Book 3 - Page 214

Join the club

Book 3 - Page 214
Comic - Book 3 - Page 214
Recent posts... (See full thread)
DVL wrote:
Mathematical models also don't have to be some snapshot of the Absolute and Complete Truth, they just need to be precise within acceptable and usable tolerances.


Not disagreeing with anything else you said, quite the contrary in fact.

However, I wonder if perhaps the mathematical models that give us erfworld unit statistics (assuming that's how this all works) might actually be perfectly accurate. Real world physics is mind shatteringly complex with immeasurable numbers of variables changing in immeasurable numbers of ways at an immeasurable rate. This creates implicit uncertainty that is the reason why we settle for "within acceptable and usable tolerances" with our mathematical models and statistics. But Erfworld's physics might be very much simplified making the assumption of inherent uncertainty unnecessary. Instead of uncountable variables changing in uncountable ways at an untrackable rate, it might be just a half dozen variables changing in one or two ways at a rate that can be observed without special tools.

And we've seen how powerful Fate is. And Fate is but one of three major axes at the heart of erfworld's mechanics, the other two being Erf and Numbers. So if Numbers has the kind of power Fate does... then might not it be possible for the statistics and the mathematical models employed by mathemancy and luckamancy and the like to be completely accurate?

But perhaps they aren't. After all if Fate's plans can be interfered with then perhaps Numbers' interpretations can be wrong.
I'm agreeing with you for the most part, it's just that folks like Count_to_Ten can't get their heads around the idea that Attack and Defense don't have to have any literal reality into themselves because the math may be dealing in abstract statistical fudge-making. It's like confusing the tally in card-counting with the actual game itself.

Time is a lot the same way, since many would argue that time isn't real, it's actually a function of displacement and not the other way around. And therefore it is wholly abstract, but is nonetheless an important idea for human purposes.

"Attack" basically amounts to, "military power is really important for human purposes in Erf." Also there's an obnoxious computer spitting out these numbers for some reason. Some say the Titans dunnit, but they always say that.

Absolutely, it does seem like Mathamancers basically have a hotline to some kind of universal computer whose entire job it is is to just do these really involved computations for some reason. It's also been implied in the Jack-Maggie-Marie linkup that there is some overlap between how Predictions are made and Mathamancy.

Instead of an estimate of likelihood, such as "50% chance," it gives definite certainties about what will happen true or false. But such is the nature of the power that it rarely can do anything more than give a few very specific details with very little actual context. Maybe Prediction eats up too much processing power.
DVL wrote:
I'm agreeing with you for the most part, it's just that folks like Count_to_Ten can't get their heads around the idea that Attack and Defense don't have to have any literal reality into themselves because the math may be dealing in abstract statistical fudge-making. It's like confusing the tally in card-counting with the actual game itself.

...snip...


Ah, I see.

I was just trying to clarify the possibility that the statistics could be perfectly accurate without disagreeing with what else you've said. But you actually seem to be offering a third interpretation that is still consistent with the underlying concept based on misconceptions about probability and statistical models. Jolly good. :-)
Personally, the prescriptive/descriptive thing doesn't bother me particularly much, perhaps because of too much time spent playing D&D and the like. Weird, arbitrary, and sometimes exploitable mechanics aren't really a new thing. But OOTS has already gone over that better than I could, so instead of driving the point into the ground let me set up and then knock down a straw man, while patiently awaiting the next comic.

If combat results are driven by underlying stats in combination with a random number generator (or, if Fate is involved, a pseudo-random one) you might wonder why units go through the motions of a complicated fight, maneuvering for advantage and so on. (Are they being forced to play out a scene to justify the pre-ordained outcome?) Well, this puts me in mind of a rule from the RPG Over The Edge -- players were expected to be creative when narrating their PCs' combat moves, and if the description was boring or phoned in, they actually got a penalty to the action. I can imagine a similar meta-mechanic at work in Erfworld: units have stats and those stats determine their odds in battle -- but the units also have consciousness and some degree of agency, and if they don't treat the battle as though every detail mattered (as we would in Stupidworld) then they get a penalty, and that could make the difference between life and death. So the net effect would be to produce a battle that looks realistic (or at least dramatic) even if it's driven by simple rules and stats.

(Disclaimer: not saying that that is the mechanism. Might be something else; might be just that things work that way period. But it's not totally crazy -- compare with Parson having to do the rounds of an empty city to get the productivity bonus, for example.)

[Edited: typo]
Mut, one thing that has been made clear is that Erfworld does not have a random number generator. It was stated as much in comic directly by Clay, and the idea has been reinforced by Rob .
As confusing as it is, peoples actions can somehow just get described as a dice roll. This is impossible in real life, but somehow possible in Erfwolrd. SO if you try to hit a ball with a bat, you just try to hit the ball. Based on how well you did, Erfworld puts a number behind it, like a 1 or a 6. It does that rather then roll an invisible dice that influences your outcome. Your choice becomes the dice role. So Erfworld units are all somehow simulated dice.

But despite that there is still other things that influence the outcome. Somehow your poor choice that was just a 1 will be more powerfull just because there is a cheif warlord in the hex, or you are stacked with 7 other units. The chief warlord doesn't actually have to give orders to make you better. You are just somehow better. And you can make a choice that is a 6, and erfworld could treat it as 1 by having something unlucky happen because you are on the bad side of a luckamancy spell.

And Clay never did describe what sort of dice are modeled. In a d20 system there is just as much chance to roll a 1 as a 20. This can be chaotic. If you roll 3d6 though the result will tend to be curved towards the middle. Something with a curve models human behaviour a lot better, as most of the time people will be consistent in their behaviour. So I would say actions that people take in Erfworld are somehow described as a roll in some sort of a curved dice system.
Sir Dr D wrote:
Mut, one thing that has been made clear is that Erfworld does not have a random number generator. [...]

The RNG thing wasn't really the point of my post (in as much as it had a point) -- it was more intended to explore a way in which agency and personality could matter mechanically at a micro level in a world in which units have concrete, well-defined stats. If you like, go back and read it assuming a purely deterministic combat system in which chance plays no role at all. Such a system would still raise the question of "why would units behave like every little thing matters in a world with simple combat mechanics", and the notion of a penalty for not acting with gusto is still one possible answer to it.
Whether or not RNG exists in Erf is irrelevant to the character's perceptions of it, and how they might act because of it. And mechanics aren't as important as people are making them out to be.

Randomness is mostly used in Stupidworld to cover up uncertainty. We're uncertain about how a thing works because we've abstracted away parts of it, and so we say it's random (yes, things get a little different on the quantum scale. I know). If you do a survey where 90% of people answer a question one way, the difference between people randomly answering one way 9/10 times and something more complex under the surface causing a higher level observer who doesn't understand the underlying mechanics to see it as a 90% random chance is moot to said higher level observer.

Given that people are surprised when they crit, a Stupidworlder in an equivalent situation might call that RNG. Given that people have a tiny chance in thousands to hit someone with a brick off-turn, that's something that we would call RNG. Now, it might be that these are entirely causal, and there are just so many things that cause it to work or not that it can be approximated at 1/4000. But to an Erfworlder, there is functionally no difference between that and a 1/4000 mechanical roll.

Whether or not Erfworld is deterministic or not, the perception of the mechanics can be described through RNG - this is essentially describing an abstraction of Erf-reality. This leads us to the question of whether Mathamancy is doing. Is it describing the actual mechanical rules, where the outcome of a situation is described based on the dice roll that occurs like a GM noting how much your roll suceeded or failed by, and noting how the event plays out? Or is it tapping into some kind of predictive model based on uncertainty (see - Bayes), which lets you simulate what a probability might be approximated as, given what you do and don't know? It's worth mentioning that this seems to be how the bracer actually works. It requires priors, it changes based on information you give it. And it can be frustratingly complex. And people treat it as a Mathamancer.

If it matters, we don't have any 100% chance way of knowing what the underlying mechanics are: are they prescriptive, telling people what works and doesn't, or are they descriptive, a way of wrapping your head around how the physics of the world works, but always being at a higher level than you could be thinking about it at. But it doesn't matter, because this is a character driven story, and anything the characters don't know is uncertainty, and uncertainty becomes RNG. It only matters if someone becomes a hell of a lot closer to omnipotence (which sure, might actually happen at some point, who knows..but right now it's moot).

I wish I could have put more effort into this, but it needs to end so I'll close off my rant with a useful reference for Tools to look into. Look near the start of the Toolbox for a audo file called Audio - Rob Balder Podcast "Erfworld is Not a Game". Rob in it says "Erfworld is not a game. It is a game-like universe". He talks about how if there is a game system, it's the metaphorical shadow on the back of the cave. It's not the primary focus, which is and always has been on the characters - if the characters care about the game system (if there is one) then what matters is their and their action's relationship to the mechanics of their world.
wih wrote:
Randomness is mostly used in Stupidworld to cover up uncertainty. We're uncertain about how a thing works because we've abstracted away parts of it, and so we say it's random (yes, things get a little different on the quantum scale. I know). If you do a survey where 90% of people answer a question one way, the difference between people randomly answering one way 9/10 times and something more complex under the surface causing a higher level observer who doesn't understand the underlying mechanics to see it as a 90% random chance is moot to said higher level observer.

Pretty much this. If you were to give a survey, how you frame the question matters a lot. Since the survey won't necessarily reflect a person's motivation for answering. If the person you're asking feels they're being put on the spot, then they'll much prefer to give a lie that they think reflects on them better. Or the survey can be constructed in such a way as to cover up much finer demographic details.
The "mechanics" are there but they aren't noticeable if all you're looking at is the end results.

Too much of Erfworld seems to be about overlooked mechanical stuff that belies a ridiculous amount of complexity under the cutesy cartoony exterior or a slick strategy UI's. One of the lines that demonstrates this consistent theme is, "Loyalty is more complex than you know."

I may be paraphrasing but Maggie used this to explain her motivations. She's technically Loyal in an abstract sense to GK, but she's more motivated by her guild membership in the Thinkamancer cabal and her love of Parson. Which is more casually significant than saying, "Her Loyalty is a 15." Managing Stanley is more like a job to her than a cherished life goal or some fundamental part of her identity. This isn't surprising, she derives a lot of pride in her craft as a Thinkamancer, as casters are wont to, and Parson is the one individual who shows he cares about her personally, if not romantically.

You can abstract this out to saying, "She has a 29% chance of subverting this sort of order from Stanley based on other Thinkamancers placed in a similar general situation." But this is a stupid attitude to take when the audience knows her motivations already. We know why she floofed Stanley in the first place to make Parson the Chief Warlord of the side.
Hell, Maggie outright said that Loyalty is "more complex" when explaining Loyalty to Thinkamancy, and ability to lie to Stanley about Thinkamancy secrets. Loyalty isn't just a stat; if it's a stat, there are multiple of them (a la Factions), it's just an abstraction, or both.
Nasty. I like the touch of Stanley's head being on a pike in the banner. Appropriate given that he was a piker. Dying at the end of a pike would be fitting.