In how many ways can you pick five books from a shelve with twelve books, such that no two books you pick are consecutive?
This is a problem that I have encountered in several different forms ("In how many ways can you paint five steps of a twelve step staircase, without any two consecutive steps being painted?", etc.) but the idea is the same. Here is how I solved it:
Consider a shelve with $13$ books, and then calculate the number of ways you arrange $5$ pairs and $3$ single books on that shelve. The answer to that question is ${8 \choose 5} = 56$. This question is equivalent to the original, because the pairs prevent you from choosing two consecutive books, and a shelve of $13$ books rather than $12$ is considered because otherwise, it would not be possible to choose the last book on the shelve.
This approach has its downside: writing a formal proof of this is not easy, as you can see. I tried to explain this proof to a friend of mine in order to give him the intuition ${8 \choose 5}$ is the right answer (which is usually easier than a formal proof) but I already had a hard time doing that.
My question is: Is there a more rigorous proof this is the right answer? Not only would that help me explain my answer, but it would probably allow me to use that technique to a broader range of problems.
Imagine the books are numbered $1$ to $12$. Let $a_1,a_2,a_3,a_4,a_5$ be the numbers of the chosen books, in increasing order.
Consider the numbers $a_1,a_2-1, a_3-2,a_4-3, a_5-4$. This is an increasing sequence taken from the numbers $1$ to $12-4$.
Conversely, given any increasing sequence $x_1,x_2,\dots,x_5$ taken from the numbers $1$ to $8$, we obtain an allowed book choice by choosing books $x_1, x_2+1, x_3+2, x_4+3, x_5+4$.
Thus there are as many allowed choices as there are ways to choose $5$ numbers from $8$.
One could use more formal language. In effect, we have an explicit bijection between the set of allowed book choices and the collection of $5$-subsets of $\{1,2,\dots,8\}$.
Of course there is nothing special about $12$ and $5$, the same reasoning works in general.