Sunday, February 17, 2013

Infrastructure Perspective

Okay, this one is going to seem strange to some people.

To me, a lot of the trouble that people have where it comes to "designing their world" is that they don't know what to design.  The default is to design things that are very small - buildings, streets, dungeon hallways and rooms, etc.  These could be termed 'microdesign' in that they serve to apply to the immediate needs of the party, in terms of what is five feet to their right and fifteen feet to their left.  Useful, obviously.  But apart from the immediate, these things don't help much where the party decides to march off in a direction, say a large open landscape where they're going to see two thousand acres in a day.  You can design one small part of one acre, but you can't really get a sense for the scale.

What is needed, then, is 'macrodesign.'  Here, I think, the game and most of the players are notably lacking, because they just don't know how to approach the problem.

To try and sketch out a method, I'm going to return to an image I once used to describe my trade system: the map of Hothior, from the old game Divine Right!

Ah, so familiar, so conveniently neutral.  I am going to use this to demonstrate a system which really works better on the maps on my world, but for the general audience I think it would be best to simplify that so that anyone can apply this.

That does mean I will need to add some information to the above to make my system work.  To start with, I need to give a population figure to each of the above cities, Port Lork, Tadafat, Lapspell and Farnot.  I also need to establish a population for the region.

If the hexes are 20 miles across, the above region is about the same size as modern day Bosnia (without Hercegovina).  Recognizing that there really ought to be more than four cities in an area this size, I'm going to identify the population at 2,000 per hex (there are 50 hexes, counting all those on the water as a full hex).  Total, 100,000.

For city populations, we'll say Port Lork has 8,000 people; Lapspell, 6,000; Tadafat, 4,000 and Farnot, 2,000.

I'm introducing these numbers so that you, in your world, can take the probable size of cities that already exist.  Obviously, the population isn't going to be distributed evenly throughout the countryside.  More importantly, neither is the infrastructure ... and what I mean to do is produce some infrastructure numbers we can work with.

Let's take the total value of the cities (20,000) and divide it into the total population.  From that, we can presume that each city influences 5x its base population:  Port Lork, 40,000; Lapspell, 30,000; etc.

Let's further divide these bigger numbers by 259 (persons per square mile).  We'll call this the BASE Infrastructure Number.  For Port Lork it is 155; for Lapspell, 116; for Tadafat, 77 and for Farnot, 39.  Let me add those numbers to the map just for convenience.  In addition, I'll color code what amount is coming from which centre:

Now, let's say that for every hex away from the core hexes, the cities themselves, the Infrastructure Number is halved.  Let's also say that if the adjacent hex is a forest, the number is cut to one third.  If hills, the number is cut to one quarter.  And finally, if mountains, the number is cut to one fifth.  The number is always rounded off to the nearest whole number before calculating the next hex.

Let's just work out the hexes immediately next to the core cities:

The important thing here is that where an overlap occurs, the numbers should be added together.  The hex between Farnot and Port Lork, therefore, adds 20 and 78 together.  The gentle reader may also notice I didn't extend the Port Lork number into the hex adjacent to the water ... that is because this infrastructure system is measured overland, except where overland is impossible.  Also, because this is infrastructure and not actual population density, the whole seashore is counted as one hex, even if not all that hex is land.

The influence of all the cities are ultimately expanded outward until they cannot be anymore.  As well, each city's influence affects the actual hexes of other cities.  So, even though its messy, let's expand all city's influences out to their maximum:

That's annoyingly difficult to read (and I hope done accurately) ... but it shows the pattern distributed from each of the four cities.  Usually I would do one city at a time and add up the numbers as I went along, but this is for demonstration.

Let's add up the numbers to get a single total for each hex:

Now, let me stress.  This is simplified from what I usually do.  I usually have more cities; I have elevations for every hex which influence the total distribution and son on. 
Also, I don't suggest that what we have here are any surprises.  Obviously the mountains and Bad Axe forest were going to be empty.  Obviously there was going to be higher numbers around Port Lork.

What's interesting here is we have exact numbers.  We could say that these were % die rolls for whether or not there was an Inn in the hex.  We could use the inverse number for the likelihood of bandits.  Or a measure for how good the roads are.  A minimum number could indicate a ferry (50) or a bridge (100).  A forest hex with 60+ could have industry; 30+, gameskeepers; and less than that, thieves or brigands. 

It's really up to you how you want to assign values, and what for.  If you want your cities to be more metropolitan, increase the penalty for infrastructure in adjacent hexes (divide by three or four or whatever you wish for each hex out).  Add a penalty for crossing a river.  Roll a die and make "evil" hexes that severely lower the numbers.  What's important here is that you find a way to measure a lot of blank hexes in some way that establishes a frame upon which you can build up all the little pieces I spoke about in my last post.  How many points does a hex have to have before there's a hospital?  A school?  How little does there need to be for the party to be able to have a free hand in the area?  Where does management and government hold the greatest power?  You have a yardstick to give all that space a greater integrity.  Call it a sort of "infrastructure perspective."

Go have fun.


  1. This is exactly the kind of thing I need right now. The party's moving from a dangerous desert wilderness area to a more populated one with a lot of cities, and I've been having trouble with creating a realistic sense of scale.

    Something I think I'd use with this is a reverse of this system for distance to monster lairs and dungeons, where infrastructure decreases with physical proximity. There could be "dead zones" surrounded by inhabited areas that are simply too full of monsters to make large scale infrastructure possible. Clearing out these monster strongholds and allowing human civilization to move in could be a good job for adventurers, especially if they own the land themselves.

    Do you plan on posting a list of infrastructure types by likelihood? I think most people would customize the system to fit their world, but it would be nice to see an example.

  2. Considering that I am in the process of building a new campaing, this is quite useful. Though I am using 6-mile hexes so I had to rejigger the reduction of infrastructure point, to 20%, 33%, 37% and 41%. Now, I just have to build system for macro-economy and I am ready for some hexmappin'

  3. You mentioned your own heightmapped version. Where/how does the elevation factor in?

    Also, I second Ozzie's hope for a potential list of things to add to such an infrastructural chart.

    My mind would be freaking BLOWN if I redid any one of your maps with an infrastructural list, especially since I was planning a campaign in Finland, which has a lot of open area to really illustrate how spread out simple resources could be.

    And of course, I imagine this is all just one more step towards your efforts in encounter design, neh?

  4. Every 400 foot change in elevation counts as 1 point of division. This is in keeping with 400 feet counting as an additional day for the trade tables.

    Thus, if one hex is at 400 feet, and the adjacent hex is 2,000 feet, then the amount of change is 5x. This means that the high mountain country tends to be more "remote" than flat areas far distant from cities.

    You have seen an infrastructure list - it is the "vegetation" map on the wiki of Croatia, or any vegetation map. The mixed is 100 points or more, the cropland is 500 points or more.

    I just hadn't thought of counting the numbers in terms of their precise effect on possible features before.

  5. Oops, that should read 4x, not 5x ... I was distracted.

  6. So, my mind was blown as predicted, and now I have to go back over all the maps, and figure out something awesome.

    Well. Done. Sir.

  7. Now this I like. And it also reminds me I need to sit down during my vacation this week and fully realize my map for the area the players have been in. Ugh.

    This is the sort of system stuff that is very useful because it creates numbers. People really do underestimate just how important "numbers" can be. There is nothing put forth here that can't easily be done via a program and it gives a solid foundation for determining so many things. I like it. I am going to have to take a crack at wedding this to something in this coming week.

  8. Alexis, while I applaud the depth of your scholarship, and the intellectual effort to arrive at the level of detail presented, I can't help but wonder how this differs from world-development focused on tiny details.

    Isn't it simply a matter of scope? I often shorthand the level of detail you're presenting by referencing existing or historical places. Places I've visited. Sometimes just linking a picture in my notes to remind myself.

    From that shorthand framework I can conjure enough details to satisfy most players.

    I understand the premise of your philosophy of logical campaign structure, but I feel that, at the core, it contains an unnecessary level of detail. You know what they say about opinions though ...

  9. Nothing wrong with that, Keith. But I beg you to consider that IF everything you draw comes exclusively from inside your own head, then you are going to repeat yourself.

    The difference between the hard numbers above and my imagination is merely that I am stretching my imagination to the framework the numbers offer, as opposed to forever crunching every outside templace into my own perspective.

    To put this another way ... the actual numbers produced by the calculation ARE a type of experience. I look at the map, I see the numbers, and I learn from what they tell me information that I did not possess before.

    Thus, before I had one conception of the way I would describe that part of my world, and NOW I have a different conception. My conception has expanded by NOT limiting myself only to the knowledge I already have.

    You have evidence that the knowledge I already have is considerable - how do you suppose that happened? Certainly not by always trusting to my previous experience to run the various moments of my world.

  10. It is the same thing that makes me draw from random generation so forces me to make things fit and make sense of items I would not have ever thought to put in the world because, being human, my imagination is both biased and limited.

    As a test several weeks back, during one of my walks on my lunch break, I took time to observe what I saw on the ground. I found a dime, a dead snake (that was weird!), a Pokemon card, a piece of wiring from some kind of electronics and a box of tic-tacs. I have no idea how any of those objects got to where they were...sure I can GUESS how they might have gotten there but one never knows. And that was just on a half mile or so of road.

    If you ever asked me to give a list of objects found on a road I can't say anything I just listed would be pulled immediately from my mind...but the world is a big, random place...and a D&D world should be a big, random place too! It makes it far more fun both for the players and the DM.

  11. Alexis and Yagami, I get what you are saying. And, framed as it was in your response Alexis, I have more appreciation for the experience that your model provides.

    But your argument assumes my experience is static. Yet, I'm still traveling, experiencing different cultures, learning. My experience is infused regularly with new stuff. People I meet, or read about become my NPCs ...

    Maybe we're doing the same thing you and I. Your world arises from mathematically-modeled data points, and mine from experience. It's not as efficient as a model, but even math repeats itself once and a while...

  12. Oh, I've made this point before, Keith. My world only appears to arise from mathematically modeled data points because those are the easiest to express as structures.

    My world arises far more from the violent passion with which I write this blog.

  13. I came back today to read this again. I am going to look at applying this rigor to a Traveller map for my Star Clans campaign... My long term goal is to be able to randomly generate entire sectors, and using a system like this to calculate economic and political influences brings that closer to reality for me.

  14. In later rings, those not immediately surrounding the origin city, how do you handle the overlap of possible origin hexes? It currently appears that you are using the average of the two possible origin hexes, but I'm not certain.

    Obviously the "overlap" rule doesn't apply, since that would make two equal hexes always produce a third equal hex in the next ring.

    I'm trying to apply these numbers to Lenglin and Breen at the moment, and the mountainous terrain is leaving me flummoxed in how to apply these tricky cases.

  15. I'm not quite sure what you're asking, Arduin.

    Treat each origin hex (city) separately, work out its distribution until the next number is less than 0.5.

    Then ADD the overlap hexes together.

    See where in the example, Port Lork and Farnot both affect the hex between them? That hex is a total of 78 & 34 (not counting what Lapspell and Tadafat add).

    Now DON'T add those numbers together and then use those as a base for the next hexes. That's not how it works. I think that might be what you're doing.

  16. I'm explaining poorly. I don't mean the origin of two seperate "colors" of data, I mean when one hex has two possible points of origin from the same city.

    Port Lork, blue, gives itself 155, then naturally the outlying flat hexes get 78. Okay, good. However, from which of those 78s do I get the following 39?

    I'm thinking of the specific two hexes, the rightmost of which contains the actual words "Port Lork".

    From what I've been able to see, you are not using an average of two values, but rather using whichever value produces the highest number in the "child" hex, as in the case of 52 and 78 both going into the Bad Axe Forest (56/3<78/3).

    My confusion arose in that, if both "blue" 78s had been applied, both giving 39 to the child hex between them, I'd have yet another 78 "blue".

    I was confused as to why similar colored numbers with multiple possible parents were not suffering from recursive calculations.

    However, if you were using merely the highest possible value of the same color, it would solve that problem.

    Apologies for inadequate explanation of the issue.

  17. Okay.

    The next hex, "39", still comes from Port Lork. The 155 is divided by 2 for each ordinary hex FROM Port Lork. If both hexes are plains, then it is 155/2/2 = 39.

    If the two hexes are plain, then forest, then it is 155/2/3 = 26. If the two hexes are both forest, it would be 155/3/3 = 17.

    If we were talking about something that was three hexes away, all plains, it would be 155/2/2/2 = 20. And so on.

    So the hexes between the target hex and Port Lork do not have any effect on the target hex. You calculate it all from the city hex ... and whatever the highest possible result, that is what you use.

  18. Where does the 259 come from? I understand that it represents the population per square mile, but with regard to what? Urban density would be at least, I think, 38k per square mile (that's from data for the early Renaissance period, so it should definitely be a minimum), while 259 people per square mile seems, to me, to be rather high for a rural population.

  19. The 259 comes from a miscalculation of the number of square miles in a 20-mile diameter hex, one I made 10 years ago and which I don't want to adjust now because it would mean redoing hundreds of hours of work.

    259 persons per hex (or 1 person per square mile according to my miscalculation) equals an infrastructure of 1 point.

    So it doesn't mean 259 people per square mile but 1 person per square mile x 259. That should sound like 'rural.'

  20. Great. Thank you! That is very helpful.

  21. Other than using elevations to limit infrastructure, are there any other ways in which your process differs from what is described in this post? Are rivers barriers in the same way as when calculating distances for the trade tables, for instance?

  22. The only barrier I use for infrastructure is elevation, and that all infrastructure for a given region is contained within that region's borders (as the population density that the infrastructure is based on is region specific).


If you wish to leave a comment on this blog, contact with a direct message. Comments, agreed upon by reader and author, are published every Saturday.

Note: Only a member of this blog may post a comment.