In this weeks' Legends and Lore column on the official D&D website, RPG Group Manager for the D&D Research & Development, Mike Mearls, stated:

"Setting aside mechanics, I think you can boil D&D down to three basic activities: exploration, roleplay, and combat."

While each of those elements certainly plays a strong role, what about problem solving? Puzzles and riddles have been staples of D&D since its inception. You don't fight them, you don't interact with them in a meaningful way, and you don't explore them. You are challenged by them, and you cannot use your combat abilities to overcome them.

Some may argue that they fall under exploration -- but exploration cannot be the catchall category for anything that doesn't seem to fit the other two. Sure, a puzzle is something you encounter while exploring a dungeon, but so is a dragon. Should a battle with a dragon fall under exploration? Of course not. Then why should answering a dragon's riddle be lumped into that category?

People like lists of three. It's a good number. Two is too few, and four feels like an incomplete list of five. Three is the magic number. Mike Mearls probably thought of three basic activities and ran with them -- it wouldn't surprise me if "problem solving," or some variant of the term, simply slipped his mind and would have been included, had he remembered.