Torment: Tides of Numenera Interview

On the official Italian Torment: Tides of Numenera blog there's a brief Q&A with project director Kevin Saunders. Worry not if you don't understand Italian though, as the original exchange in English is also reported:
I really love the idea behind (Crises), but I'm curious to know how exactly are you guys planning to handle them? Mainly trough text - in choose-your-own-adventure fashion - or through (normal) gameplay? Will they incorporate puzzle-solving elements, instant death events and other challenges? What are the pros and the cons of Crises compared to a more (system-oriented) approach to non-lethal gameplay ( i.e. stealth/hacking in Deus Ex/Vampire: the Masquerade - Bloodlines)?

Crises occur in the framework of normal gameplay, but are separate from exploration or conversation gameplay, much like real-time exploration is separate from combat in some classic RPGs. For example, we expect for the interface and some of your options to change when you're in a Crisis (e.g., we may find that different game camera options are best for Crisis gameplay than for when you're exploring the world). They won't be choose-your-own-adventure style; text will be a part of them, but in the same way as it is throughout the rest of the game, with you potentially engaging in (limited duration) conversations and examining items in the environment. I think it's accurate to say that Crises will include puzzle-solving elements. We won't blindside players with unpredictable instant death, but you'll pay the consequences if you're careless.

The main reasons we are looking at this approach (as compared to the approach in the examples you give) have to do with compartmentalizing these experiences, which provides benefits such as:

  1. We can safely create the environments throughout all of the parts of the game that don't have Crises even while we're still iterating on Crisis design. So we can take the time necessary to flesh out our Crises and best understand what they require from their environments without putting the rest of environment creation on hold. This is especially important because we expect iteration on our 2D prerendered environments to be more challenging than for a 3D game.
  2. We can consider more extreme abilities and reactivity (i.e., choice and consequences).
  3. This approach allows us to handcraft the major encounters, simultaneously blending non-combat elements into them. We promised early that our encounters would be thoughtful and strategic, and Crises allow us to live up to that promise.

The primary potential disadvantage is that the Crises (and thus combat) will feel less seamless. This affects the pacing of the game in ways that could be either good or bad depending upon the specifics. But we do risk some loss of immersion.