Friday, December 22, 2017

Scratch Paper Ideas

I'm tidying up notes from around my desk. This turns out to be my 50th post!


Costs associated with getting your way. Spend points from (a pool based on ability)?

No. Very story-gamey. But, losing points from a pool which is based on your ability scores *is* fine, so that idea could be used for something. Suppose that if you have a Int/Con/other ability core of 15 that you also have a pool of 15 points which can sustain damage, similarly to HP.

(well actually that's very finicky. Revisiting this one a day later: I don't think I like it, but it's fod for thought.)

Also, spending points from some pool to do something *can* be allowed in limited quantities as part of benefits earned from the skill system. But mostly I want skill system powers to be always-on.


Dice roll probabilities are difficult to estimate, especially for a pool of multiple dice (possibly of different sizes). THe gam can have sophisticated use of random chance, and it may internally be represented by combinations of die rolls, but player- and DM-facing views/interfaces should also display percent chances for failure/success, or for the various outcomes on a more complex type of risk.


Having one die roll determine whether or not you can do X in this situation doesn't work fo rhte fact that many ations are predictably either very hard or very easy.

Thus, most talents shall provide always-usable actions or knowledge (senses!)


*Actions* and *senses*. One is for affecting the world, and one is for filtering the information coming in and determining which of that information reaches the entity.

That's a useful paradigm to think in whether one is making a computer game or running a traditional game. What can this entity perceive and how does that shape its capabiilities? what can't it perceive and how does that present weak points?

Senses are typically things like sight, smell, touch, taste. magic sense, blind sense, tremor sense, and so on. These could take parameters indicating how far out they extend, e.g. vision 30, magic sense 50. Senses will otherwise binary: you have them or you don't.

What about senses like proprioception, the awareness of one's own body and movements in space? Well, no need for anything fancy right away, so we say that normally people have it. for someone who lacks it, a lack of it would be modeled by penalizing a character's attacks, balance on two feet, coordinated actions, amount of action points available, and so on.

Magic sense, blind sense, and other unusual senses can probably just be effective no matter where [...]

It makes sense for the parameter giving the range of the sense to indicate the maximum range at which that sense can be deployed with no errors. Then, calculations be made to detemine how far away the sense can be deployed with misdetections (false positive, e.g. seeing something not there, or false negative.)


Use plain probabilities. 0% to 100%. Store them as ratios of arbitrary-size integers internally, but round down (always!) to the nearest whole percent when *applying* or *displaying* them.

For ordinary tasks that we consider everyday but which could be difficult for others, don't merely assign a die roll chance to pass/fail. Set *ability score thresholds*, and from there set the degree of difficulty (it fmight be automatic or require a die roll.) Maybe two thieves can both pick the same types of locks, but one can go a little faster because of high Int. And so on.

e.g. Open Door [requires one hand, no Int requirement]
takes a certain amount of time based on Str AND may have chance to fail outright afte spending that amount of time
Strength 1-2: takes 10 AP
Strength 3: takes 8 AP
Strength 4: takes 6 AP
5: takes 5 AP
6: takes 4 AP
7+: takes 3 AP

We can also allow for thresholds which are higher than those an ordinary human would have. e.g. a large rock for which the minimum strength to move it at all is 25. IF there is room for a few people to stand near it, they can move it a little. IF ther es only room for one person he'll need super-human capabilities.

Or maybe he just needs to be non-human.


A fact is a binding of a name to a list of tuples List (Predicate, List Components). A predicate takes an entity and calculates True or False based on the entity's qualities (usually his senses, but not always.) If True is calculated, then the Components listed in the Predicate's tail are perceived by the entity which was passed to the predicate. These components are the salient perceptual details of a fact.

One easy way to model "components anyone can perceive" is for the predicate to be a function which always returns true no matter which entity can see it, e.g. the Clojure function (constantly true), which takes any arg(s) and always returns true.

The presence of an entity in a scene is a fact, and it would be detectable by the other entities which had the appropriate senses. But what senses are appropriate? Hm. Perhaps each entity needs to list which senss it's visible by. Obviously defaults could be given.

Thus the relationship of predicates to component lists is like the relation of predicates in a cond expression to their corresponding result expressions, except that

An example fact, in pseudo-Clojure.

(Fact "presence of mushrooms"
     {(#(> (mycology %) 10) "There are GREEN MUSHROOMS here. With your skills you could gather 4 lbs in 8 hours.")
     (#(> (mycology %) 20) "The mushrooms are healthy to eat. You could gather 6 lbs in 8 hours.")
     (#(> (mycology %) 40) "You know how to prepare these mushrooms into a healing stew. 12 lbs mushrooms and 1 hour time to prepare quart stew; one quart heals 1d3 HP. You could gather 8 lbs in 8 hrs.")})

Fact is a crappy name, but all of this is mostly crappy. I'm just trying to put ideas out on the page.

Also, the "presence of mushrooms" fact should be conditional on the entity having the sight sense. Otherwise they can't see the mushrooms. But if one with sufficient mycology were TOLD about the shrooms, they would recognize them for what they were. It's a tricky thing to adjudicate. This stuff is where computer modeling "knowledge of facts" is super hard and maybe not worth digging into. It's a rabbit hole. Fuck.

1 comment:

  1. As you specifically asked me my opinion, I'll write some thoughts.

    1) Haven't we learned our lessons about point-buy systems yet?

    2 to 4) These seem very vague thoughts. I don't think I'm part of the conversation.

    5) I dislike probabilities that determine what probabilities to roll. Seems counter-productive. Figure out a scheme that compares the strength to the task and settle on a consistent roll for each level of difficulty.

    6) I might be able to parse this language, but without context I'm not motivated to try. It is a lot of "needing to be made more human." See Steven Pinker's comments on this.

    On the whole, you're in your own head more than you're writing things to a reader. You'll get more attention when the post is less about you and more about them.