You know how to find the game. You can heighten it. You know when to rest. That's Level 3 work. Here's Level 5.
The first thing to understand is that "game" isn't one thing — it's a taxonomy. Behavioral games (a character tic), emotional games (inappropriate feeling), verbal games (speech pattern), physical games (body-based pattern), status games (power dynamics), mapping games (treating X as Y), scenic games (the space generates the pattern). Each type heightens differently, rests differently, and resolves differently. Knowing which type you're playing lets you make more precise choices.
The second thing: games don't just get bigger. They evolve. A game can pivot (shift to a new expression of the same underlying pattern), invert (the mechanism reverses), combine with another game (two patterns merge into something new), or transcend (the mechanical pattern reveals a genuine emotional truth underneath). Game evolution is what separates a clever scene from a moving one.
The third thing: sometimes you need to break the pattern. The rule of three — two iterations establish, the third subverts — is the simplest structure. But the discipline is knowing WHEN to break. Break too early and the audience feels cheated. Break for cleverness rather than truth and it reads as abandonment. The best breaks reveal something about the character that the pattern alone couldn't.
The fourth thing: games travel. The analogous scene transplants a game into a completely different context — different characters, different world, same underlying logic. This is the Harold's second-beat engine. The key is abstraction level: not "the doctor who lies" but "a trusted professional who betrays trust." Abstract enough to generate fresh scenes; specific enough that the audience feels the connection.
Finally, "if this is true, what else is true?" remains the master question at every level. It's just that at the mastery level, you're asking it across scenes, across beats, across the whole show — not just within a single moment.