Work over the finite field . Reduce polynomials modulo , so their monomials are square-free. The binary Reed-Muller code consists of the evaluation vectors of polynomials of degree at most at all points of , in a fixed order. This is a linear code of length and dimension , for . Square-free monomials give independent evaluation functions, as follows inductively by splitting according to the last variable.
The minimum Hamming distance of a linear code equals its least nonzero Hamming weight. The result is
Here is the repetition code, while is the full binary vector space and has distance one. For the induction step, write , where and . Its evaluation vector is , the Reed-Muller bar-product recursion. If , its nonzero Hamming weight is at least , with the full-space endpoint treated as above. If , every position where contributes exactly one to the sum of the two weights; hence the weight is at least . The monomial is one exactly on points, attaining the bound. Degrees greater than give the same full code; the zero code at negative degree has no nonzero word whose weight could define a finite minimum.
The binary symmetric channel with crossover probability has channel matrix
Maximum-likelihood decoding chooses a codeword maximizing , while minimum-distance decoding chooses one minimizing the Hamming distance . For a block of length sent through a binary symmetric channel,
When , the final factor decreases strictly with the distance. Thus maximum-likelihood and minimum-distance decoding agree.
For the length-three repetition code , majority decoding is wrong exactly when at least two bits flip. Therefore