Let $f$ be an arithmetic function that counts the number of consecutive integers between $1$ and $n$ (inclusive) such that both integers are coprime to $n$. More formally,
$$ f(n) = \sum_{\substack{1 \leq t \leq n \\ (t,n)=1 \\ (t+1,n)=1}}1. $$
An immediate observation is that $f(n)=0$ when $n$ is even. After playing around with this function, I suspect that it is multiplicative. That is, if $(a,b)=1$, then $f(ab)=f(a)f(b)$. However, I am unable to prove this.
Is this function really multiplicative or are there counterexamples?
Also, what happens when we modify $f$ so that it counts the number of consecutive triplets that are all coprime to $n$?
Suppose $\gcd(m,n)=1$
Then let $(i,i+1)$ be a good pair for $m$ and let $(j,j+1)$ be a good pair for $n$. The Chinese Remainder theorem gives us a unique $k$ with $1≤k≤mn-1$ such that $$k\equiv i\pmod m\quad \&\quad k\equiv j \pmod n$$
Clearly we have $$k+1\equiv i+1\pmod m\quad \&\quad k+1\equiv j+1 \pmod n$$
So $(k,k+1)$ is a good pair for $mn$.
It remains to be shown that every good pair for $mn$ arises this way. But if $(k,k+1)$ is a good pair for $mn$ then remark that $tm<k<\left((t+1)m-1\right)$ for some $t$ so $(k\pmod m,(k+1)\pmod m)$ is a good pair for $m$ (and similarly for $n$) and we are done.