A bit-flip error channel maps a finite input vector to a finite output vector. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.
highlighted = computed this step
One fifth of zero shots flip
The declared bit-flip channel moves 1/5 of the 100 zero shots into the one bucket.
100⋅51=20
The rest remain in zero
The remaining share is 4/5, so 80 shots stay zero and 20 shots become one.