Yes, now that I’ve got an interpretating framework, I can dig the datamine.
Here’s what’s new :
1 - I found I need to make the encounters semi-isolated systems. This means codifying plot trees (Cf Pridwayn railroad campaing), “getting away”, and “getting there”.
2 - Regarding “getting there”, the use of trap tells in particular needs to be heavily codified. Not the qualitative signs themselves (I’ve distilled them all from here!), but the question whether there is, or whether there can be a clue, and who gives it, and things like illusion magic encounters or deceptive strategies.
3 - the encounter mechanic and tactic goals can be refined to infinity (NND, RPS, block or dodge, sweep, bash or thrust, etc.) They don’t need to be though. However tempting that is, and even though it’s “mechanical”, that’s diversity in colour and shades, and very easy to ad lib. Resource economy will benefit from conversions 1-1, with one big cost : turns, and one small change cost : options in a turn (not unlike the Lost Worlds duel booklets). This 1-1 ratio makes harm and narrative conditions a light brain load, too.
4 - I’ll focus on imitating the original material (HxH), include “strategy tropes” and ethical considerations (goals < motivations < values)
5 - and problem solving difficulty scaling, with points number 1 and 2 coming into account (hiding the box obviuously makes it harder to think out of it). Level design appears as a a combination of these factors along the plot tree. I don’t know why, but this seems like the easy part, like I’ve cracked this nut and all the players will have to do is eat it.
6 - trap investigation is still a bit vague. I leave a lot to players here… What I have is mostly a number of investigative steps for the most complex problems (distilled from corporate articles)
7 - and simple nice locpicking steps.
8 - and a collection of Perception moves (SoP etc.)
Now the hard, tedious work begins: datamining and meticulous redaction.