Friday, June 4, 2010

Engineering II

I am posting this second version of my Civilization Posts, because for me this is what I'm thinking on.  I have been giving a lot of thought to the one I ought to be writing, "Theology," but it will have to wait for awhile.  I run the blog, it does not run me.

Here's the thing.  I was raised the son of an engineer who became very disappointed when he son decided not to go into engineering, but instead into writing and other Bohemian pursuits, such as acting and journalism.  But as it happened he did manage to instill into me certain characteristics that, clearly, continue to define the manner in which I approach a problem ... such as, how does one establish a trade system, or how does one define the qualities of an NPC.

There is an old joke, that goes as follows:

A mathematician, a philosopher and an engineer were asked, what is the sum of 2+2? 

The mathematician replied that the answer was '4', but specifically it could be defined as '4.000' or however many zeros you wished to add to establish the accuracy of the reply.

The philosopher answered that we could not be certain of any particular definition, since reality is defined by perception, but we could reasonably judge within a common framework that the answer might be given as more than three and less than five.

The engineer answered, "What do you want it to be?"

When I told this joke to my father, he did not laugh.  Instead he answered, "Damn straight."  I remember we had a discussion afterwards in which he explained that when he was in university (he went to the Colorado School of Mines), those who were flunking out of engineering courses were not those who 'couldn't handle the math,' but rather those who just didn't get the fundamental principle that engineering is not about assessing, it is about doing.

With regards to the disaster in the Gulf, I have spoken to a few engineers who assure me that a) the technology in existence operates like it does because it is what the money wanted built; and b) the solution is straight forward and practical - but it does involve a second hole and it will take months.  And finally, that the second hole could be drilled beforehand, as a precaution, for every deep water platform in existence, but that no one wants to spend that kind of money.

Money is the universal limitation on engineering.  I am told that everyone in the field knows this.

Well, okay.  I make a D&D world, so I don't have to worry about money, and I don't want this to devolve into a long discussion about social political and technological recent events - so please make the comments relevant to D&D, at least obliquely.  What I want to talk about is the inaccuracy of my world, and why it doesn't matter.

I did touch upon this with my last post about peasants and so on ... but reading through the link on the last post, about the sociology of political divisions, I am struck by my approach vis a vis towards producing a solution to a world-design problem.

Let's take something not so simple, that I have not yet solved: given a hex size of 20 miles across, with an approximate area of 309 sq.m., how many monsters of a given type ought to exist in the hex?  This is to say, IF the party is going to 'clear out a hex,' as suggested in several places in the original DMG and Player's Handbook, how many creatures will the party have to fight until they are a) mostly eliminated; or b) completely eliminated?

Now give me an answer that applies to all types of hexes, from tundra to jungle, that takes into account the increase in animal size with regards to open-vegetation environments, as well as animals which are not large enough to be considered 'monsters.'  At what point is the hex 'clear'?  When there are no dangerous creatures larger than a dog?  Are giant centipedes, throat leeches, ear seekers or rot grub to be cleared, or not?  Can you answer?  I mean, without pulling the number from your ass?

Fact is, you can't.  No one can ... not that I've seen so far.  The answer ought to lie in the biomass of a zone - for which I can find total tonnage/hectare - but how precisely is this biomass deconstructed so as to give numbers for elephants, sphinx, dopplegangers, ochre jellies, ixitxachitl and dragons?  Ideas, anyone?

The general point of this post is to emphasize that a portrayal that accurately gives answers to any of the questions above is a matter of entirely no relevance whatsoever.  An correct number for dragons?  Huh?  Define correct.

It is exactly that which justifies the rectally derived numbers that virtually no DM, anywhere, ever bothers to write down for any part of their world.  Why bother?  Why not just say, when the party has to clear the hex, "Uh, sure, there's a dragon and a keep that has six ogres, two roving bands of brigands and a few hundred giant rats."  There, done, problem solved, please go away now.

And yes, this is engineering, by definition.  Problem solved.  It is rather piss-poor engineering, something like dropping two boards over a chasm and telling the driver that as long as he keeps his car wheels on the boards, no problem.  My father tells of how he did exactly this over a Colorado gorge back in the 1950s, and it obviously worked, he lived long enough to make much of my life unpleasant.  But even he wouldn't recommend it as a standard.  It would be better to build a bridge.

Where it comes to building bridges to solve problems in your D&D world, it helps to remember that there are a wide range of solutions, ranging from the two boards to complicated archways designed to handle all manner of traffic.  The principle to be followed is this: what does it do, and what does it allow?

The above example of what is in a hex, dragons and ogres and what not, answers the question, and lets the DM get on to other things.  I personally don't see the benefit of that.  I'm not anxious to 'get it over with' and move on.  Instead, I want to make THIS the interesting thing, the thing that draws the party in and compels them to burrow around and find every last fucking thing until they KNOW the hex is clear.  I want them to feel like they have accomplished something outlandishly difficult, so that when they are sitting by their hearth in their home, they FEEL like they've goddamn owned their land.

So I want the structure of my full hex to be better thought out ... hell, I want it to be randomized to the point where there are things in that hex that I, as DM, don't want to be there!  I want to roll up a lich and shudder to myself, thinking, uh oh.  Be damned if the party is ready for it.  Or if I'm ready for it, for that matter.  And when I find out that the lich has a vampire buddy, I want to be thinking, oh shit, oh shit, oh shit ...

Obviously, I also want a table that spits out the result that the only relevant thing in the hex is a set of caves populated by only kobalds.  I don't want the table to be necessarily brutal.  But I want the possibility of brutality, and when the dice comes up that way, I want to point at the table and say, "Guys, I hate to tell you this, but the Gods don't like you."

Okay, that's what I want a system to do.  What do I want it to allow?

A nice thing about an overall system is the continuity it provides.  Many DMs don't think it's true, but a party can tell when they're being jerked around by an irrationally constructed world, and it gets exhausting.  Oh, look, we thought we would be safe here in this peasant village where we came to rest, but there's five wraiths feeding off the peasants.  Oh, look, I've spent months building up my fortress, and now its being attacked on the day its finished by a huge army that has conveniently appeared on the horizon.  Oh, look, we thought we cleared out that dungeon but a whole group of demons and devils have popped out of the ether and made it their new home.  Oh, look, it doesn't matter what we do, everything we've created is being torn down for the sake of story.  Oh, yay.

It is comforting to the party to know that if they cut down or burn down or blast away everything that lives or moves in the adjacent hex to their own that the number of attacking creatures would actually go down and stay down, and not ramp up on the whim of the DM.  It is comforting to the party to know that if they deviate from their route to avoid the dark woods, that they won't be attacked anyway - you know, because this hex was presumably cleaned out by someone, at some point in the past.  It would be nice to think that since every civilized hex is occupied by various lords of the manor, that the lords of the manor would do something to kill the nasty horrible creatures living in their own domain, so they couldn't attack random travellers on their way through.  You'd think that would be something the local lords would want, you know, to facilitate trade and such.

I want my parties to have the reassurance of that.  That going from wilderness to civilization is the sort of thing that makes them feel safer - that falling prey to a nasty creature living in the sewers of a city is something that doesn't happen every time they drop into town to buy supplies.

So I'm thinking that an engineered system that took account of these things would help provide a stronger structure to my world, and create a balance for the players so that they knew that if they did this, or went there, they could expect such-and-such.  As opposed to a DM who's bored or whimsical or prejudiced in favor of one kind of monster (Another otyugh? For shit's sake!), who forever plays with the results to satisfy their constant need for self-importance.

(What am I saying?  I create systems to satisfy my constant need for self-importance.  Bad Alexis!  Bad!)

In any event, the long and the short of it is engineering some kind of system that works and provides a valuable result.  Engineering as a technology is more than the sum of the mathematics that goes into the construction of a bridge or the development of a methodology for the removal of oil, or what have you.  It is a mindset, wherein the end result is the issue, and not necessarily the minor details.  This has occasionally had some disastrous effects where it comes to building things that one would hope to remain standing, or applying efforts to extract things that one would hope not to leak out all over the environment.  Such considerations are necessary for the engineer who doesn't want to kill people and such.  But engineering itself is not defined by competancy.  It is defined by adequacy.  A bridge that remains standing without falling down is the minimum degree of adequacy desired; just as an effort to model something in order to determine how much money the project will need is, equally, not made inadequate by its degree of inaccuracy.  It is accurate enough.  It serves the purpose for which it was designed.

That is the only measure.


Carl said...


I now have a good idea of your design parameters, and you have a few requirements defined along with a few possibilities for how those requirements can be fulfilled.

All we need to do now is break those requirements down into atomic statements, create a technical specification, build a mock-up of the system (usually a storyboard) and do gap analysis. Once that's done, we can add in missing requirements and create a prototype. Then we'll continue requirements analysis, technical design specification development and prototype revision until we get a functioning system.

You nailed one side of the engineering project triangle: money. I refer to it as "resources" since money is a medium of exchange and you can't program a computer or build a bridge directly with dollar bills. Resources includes the budget, equipment and people.

There are two more sides to the triangle: time and quality. You only get to set two of the three.

Symeon Kokolas said...

Would it be feasible to build up a table of residence, similar to your table of references? Presumably, monster type depends on climate, available prey, available predators. The initial pass would be to generate base prey animals, everything from rabbits to mosquitoes based on the available geography, climate, and vegetation. The next pass would add predators compatible with the environment, at a density determined (at least partly) by available prey. A third pass could be used for monsters, if they are to be treated separately from predators. This would be reasonable since intelligent monsters have some differing needs/desires from base predators. These raw values could then be modified by human populations nearby, where humans compete for prey animals or other resources and seek out/kill monsters to improve their neighborhood. More passes could be performed if you wanted finer control over levels of the biological web, but the distinction would be arbitrary for most MM creatures.
I would expect this to produce fairly realistic distributions after a little trial and error. The results could then be compiled into tables based on geography, climate, vegetation, and human population for quick and easy rolling. The difficulty in compiling a table would depend partly on the results from the simulation and partly on the degrees of freedom under each influence.
It might also be possible to build a tree instead of a table. The first roll (or rolls) might determine prey animal(s), and the next roll(s) would be drawn from compatible predators of those prey animals, and so on until the web is filled.
Some of the top-end predators or monsters would have ranges exceeding a single hex. Especially, any good-sized winged carnivore would pose a threat to (and draw food from) a huge range. Having a dragon resident in any given hex is extremely unlikely, but having one use that hex as an occasional hunting ground is much more likely.

Hopefully you guys can pull something useful out of that mess.

Menace 3 Society said...

I would break the problem down in terms of ecology, making the following assumptions to simplify things:

1. but for the players' involvement/influence, each hex is at a sort of dynamic ecological equilibrium, with constant populations of everything over time.

2. everything has to eat/live off something else.

Since you (presumably) know the type of biome in the hex, you can use that with principles 1 and 2 and some real-world statistics to start figuring things out.

For example, let's start with a forest that has no large herbivores other than deer. Complete forest seems to be able to sustain up to 64 deer per square mile; mature deer have 1-3 young per foaling and mature in about 2 years, which means we should expect something on the order of 1/3 dying per year to keep a sustainable level. So that means, in toto, we have about 6500 deer per hex per year for other inhabitants to live off of. Figuring one deer yields about 60 pounds of meat, or 41600 Calories--figure about half again if you include excess fat and bone marrow maybe? So, about 60000 Calories per deer: about enough to feed one person for 25 days, or a grizzly bear for 12 days. That means one hex of forest with deer in it can sustain 445 people or 213 bears, or solutions to the equation 14.6 *people + 63.7 * bears = 6500 (deer).

Figuring that humans are about the same metabolically as orcs or hobgoblins, we can basically swap them on a one-for-one basis. 445 orcs would be one large lair, or maybe two small ones. Orc societies in 1e AD&D had 2:1:2 ratio of adult males:adult females:young, so sustainable population losses might be about 40% of that per generation (say, 15 years), or about 12 adult orcs per year. Let's say 75 Calories per pound of body weight, or maybe 10000 Calories per orc, or total of 329 Calories of orc available per day for ogres, giants, and scavengers. That means one ogre would need 15 hexes to sustain himself, and a hill giant 30. One man-sized scavenger could live on six hexes.

Obviously this gets thrown out the window once they start practicing agriculture, but remember the Calorie yield of grain is the same whether it's eaten by orcs or humans. A person needs about 24 bushels of wheat per year, with a per acre yield of maybe 5 bushels, with a year fallow in between. That means, absolute maximum, 20,000 man-sized entities could live off the produce of one hex of land devoted to agriculture, including both peasants and monsters that feed off of them in the surrounding hexes.

Note also that cold-blooded creatures need about 1/10 as much food, so a dragon could presumably live on less land than a hill giant (or, on the same amount of land but feed 1/10 as often).

For less simplified statistics for wilderness areas, here are estimates of animals populations.

Carl said...

Symeon & Menace,

Thanks for the notes! I like the idea of an encounter tree, and the caloric needs of a creature by mass would serve as a starting point for constructing that tree.

We need caloric yields for terrain by type to get to the next stage. Terrain type is broken down in the encounter table of the DMG like so: Plain, Scrub, Forest, Desert, Hills, Mountains, Marsh. We can expand this list to include additional types like Cropland, River, Town, et cetera.

Next, using Alexi's mass numbers for creatures in the Monster Manual we can calculate their approximate annual caloric needs. We can organize them into a tree structure ordering the creatures by terrain type from lowest to highest caloric requirement.


Alexis said...

Menace 3 Society,

I think there’s some excellent reasoning here, and I’d like to troubleshoot it a little.

1) I don’t think that the numbers provided by Yellowstone National Park are the best possible, given that the national park is a groomed landscape, very far from natural and very far from that which would be peripherally affected by various aggressive humanoid tribes (the deer in Yellowstone are not hunted).

2) Basing the final numbers upon specific number counts of animals is problematic, as the number of birds, amphibians, insects and so on cannot be extrapolated from the numbers of mammals. I mean, sure, they probably could be, but I have problems with it. I would rather somehow base the number of creatures upon the amount of biomass in the area - this being a number that would then be universal for every hex, even underwater hexes, since work on estimating the biomass in shallow seas and lakes has been done (sorry, haven’t gone looking for figures - I plan to post on this, so I’ll see if I can obtain some figures by that point).

3) I can’t agree that humanoid needs can be compared to other animals, as hunting is done to obtain more that only food - having a more devastating effect upon a biome than any other creature.

4) I’m a little confused by your calculations as regards hexes; a man-sized scavenger needs six hexes? Seems odd. I’m sure the combined mass of rats within a given 100 square yards within a city would amount to the body weight of a man - and since rat metabolic rates demand more food per mass to be eaten daily than a usual man-sized creature, there must be more available food per square mile than you’re suggesting.

Please understand, this is not meant to disparage your approach; I only describe the places where I would have issues with applying it to my world.

Carl said...

I came up with a list of 50-odd terrain types, including bodies of water (river, lake, ocean shore, ocean deep, ocean abyss, et cetera). I then copied all 400-some creatures from the Monster Manual into the sheet.

I like the idea of tying creatures to biomass, but this is where I was headed anyway. What we need is some kind of connector number to help us determine how many creatures can be supported by a biomass of a given terrain type.

Personally, I'm not that concerned with determining population numbers for creatures outside of the Monster Manual. At least not yet. Models can always become more detailed so right now I'd like to concentrate on developing a system that can be put into use immediately.

One last comment -- I'm not interested in specific number counts as I am the likelihood that a particular animal will exist in a given location. I'm looking for the classic encounter table, but made using some actual biological data.

Alexis said...

I'll put a list together of vegetation types I want to apply to my monster tables, Carl, and you can compare them with your list. My list, as I said a few posts ago, is based on Kuchler's system. I'll try to post what I have of that system tonight.

Also, have you seen this?

I have more on it, and I'll tidy it up and post it sometime this week. But ... and this is a sign of me dragging my feet on this particular subject ... I haven't even started on carnivorous, semi-intelligent monsters yet!

Maybe this is the inspiration I need.

Menace 3 Society said...


You're right about scavengers: I calculated that they ate only the carcasses of large dead mammals, which is obviously not a good assumption.

But as for the rest, I more or less stand by it. Deer in Yellowstone, for example, ARE hunted by bears, wolves, and coyotes. They aren't hunted by humans, but part of the point of my calculations was that everywhere organized hunting by humanoids occurs, it must displace predation by natural inhabitants (whatever that might mean in a fantasy world). Up to 64 deer per square mile is a statistic for Pennsylvania, which has essentially no other wild animal herbivores, making it a good source for the simplistic model.

Using raw biomass as a way of figuring out how many things can live in an area is itself tricky, since not all biomass is of the same quality. I think it makes more sense to use types of biome (temperature forest, desert, jungle) and use where possible actual numbers for wildlife type and density.

As for the dangers of non-food hunting, I think it's not worth worrying about in a medieval fantasy world. Prey creatures hunted even for ritualistic purposes usually get eaten, and the targets of ritual hunt that don't get eaten are usually apex predators, so this would fall under the category of competing for resources.

As for rats, you're right--a city hex is going to have a lot of food being imported, which can support many more creatures than a temperature forest wilderness hex (which is what all my figures in the previous post were for).

I suppose the real theme I was exploring is that you can't really expect lots of big flesh-eating monsters living in the deep wilderness: there's too little for them to eat. Philosophically as well as ecologically, they live just beyond the limits of the everyday human world.

Alexis said...

Now Menace, why would you argue that?

My figures have it that there is about 2 trillion tons of animal life upon the earth; if we discount the 99% of that which exists in the volume of ocean, we're left with about 20 billion tons that dwells on land. If this were distributed evenly over the 148,940,000 of land, it works out to 134.28 tons per; if we consider only the 13.31% of which is arable, this is 1,008 tons per, or 10.08 tons per hectare.

Now, mind you, this is animal life only. The plant life in that same hectare is estimated at 10 times the volume of animal life.

I dwell in a country where, up until four centuries ago, there were an estimated 30-70 million of a particular animal weighing up to one and a quarter tons, scattered over an area of about 8.3 million (4-8 animals per square km), of land that - according to the links I posted this morning, has only 4% of the total biomass of the densest vegetated area available. I am speaking, of course, of the buffalo, on the comparatively less food-rich great plains.

You are making hugely fallacious assumptions. Return and check your premises.

Symeon Kokolas said...

Since there are no comparable ecosystems for D+D monsters and they do not necessarily drop straight into an existing niche, it makes more sense to me to use raw biomass as a starting point. The amount of available biomass would be derived from the raw amount modified by a quality factor (Fq) and a regrowth factor (Fr). The quality factor would be based on the biome or environment type and the regrowth factor would represent the percentage of biomass that can regrow without significant damage each year. For grasslands, Fr should be around .95 and Fq perhaps .3. I would expect perhaps Fr = .2 and Fq = .8 for temperate forest. These are baseless assumptions, but it is possible that this idea has already been researched.

Perhaps it is the division of animal life that is messing up calculations? On the cold northern plains when bison were the dominant herbivore, the vast majority of available plant mass was consumed by bison. There were small populations of insects, birds, and rodents as well as the predators they supported but the overwhelming leader in animal biomass was the bison.

In contrast, a temperate forest supports a dizzying array of life feeding on the plant energy. This is everything from fungus to insects, up through large mammals like deer. In the forest, though, the percentage of plant mass available for consumption (with all things in balance) is lower than in cold plains. The trees themselves could be harvested, but it would take a decade or more for the ecosystem to recover. Grasslands replenish themselves every year and do not suffer unduly from extensive harvesting. Besides, the available energy is distributed amongst a much greater diversity and results in smaller individual populations. Predators tend to need versatility to survive, in contrast to dominant-species environments where predators tend to be very specialized. This is an excellent environment for omnivores like bears, where they can compete against several unrelated groups for their food energy and take advantage of any transient population imbalance.

One resource available to semi-intelligent monsters would be the number of people traveling through their hex (and their supplies). Roaming food resources are more concentrated near cities for these creatures. The risk of being hunted becomes higher the closer they get and the more people they take, though. If there were no patrols and no adventurers, most cities would be surrounded by a ring of monstrous predation. Since cities do have the resources to run patrols and organize hunts, it would make sense to see more monsters in the middle of long traveling routes.

While it may be unlikely to encounter a big group of intelligent carnivores, it is still possible. It is also very likely that nearby human settlements will have at least some idea that there is grave danger along a particular route or in a certain area. Seasonal variation would play a part as well. From late spring through early winter, food is plentiful. Outside that time, food is scarce and predators range much farther than normal. Late summer is probably one of the safest times to travel. Food is easy to find, predators are well-sated on their preferred prey, and the weather is nice (ish). Late winter would probably be the worst.

(thanks for reading another rambling post:))