A couple of further thoughts:
* This was a big session for Golin from the point of view of PC build. His Will improved to 4, and he opened three new skills - Lore Master, Manipulator and Persuader. This all resulted from the social-intensive action of the past few sessions. He also improved his Cook to 3.
* @Manbearcat and I have talked in the past about the adventure-building rules. As I posted in the OP of this thread, I find them very good for their focus on fictional situation and mechanical obstacles. But compared to (say) 4e D&D, I find them a bit trial-and-error in terms of getting a sense for how easy or punishing an adventure will be.
But what has become clear is that adventuring in town, with no need to journey to or from the adventure site, is much more forgiving than having to travel. Which the rulebook says; but I think it might have been helpful to have another few sentences or a paragraph explaining this a bit more - I think one reason for this is the splitting of the rules across the Scholar's Guide and the Lore Master's Manual, so the former doesn't explicitly include Journeys in its discussion of adventure difficulty, while the latter's Journey rules don't link back to that discussion either.
As I said I think our next session will involve delving beneath Megloss's house. I would then expect the next town phase to be a Respite. And then I think it must be time for another journey, eg back to the Tower of Stars which remains unexplored by the PCs to a significant extent. There are also the NPC bandits they're interested in, and Korvin has an enemy (Nob H the Bandit Lord) and Golin has two of them (Ebenezer the Erudite, and his rival in the Forgotten Temple Complex also called Golin) who haven't figured yet. Plus there is the rumour form last session about how Vaccin may have known Golin's parents (I imagine Golin's father must have been named Gilim).
If the players remain interested, I think there's enough stuff there to support several more sessions at least.
4e's is "better" than TB's for a few system architecture reasons I'd say:
* Less parameters governing the difficulty of the throughline of play in 4e than in TB.
* The Milestone mechanic is a negative feedback loop which ameliorates the hardship of "pressing on" within the Adventuring Day paradigm.
* Torchbearer's attrition model is a continuous, layered positive feedback loop which continues its work beyond each singular loop of play (Journey > Adventure > Camp > Town) whereas 4e's is overwhelmingly siloed with little in the way of mechanical knock-on effects (unless the Disease Track is employed in particular ways).
With TB, you've got a host of converging elements beyond the Adventure Design parameters (# of Problem Areas based on Adv size + Obstacle difficulties and #s divided into tiers, Rest areas, Proximity to Town, Challenge Types, among other pieces of clear and excellent direction):
* Journey phase as multifaceted, variable attrition element.
* The incentive structures inherent to advancement (folks are going to make Tests that they aren't good at and this is going to yield Conditions/Twists that diversely complicate the model of play).
* The interplay between Grind Clock + Light Clock + "Must Secure Checks to Camp" + the complex decision-tree that Team PC must navigate around the question of "when do we need to camp and can we afford to make a Survey test for Camp amenities/Events Roll etc + this area is Dangerous...do we want to push on to a less Dangerous site", etc, etc. This, by itself, is so deeply complex, that it makes trying to accurately model and predict play in Adventure Design deeply difficult if not impossible.
* The importance of Gear and Tools and Supplies in the course of play.
* The role that Twists play in significantly perturbing (difficulty, length, the overall attrition model, the downstream role in impacting subsequent players' decision-trees) every single Phase of play.
* The decisions players will face and the variance with which one subset of players will make (vs what another subset might make) when it comes to managing Order of Might interactions/decisions and the same for Precedence.
* As mentioned before, the overall positive feedback loop of play is such a powerful parameter to all of this. Its powerful holistically and its also powerful discretely (my Resources sucks...oh boy that is going to be a problem loading out and Paying Debs in Town Phase).
So I guess my position on this is:
4e's model is just profoundly less complex than TB's so its inputs into "The Adventuring Day" are going to inherently be more reliable (GM perspective) whereas TB's are so multivariate with so many consequential short and long-term feedback loops that the best you can hope for is what they've provided. And with that in consideration, what they have provided as absolutely excellent in my opinion. Wonderful design that works quite well. When I contrast it with other systems that are considerably less complicated that offer both lesser and worse instruction at "the game layer" (in terms of reliability and durability)...let us just say that Luke and Thor should win alllllllllllllllll the design rewards and be heralded for their work!