Detail XIV: Benefits

I have put the world project on hold for a bit to get refreshed with other work.

Over at the Tao, Alexis has been reworking his hammers/coins/food system. It's a brilliant system with endless opportunity.

My own use of programmatic aids lets me break a 20-mile hex down into not just 6 6-mile hexes, but 400 1-mile hexes. This is a lot of detail; not something to be done by hand. And I do think the 20$\to$6$\to$2 system generates very beautiful results. I want to consider working on this for a while to see where it leads. By saving the random seeds that the model is built from, I can replicate a nice result without storing too much data.

This was the result from last year:

Last year's detailed hex

My latest version has some upgrades with color, but also additions of hex-type and benefits (hammers/coins/food). One major change is that the scale will be much different than the Higher Path: with many more discrete points, the number of benefits will be correspondingly higher.

An example (not of the same hex as above, I couldn't find the old version of this one) of a 20hex with an infrastructure of only 20. Yet, due to its presence in a forest, with some substantial clustering of hexes (producing a large number of high hex types), it generates a total of 134 hammers, 168 food, and 78 coins, calculated for each 1hex according to a binary system such that $B = 2 ^ {b - 1} + 1$, then added for the whole 20hex. This doesn't yet consider any benefits that would be added for the presence of a trade reference.

Updated hex

There are two paths forward. The first is to adjust the benefits conferred by each hex type such that the numbers become more reasonable. The 20hex total may not be as important as the individual 1hex number: eg, the settlement hex generates {'hammers': 5, 'food': 5, 'coins': 4} in the binary system, or {'hammers': 15, 'food': 5, 'coins': 4} in the decimal system. Without context, it is impossible to determine what constitutes a "reasonable" number. The second is to adjust the results of those benefits.

If a system is to work properly, it has to generate desirable and consistent results. The advantage of doing this programmatically is that I can implement totally different systems at the drop of a hat. The design loop is much tighter than hand calculation (although I am trying to avoid something that is so opaque it can't be replicated by hand).