Mark Nau gave me this puzzle book by Raymond Smullyan. (Well, actually, he loaned it to me, and then I spilled tomatillo salsa on it - this tomatillo salsa stuff is the bomb, btw, you can get it at Pachanga, if there's one near you, the stuff is ambrosia - and then he said, "It's yours.") I'm enjoying it a great deal - I've mentioned before how much I like puzzles - and it's given me some thoughts about game design.
These puzzles aren't your standard chess puzzles. I don't think solving them is helping my chess game one bit. Smullyan will give you a fairly contrived chess position, and it's up to you to figure out how that position came to be - often asking questions such as, "What was the last move?" or "Has anybody promoted yet?" or "Can white castle?" Totally different from "White to play and mate in two."
Which just goes to show: you can take a game that has been strip-mined for content, turn it on its head, and get a wealth of new material out of it. Here's where my thoughts get fuzzy: I feel like I've seen just about every possible Zelda puzzle, for example. Is there some way to turn Zelda on its head and get a whole new set of puzzles out of it? Maybe you're creating the puzzles for somebody else...or maybe you have some control over time and you have to do the puzzles backwards...
In general, trying to think with this mindset could be a useful tool to have when you're asking yourself, "How do I take this genre that's been totally tapped out and make something interesting with it?"
There's a number of other straight ahead game design principles that Smullyan obeys: the puzzles are presented as part of a story; he acclimitizes you to the puzzles slowly, introducing concepts ("imaginary check", pieces promoting as something other than queen, using pawn captures to analyze the history of the game...) one by one.
Also, any puzzle book allows for nonlinearity - you can always skip ahead to the next puzzle if the last one was too hard.
Here's a thing, though: puzzle books always come with the solutions. A puzzle book that doesn't also have the answers is unheard of. Nobody sells the puzzle book and a separate strategy guide to the puzzle book! Why do puzzle videogames use this model? I asked Mark Nau (I really would have to invent him if he didn't exist) and he said, basically, precedence.
Talk about an antitode for shelf level events. I'm imagining a game where every time you fail a mission it asks you if you want a hint. The hints keep getting more specific until they give it away. Why not? Why the hell not?
My only nitpick about this book is that the chess games that you're doing this detective work on are ridiculous. How come these players who are so smart that they can dissect a chess game in minutes play so badly? It's a ridiculous complaint - the only way to make these puzzles interesting is if the players play strangely, so we need to grant Smullyan poetic license.