View Full Version : Plausible Distances & Populations

11-10-2008, 12:03 PM
I know I'm nit picking but I tend to worry about the distances between towns and if the land around them is plausible to support their populations. I like to draw as much as I can from the geography of a map so when I make one I try to put some thought into it.

Along those lines I found a great website about Herefordshire


Its very approachable and has some good info. This seems like a coherent little snapshot of medieval town development.

Chiefly, I'm looking at the little map attached to this post.

The first map is from the Herefordshire website and the second is from wikipedia. So you can see where in England it came from. Herefordshire was a center for the Anglo Saxons and Alfred the Great. It bordered the Welsh and has its own illustrious history that I will largely skip over. I'm going to try and concentrate on physical map elements.


There are many on this site who know far more about this that I do. Someone living in Herefordshire is probably reading this right now :). I appreciate gentle correction and considered opinions. I've never spent any time in Herefordshire. I'm putting this together from websites and google. My apologies for mystakes or false assumptions. If you want to pontificate on this region and what it has to teach us about map making, you too are welcome.

11-10-2008, 12:34 PM
If you look at the greater map of England in the previous post, you can see that Herefordshire was by the middle ages a border region. It was not a center like London, York or Paris. Although fully populated it did not have the same crush of people that London did. For the sake of this treatment I am going to assume it is equivalent to a role playing average or smallish holding (King,Knight,???). Depending on the world and the game needs I'm sure the leader could have any title. I'm using it to create a middle of the road rule of thumb.

First, the over all size.

At its longest its about 44 miles. It is roughly 40 miles in diameter. Looking at the list of possible towns I count 24 identified sites. I'm not too hung up on details (This is a game after all) so I'm going to take this as a maximum of concurrent towns in a region.

There were only 8 Medieval towns present after 1500 in the region. Of the 24 possible I'm going to reserve 8 as dominant and either town sized or better.

(see first pic - this post)

There are a known collection of 13 failed towns. I'm going to reserve that many struggling or really tiny towns. This will include little Abbeys or wierd game elements that aren't supported by trade and commerce.

(See Second pic - this post)

Now looking at the surviving towns in 1500 they have a noticable space between them, whereas the failed towns are sort of clusterred. This might be for all sorts of reasons but I'm going to interpret that as a radius necessary for a successful town.

This really is very arbitrary. Population pressure can make bigger cities out of close small towns - as happened in London. The failed towns could have been grouped around resources or leaders that were unsuccessful etc... Stil this will be my rule of thumb.


11-10-2008, 12:59 PM
Herefordshire is the seat of the region. I like capitals that share the same name as a region - its one less name I have to make up :).

The Distance between it and its 4 closest neighbors in the surviving town map is approximately 12 miles (See circle in Picture). This, I have been told, is a common distance between European cities because (I am told) it is comfortable for horse travel.

So be it. 12 miles between larger villages\cities etc... Less space and they are probably competing for resources and influence. More space hampers communication and trade. Without a rich resource or an exceptional element two cities inside of 12 miles are both going to be smaller or on a collision course.

Herefordshire is also on a river. But only 2 of the 8 surviving towns are on rivers. Notably none of the towns are on the coast. (This was a frontline between Norse and English for a while.) I can imagine that the bonuses of more trade might be necessary to achieve higher town sizes (and bring in more food) but they increase the ease of travel for invaders as well.

Wikipedia gives me this....

Herefordshire has always been esteemed an exceptionally rich agricultural area, the manufactures being unimportant, with the sole exception of the woollen and the cloth trade which flourished soon after the Conquest. Iron was worked in Wormelow hundred in Roman times, and the Domesday Survey mentions iron workers in Marcle. At the time of Henry VIII the towns had become much impoverished, and Elizabeth in order to encourage local industries, insisted on her subjects wearing English-made caps from the factory of Hereford. Hops were grown in the county soon after their introduction into England in 1524. In 1580 and again in 1637 the county was severely visited by the plague, but in the 17th century it had a flourishing timber trade, and was also noted for its orchards and cider.


So I am going to assume that these town placements are largely agrarian - farms and such. Especially in the middle ages large populations need their food producing regions to support their higher population. Only for the system to work the food producing regions have to export their people along with their food or they'd have nothing to sell :).


So this is the running rule of thumb -

Small to Medium Territory
40ml diameter
1 major river
Good farms and lands but no exceptional industry.

2 Larger cities - probably one dominant keep\castle.
Influence\support radius around larger cities 12ml.

6 Villages
As many as 16 further sites of interest, ruins, failed cities, manors etc..... (this is only a rule of thumb.)

I can't say this is historically researched but I'm content that its vaguely plausible based on this very small sample.

Steel General
11-10-2008, 02:52 PM
Very interesting.

11-10-2008, 04:16 PM
Now the application. This is not a science. I'm just hoping my rule of thumb will give me an option to keep or depart from.

This is output from Fractal terrains without rivers or trees. I have modified it heavily in Photoshop. I am trying to be accurate because I carve off pieces of the same world for the same group. That way FT lets me consider accurate distances, even elevations.

Personally I really prefer to 'discover' land rather than invent it. Random gens are brilliant because I recognize the right sort of thing but second guess every step if I have to plan it. Besides it just feels more natural to me if I have the story be influenced by the geography not the other way around.

These are stats for the blank map.

1km = 20.9px w
1km = 18.362 h
77.98km across
57.60km Top to bottom

Settling on a map scale of 19px = 1km

Roughly 82km Across & 56km Top to bottom

Highest 107.2708M
Lowest 6.9501

I can take little credit for the accuracy of these figures - they come from Fractal Terrains or Wilbur

I hope the elevation, scale and coordinates of the piece are clear. Unfortunately with little ground detail its very hard to intuit how big this map is or what it might hold.


11-10-2008, 05:50 PM
I like the way you think or, in this case, deduce. I do my maps based on my surroundings of the the St. Louis suburbs where every named township is about 5 miles apart and larger cities (approx 40,000 people) are 15-20 miles apart. This corroborates with the first pic you posted. So I agree with you 100% as to the way your thinking.

In college I drove back n forth frequently between home and Mizzou (exactly 100 miles from my driveway to my fraternity or about 90 minutes). Everything seemed to be spaced about 15 minutes apart. Starting in St. Charles, it's 15 to O'Fallon, then 15 to Wentzville, then 15 to Warrenton, then 15 to High Hill, then 15 to Mexico, and 15 to Columbia...roughly. Thus, I do my map distances almost exactly the way you do...villages are 5-10 miles apart, towns are 10-20 miles apart, small cities are about 45 miles apart, large cities about 90, etc.

I also liked studying this phenomenon since Lewis & Clark started here, St. Charles was the first capitol of the state, and Daniel Boone lived in Warrenton after his famous days (he was buried there too until like 15 years ago when he was taken back to his home state), lest I forget to mention Jesse James and his gang. Since there were no cars in them thar days I had always just taken the 5/15 thing as some sort of human nature rule of thumb.

From downtown St. Louis to Columbia is 2 hours and then another 2 hours to Kansas City (by car at 70mph). The actual mileage is about 250 so the multiples of 15 would be 240 (same as for multiples of 12). So I'm thinking that the 5 mile rule for villages might be true and the 12-15 range for larger towns might also be true. Between St. Charles and KC it's all pretty much farmland, except for Columbia, so these distances might only be true for rural settings and cities might be different and more clustered.

Looking to seeing more of this.

11-10-2008, 08:44 PM
One question for me is always what level of detail do you include with a map at a given scale. Since you start with a largely blank map I always wonder what is too much.

Thankfully, with a real world example we can compare samples of land and detail. These are modern samples from Google, map & satellite images of modern Herfordshire. I've roughly scaled and placed them inside the border lines from the earlier maps.

As you can see there are quite a number of hills and a fair amount of detail on the satellite shot, event the 'Modern Map' has quite a bit of detail.

The Modern map is very similar to the historic one (the map itself is a little better) similar number of towns & most of the old towns survive.

If I overlay the Hereford map on the FT blank, at the appropriate scale, you can see that my blank needs a great deal more detail and a great deal of room. I couldn't imagine that all my and gaming would be drawn from this one map but I'm sure there are real world people that had great exploits and never left Herefordshire.

My little square of land doesn't look so small anymore. Based on my real world example it's probably big enough for a small kingdom and some wilderness, even 2 different kingdoms\regions.

So the land currently in my map is very very flat and bigger than it 'feels'. This might be fine for a political map but it doesn't really have enough detail for my liking. If you look at the modern map of Hereford - it has far less detail than the satellite but more detail than my map right now.


11-10-2008, 09:31 PM
A word on Fractal Terrains

FT is a wonderful program. It generates a random globe and lets you cut it up for your gaming.

Unfortunately, in order to scale to the whole globe in reasonable time with reasonable computer resources it doesn't do every available bit of land.

This selection is a tiny, tiny part of the world I generated. The red square in the middle of the globe (see attached) entirely covers my blank map. I'm adding detail to an infinitesimal bit of game world - but at least I know where it goes. Some day I wish there would be a way to modify regions and paste them back into FT, or another global program. As it is, I keep directories based on continent and locations on a modified google earth.
The ideal solution for me would let each detail build on the previous ones without having to tear anything down to do it. These simulations have so much detail that you can never add it all at once.

Anyone familiar with the Harn game system and world? I think they have a wonderful model. Everyone shares a world that has a history but stops background development on a given date. You can add all the depth you want but don't step on other efforts and don't move the world into the plot. If you have earth shaping events plan them after the given date. I think that frees up world design to be descriptive but not reactive or mutually disruptive. It maximizes people's efforts and encourages people to spread out and develop new areas.

On the first globe, the Red Dot towards the center of the globe covers my whole 'blank map'.

I think its pretty amazing you can go from orbit down to a selection 60 km high. When you get there though you have to add, or generate, some land detail if you want to approximate the new scale.

Lastly, for those involved in the Cartographers Guild World project, one of the mapping squares with an appropriately sized Herefordshire. (Look in the far South East by the scale and compass. No thats not a lake, its a whole shire! As you can see there is a lot of room in one of those squares :).

Accuracy - Trying my best.

Other than the stats from my sample map (for which I have FT to thank) the rest of the scales have been set mostly by eyeball. I'd be surprised however if they're more than 10% off. The biggest source of inaccuracy is probably the shape of screen pixels which are wider than tall.

If anyone spots an error or has something to add they're welcome to speak up.


joćo paulo
11-11-2008, 05:32 AM
:compass:Sigurd ,you are the man, this post should go to the E-zine.

11-11-2008, 02:30 PM
Thank you all for the comments.

Steel General - I appreciate those who take the time to read it. I'm trying to make it visual and approachable. I'm thankful to the site that I can post the pics and text together so easily! Scale is hard to work with unless you've 'seen' it. The only measurements Im super familiar with are small ones. I have some trouble pinning down a 'Mile' or a 'Kilometer' when I write\map.

Ascension - appreciate the compliment. :) I'm hoping if I can settle these questions for myself, others might be interested in my rule(s) of thumb. They're going to be hugely imperfect but I hope they're plausible.

joćo paulo - Gotta like that kind of compliment :). I'm just happy its been provocative or interesting for someone else.

Thanks everyone.

I'm not done yet. Just have to plan the rest of the journey.


11-11-2008, 04:37 PM
Kudos for doing this research! This kind of depth is, to me, always worthwhile.

11-11-2008, 08:36 PM
Check out the website from the first post for a list of over 100 a hundred castles in Herefordshire. Some have maps and pics. Most of the pics are of surviving sites. Its actually a really great site.

So Back to detail.

We've talked of Towns. For our purposes Towns are bigger than villages and have some sort of recognition. "Town Air is free Air"

The first is a map of over 100 fortifications.

The second map shows locations mentioned in Domesday Book. This is before the plague of the 14th century. 312 separate places are mentioned for Herefordshire.


The population is listed as 4453. The Domesday book was a taxation book and its likely that number is low. Who doesn't want to hide from the tax man? But as it is, that number gives us a round figure for a civilized land. 40km diameter - and 5000 people - approx density. Don't forget this is after a war and without magic.

I found this quote very interesting.

The Old English word "tun/e" is often attached to a placename and denotes "settlement", "village" or "enclosure", eg. Winetune. "-tun" has also been interpreted as "outlying farm", which might account for the fact that some Herefordshire Domesday places are today only individual dwellings or farms.

11-11-2008, 09:59 PM
So what is in this 40 mile diameter piece of land?

If we overlay all the 1500 or earlier locations from all the maps together, and place them on top of a modern physical map, we get this....

Pretty clear that there is a heavy stamp of civilization on this shire.

If we look at the development we see some standard things.

1. Most of the fortifications are on the borders of the shire with the capitol in the center. Mottes were built mostly on hills and more frequently than castles (They're cheaper and easier to build). Norman invaders built Mottes and then eventually Castles - the Saxons didn't build castles. If a site was important enough for a castle it mostly stayed important - most castle sites are on or near modern roads.
This lets us reason a defensive plan. The largest castles are in the center of the region and at the end of river systems connecting to the main castle. Mottes are built readily where there is high ground especially if the land borders the regions opponents (In this case the Welsh).

2. Domesday sites are mostly placed in valleys without much regard for castles or fortifications. The land qualities were the deciding feature for a successful farm or manor.
The exception seems to be river banks. The southern border has a river snaking through a mountainous region. Domesday sites are common there as well.
I think this sort of reasoning is helpful if you remember that the Domesday book was a tax document. These sites might not have been grandios or militarily important but they produced revenue and contributed to the economy of the region. I imagine a more suppressed population would need more, more local fortifications.

11-12-2008, 10:55 AM
If I continue my blank map and paint on friendly trees to my own personal sense of scale I come up with....


This is a pretty map, if I say so myself. But consider the level of detail from Herefordshire. If I overlay the collected detail at the same scale I see that my pleasant map is way too serene and empty. The circle is 39 mile diameter or approx 62 km.


Unless my forests are gigantic and my earth incredibly flat (could happen - this is not real) I need to rescale the map or increase the detail to be plausible. I have to rethink how I represent detail to fit the scale. My tendency is to draw forests as an area where trees are distinct from grasses. Clearly if I'm drawing at this scale I have to talk about mixed plant densities. Most stands of trees will be indistinguishable from the surrounding grass.

If I want something like the same level of detail for my maps its pretty clear my serene map should be only a couple of km across, not 82!

Detail is Important.

High ground is important for defensive structures.
Valleys and river banks are important for manors and farms.
Depending on your scale, you should have both.

Mountains and valleys are different than plains. They concentrate water in regions making it easier to farm. My broad friendly map would probably be very dry if was very flat. Its not enough to have flat land for farming. Besides being visually appealing, hills and mountains suggest usage patterns for a region. Compare the dry prairies of Canada to the temperate rain forest of British Columbia.


11-12-2008, 03:26 PM
I think there's no getting around the painting of the ground to make detail.

To make the job easier I made up some more cut up trees to give me something to work with. I'll think twice before I start cutting up or working on space below the trees. :).

The problem of course is that the better this region looks the more difficult it will be to combine it with the other regions at other scales.

I am going to select out the plateau at 60m for the primary fortification I will start with. The plateau is 2 or 3 km across so it is much bigger than it appears on the regional map. If I zoom in I can begin to think in different closer scales. I want to work as big as possible but achieve fine detail. The broader I can work the file the easier it is to reuse or adapt. The more specialized I make any individual piece the better it will look, by itself.

I have been doing all of this at 72 dpi. This is where I begin to pay for that.


11-14-2008, 03:34 PM
I still am not correct for my scale.

If I capture the rough outline of Herefordshire and scale it appropriately I get this. Both of these images are at the same scale. Look at how bumpy Herefordshire is. This is a physical map - something like what I'm aiming for.

Notice that the bumps in the google map are small and plentiful. My map is smooth with a contour to show me general outlines. The contour is a start but you can see that, overall even the contour is too flat.


11-14-2008, 11:05 PM
The default FT settings do not function well for maps much less than 100 miles across or so. The target purpose for the product was world and regional maps.
You can sometimes get better results by increasing the number of octaves in your generation. Use Map>>World Settings to bring up the World Settings property sheet. On the Fractal Functions page, click the Parms button. Change the Octaves settings on the dialog and click OK, then click Apply on the World Settings property page.

The below picture shows the before and after of changing the octaves from 13 to 20. There is a slight increase in redraw time when doing this as each pixel has the fractal algorithm iterated 20 times instead of 13 times.

11-16-2008, 12:28 AM
I will try that, thanks.

I've been attacking the situation through Wilbur and Photoshop. Some excellent tutorials and grim stubbornness. I'm even READING MANUALS!

Wilbur has an 'other' map that draws rivers. I found at this scale, I was better off exporting the rivers as lines and then working with the rivers, the height map and some noise to build a more detailed bump map.

Bottom line, for anyone reading this, you can't really paint a bump map with any cleanness. Photoshop's lighting affects - or similar products give you a believable texture.

FT does a great job at the scales it was designed for. As this thread has stressed though, different scales give you new challenges. Part of my problem is that I'm not really using FT and Wilbur fully.

Here's the current more bumpy beast.... Hopefully you can still see the original elevation scheme in the ground. (I removed the contour lines because there was just too much data.

And another bump map.

Any Constructive crits welcome. I Can already see that I'm going to want to vary the tree foliage colour more and...


Steel General
11-16-2008, 09:19 AM


Good God Man are you INSANE! :P

This whole process has caught my attention, usually I just kind of skim over these as I'm not overly-interested in the "science" behind it. But this time its different.

Looking forward to seeing more.

11-17-2008, 12:23 AM
Took way too much time. I had to depart from my tree method and lots changed.

This however looks like the scale. Its mostly accurate to the original, for comparison.

11-17-2008, 10:53 AM
Hmm. At this level of bumpiness, the very-smooth looking coasts are kind of out-of-place looking. Can you fractalize these coasts a little more using similar methods?

11-17-2008, 01:39 PM
Karro - I know what you mean. Problem is the file is eventually an overlay for a bigger world image. I keep the coastline to help me orient the files. I think I'm going to leave the coastline the way it is. I have a game to prepare for.

I've included small versions of the player files. The masked file is what the players have seen as they traveled along the river. They have visited areas 1 to 4 and are now exploring 4.

Now I'm going to use my scale to draw a new encounter area of a much smaller region around area 4. I know that 1 mile = 30 pixels on the big map. (Like to say I planned that but it was a welcome fluke.)

I think the next set of maps should be 2 miles high by 3 miles wide. So I'm going to make a box on a new layer. I use the Marquee tool to draw a square (approx size) and then stroke it with 1 pixel. Then I'm going to transform that box into a shape precisely 60 pixels high and 90 pixels wide.

Then just for ease later I make a new layer inside the square and use the same process to draw a white box 3 pixels high and 30 pixels long. Then I copy that box, shrink it to 15 pixels long and colour it black. I place the new scale inside my view box and then link it to the frame.

I now have a little window I can float around my big map and look for likely encounter areas. I touch up the area around 4 and decide on an area.

See screen cap. (Notice how my dpi setting is beginning to make things blocky)

Now I can find any area of my big map. Copy the file to a new name and then crop out everything that isn't in my frame. I could copy but If I crop I keep the layers separate and keep my options open.


11-17-2008, 10:20 PM
This is the new starting point for the local encounters. The images are not pretty but they are a lot better beginnings than nothing at all.

I take the very small image and multiply it by something I'll remember. In this case 500%. If I want to try and reintroduce items to the big map I can shrink them to 20% for an approximate. (In practice this is not always feasible but I like to keep my options open. The new square is taken from the tiny 3x2 mile box on the larger image. I'll leave the encounter area locked and invisible on the larger picture file, just in case..

11-18-2008, 10:02 AM
Karro - I know what you mean. Problem is the file is eventually an overlay for a bigger world image. I keep the coastline to help me orient the files. I think I'm going to leave the coastline the way it is. I have a game to prepare for.


Maybe there's insufficient time, given your game-time needs, but one thought I had was: make sure you mark where the coasts go off the map (and join the larger image of the worldmap), then fractalize the coast using whatever method works best (something that uses the existing coast as it's primary basis, of course), then go back and manually fix up the edges of the coast to make sure it will stay flush with the larger images.

11-19-2008, 03:56 PM
I took the tiny selection from my larger map and multiplied it by 20.

(If you crop a selection watch your placed files - they don't get cropped. My tiny map was 2.6 gig in size after only being multiplied by 4!. I had to rasterize and crop the extraneous files and my file x20 is 71mg)

So after looking at a couple of other files - a peaceful village without fortifications is about 250' in diameter. This does not include fields etc...

Here is my new map with the space for the village as a square in pink. That's not one house it's 250' square. 1 mile = 5280 feet

You can see that what I essentially have is a second local 'regional' map. Things would be too small for an illustration here. The physical layer indicates a ridge to the west of the town and then a depression. I think that depression will be a river for the town and the ridge will have some sort of fortification. Its pretty obvious that this detail would be way too small for my earlier maps.

Just looking at the space. I might be able to fit as many as 3 villages in this space. I probably have to reintroduce some trees and hedge rows to the smaller map.


11-19-2008, 04:31 PM
So I take a new cropping centered on my town (250 ft) and reconstruct the scale.

I multiplied the last zoom factor by 10 to work on it (making my scale 1 x 20 x10 or 200 times the original section from the first map. The selection I'm including here is 1/4 size (for space reasons) so you can consider it 50 times the size of the original selection.

As you can see its still not ideal for detail of the village but its appropriate to illustrate the general relationship of things. The relationship of town and fortification, river, and fields are clear. Now I need to clean this up and have 2 detail maps, Town & Fortification/Docks

With a bit of refinement, we have our first player map. A relational map for the Valley of Loln. In the attachment you can see the numbered sites in the valley. 1. the keep 2. the south bridge 3. the north bridge 4. the Town itself. 5 a secondary fort and 6 a keep for a competing noble.

Using our standard ration of 2 high to 3 wide, the maps are now 500' high and 750' wide. I have a couple of possible maps to do.


11-27-2008, 03:40 PM
I'm wondering if my little maps aren't too small...

Anyway. Here's a really cool site for anyone who's read this far :).


11-27-2008, 04:32 PM
Interesting project there. I wonder if they had people move in and live there permanently like this one show I used to watch about the early American colonists (don't remember the name of the show at the moment).

02-05-2009, 08:59 PM
So I am having problems with the city at this scale. My plan was originally to have an estate in the north east and the city in the SW. Too little room for the estate and the city is looking very village like. The biggest building here is no more than about 30' at its largest measurement. Still I should be able to make this scale work.

This is actually a very intimate scale for buildings and neighborhoods. On the bright side I'm learning to use brushes for colouring the roofs. On the down side the ground layer is far too basic and noisy without being very responsive to the way people are using the land.

Its a WIP.


02-05-2009, 10:58 PM
Wow, I'm impressed!

Steel General
02-06-2009, 07:35 AM
Good to see you getting back to this Sigurd... unfortunately I don't have any suggestions for you.

02-08-2009, 12:08 PM
Methods don't develop without suitable chaos first.

I'm still in the process of doing 30 mostly useless things to discover the 5 that work. The little town is beginning to feel more warm.

Painter directional brushwork shows promise as does a renewed interest in heightfields.

Just went back through my layers and stripped out what wasn't really working.