Build a solution one choice at a time. Choose, recurse, undo. It's DFS on an implicit decision tree.
What Is It?
Backtracking explores all possible solutions by building them incrementally. At each step, you make a choice (e.g., include this element), recurse to make the next choice, then undo the choice (backtrack) and try a different option. Think of it as DFS on a tree of decisions. Each node is a partial solution, each edge is a choice, and leaves are complete solutions. You traverse this tree, collecting valid solutions. The "undo" step is what makes it backtracking and not just recursion. You literally reverse your last choice so you can try alternatives from the same state.