Friday, December 22, 2017


In the ACKS repo, I'm working on a branch to change the way downtime works.  One of my issues with ACKS is that it doesn't really support open-table play well outside of low levels.  Domain XP slows down (or stops) level convergence within the party, and characters with domains tend to monopolize out-of-adventuring time and DM attention.  My intentions with these changes include:

  • Give everyone across the level range useful things to do during downtime.
  • Reduce overhead / level of detail / granularity in tracking downtime (two-week "downtime turns")
  • Create a structure / cycle of play around downtime, such that it happens in regular, predictable chunks alternating with adventures rather than "OK, Mr. Mage needs three months to make his item and we dare not go adventuring without him, so we're all gonna sit on our butts for three months."
    • As a side effect, reduce the variance of magic research by cutting it into multiple downtime actions and allowing partial progress
      • TODO: standardize monetary outlay between spell research and magic item creation
  • Add some mechanics which accelerate the convergence of low-level characters to the party mean (mentorship bonus)
  • Throw some bones to players who miss sessions (extra downtime actions)
  • Mix up the general proficiencies metagame.  Currently, a handful of general proficiencies (Healing, Alchemy, Bargaining, and Diplomacy/Intimidation initially, then Navigation and Riding in wilderness, then Military Strategy and Leadership when mass combat starts happening) typically dominate discourse and use in play in my groups.  Unsurprisingly, these are the general proficiencies that actually do useful things.  The trick, then, is adding really utile uses to some of the job-type general proficiencies.  So far, this has mostly been by increasing effective market class for goods related to the proficiency practiced, or by adding ability to find henchmen of certain classes (and possibly outside the typical henchman recruiting level range).
    • Relatedly, reduce the pain of operating in small markets and deprecate/distribute the Venturer / Varangian's core ability.
    • TODO Art - ???
    • TODO Labor, especially Mining - bonus to detect traps underground (probably only +2, smaller than dwarf bonus or Alertness), high-risk-high-reward downtime activity, ability to recruit dwarf henchmen.
    • TODO Profession, especially Lawyer - bonus to rolls on sentencing table
  • TODO: make sufficient quantities of carousing a downtime action, make the carousing table canon (that table was a regular source of fun in the first campaign)
  • TODO: rework mercenary hiring, rather like this.  Mercenary wages, henchman wages, and cost of living need some thought in open-table mode.
  • TODO: integration with Apocalypse World-style clock system, for easy running of living worlds.  I don't need to know a whole lot of details, just that this disgruntled NPC is going to try to come after the PCs in N sessions.  I'll fill in the rest when things get closer.
  • TODO: fix / redesign hijinks to provide adventure hooks and intel on running clocks.  Spying actually makes sense when people worth spying on are doing secret things in the background.
  • TODO: simplify domains.  Starting for now by taking this and working it into the Campaigns chapter, and by removing domain XP (mass combat XP can stay - that's combat.  XP from sacking domains can stay - that's treasure earned adventuring).  I have some vague notions for some Lords of Waterdeep-style mechanics, where domains and guilds provide you with anonymous, disposable "adventurer" resources that you can send on quests to delay clocks, but I'm not sure how to make it mesh with ACKS-as-it-is, nor how I really want it to look at the end of the day, so I'm going to put that one off.


Koewn said...

One thing I've done recently is to apply a summation calculation to magical research - so that instead of binary success/failure, there's just extra cost and time applied to the process based on the magical research throw - so a roll of 14+, for example, means a potion takes you 1,429 of base cost and 1.43 weeks, ignoring the chance to fail in exchange for upfront cost.

At 14+ the 85% success guarantee happens around the 8th roll; 99% around the 11th roll - so it's certainly advantageous; and it kinda approaches base costs at 4+ (588gp/.59 weeks, these numbers with with formula/sample)

The other thing it kinda showed me is that L0 apprentices leveling to Mage 1 via successful magical research is expensive, evidently, anywhere from 20x-4x expected costs at the high MRTs. I've modeled apprentices instead granting a +1/-1 on a successful throw to the base caster's chances via summation, and then split the XP accordingly by individual successes, the main pool going to the leader; and it 'feels' ok, at least.

The consistent summation process would grant some verisimilitude to projects one can stop/start; maybe with the requirement an apprentice has to watch the oven, so to speak; and it has the advantage of at least singing along with the default assumed costs of things, if not fully in tune.

I've bookmarked your fork; I'm glad someone's thought to do it.

John said...

Interesting! How did you get 1.43 / what summation did you do, exactly? I would expect 14+ to take something like 3x base cost and time (oh, divide that by 2 with 10+ as your baseline?).

Yeah, we've never really considered apprentices viable; our 5th-6th level casters have enough trouble getting any research done. I really like having apprentices add a bonus to the master; I think that would make them worth using.

I am explicitly not worried about preserving the continuity of projects / I like the idea of freely stopping, starting, and switching projects a lot. It's a poor simulation of real research work (because, like, mental cache effects), but v0v.

Koewn said...

Convergent series, evidently? Had I known I'd be using this stuff some 20 years later, for the silly elf game no less, I'd have paid more attention in my various college math classes.

((1-p)^k * p) * (f + f*k)

where p is probability, f is the factor (base cost, base time), k is the answer we're getting.

sum ((1-0.35)^k * 0.35) * (1 + 1*k), k=0..infinity

in Wolfram Alpha will give you the goods; .35 being the probability of success at 14+, solving for time (1 week base) here.

Being an unsubtle man, I brute force it in Perl when I want to generate a whole table.

John said...

That method makes sense, but it definitely doesn't converge to 1.43. The first ten terms sum to 2.68. I get around 2.86 with ten million terms, off by a factor of 2. Wolfram seems to agree, at*+0.35)+*+(1+%2B+1*k),+k%3D0..infinity ?

Koewn said...

Bah - sorry. The table I was looking at had the 'half time for formula' applied to it. (2.86/2==1.43), and I had meant to write 2.86 rather than 2.68. Severe posting failure :)

John said...

Oh, ok - that makes a lot of sense in the context of the apprentice discussion.