A good couple of the set $\{1,...,n\}$ is a couple of the type $(k,k+1)$ where $k,k+1\in \{1,...,n\}$. Two good couples $(k,k+1)$ and $(j,j+1)$ are disjoint iff they don't have any element in common.
In how many ways can I select $t$ disjoint good couples of $\{1,...,n\}$?
My attempt
Let's suppose we have $n$ balls $$\bullet \bullet \cdots \bullet \bullet$$ representing the numbers $\{1,...,n\}$. I just have to select the first element of each couple, so I have just to select $t$ elements properly. When I select the $i$-th ball I substitute it with a bar $|$. The rules are that there cannot be adjacent bars and the last ball cannot be selected. To create symmetry I can add an auxiliary ball at the start so that it can't be selected just like the last one $$\star \bullet \bullet \cdots \bullet \bullet $$ Basically this should create a bijection (through stars and bars) between the number of ways to select disjoint good couples and the number of compositions of $(n+1-t)$ in $(t+1)$ positive numbers. So according to this method the answer should be
$$\binom{n-t}{t}$$
Suppose we have $n$ identical balls and $t$ boxes, each of which can hold two balls placed side by side. Place $2$ balls in each of the $t$ boxes. That leaves $n - 2t$ balls. We have $n - t$ objects to arrange, $t$ boxes each containing two balls and $n - 2t$ other balls. Choose $t$ of the $n - t$ positions for the boxes. Now number the balls from left to right. The numbers on the balls in the $t$ boxes are the desired set of $t$ disjoint pairs of consecutive numbers. Hence, there are $$\binom{n - t}{t}$$ ways to select $t$ disjoint pairs of consecutive numbers from the set $[n] = \{1, 2, \ldots, n\}$.