The best hint systems behave like a good teacher: they answer the question you actually have.
Most games treat a hint as a small defeat. You press the button, the answer appears, and the interface politely suggests you keep it to yourself. That design assumes the player wanted a solution. Often the player wanted something much smaller: to know whether their current plan is dead or alive.
A good hint behaves like a conversation. It responds to where the player is stuck rather than to the puzzle as a whole. If you are stuck at the first rung, the useful answer is about the opening shape. If you are stuck three rungs in, the useful answer is about the branch you chose. The two answers are different, and a single spoiler cannot serve both.
That is why tiered help feels so different from a reveal. The first tier can answer a narrow question without touching the rest of the route. The second tier can point at a region. Only the last tier is a spoiler. Used well, the tiers let a player stay inside the puzzle while getting unstuck, which is the whole point of help in a leisure activity.
There is also a social benefit. When help is graded, asking for it stops being embarrassing. Players compare which tier they needed the way they compare scores. That comparison is honest in a way a completion count is not.
I judge a puzzle's help layer by one test: can I take a hint and still feel like I solved the puzzle? On a ladder, the answer is yes when the nudge is small enough, which is why the graded hints are worth studying as a design. If you want to see the difference for yourself, read the hints on a puzzle you have already finished and ask whether each tier gave away more or less than you needed.
The conversation framing also fixes something in the player. It turns help from a verdict into a question you asked and an answer you received.