {1}{2}}n\epsilon ^{2}((n-1)-(n+1))\\&\approx -n\epsilon ^{2}\end{aligned}}$

This result is quadratic in $\epsilon$ which is why it did not appear when only the linear in terms in $\epsilon$ were kept.

Magnitude of error

The numerical value of the error of the binomial approximation for selected values of $x$ and $\alpha$. The error scales approximately with $1.5{\sqrt {\alpha x}}$.

An epidemiological modeler on Medium analyzed overprediction by the binomial approximation in the case of $x<0$. She found it scaled approximately with the square root of $\alpha x$ and produced the error lookup table included at right.