Getting closer to having a simple game that will help us gather data

Development 3rd March 2013

Development 3rd March 2013

I think this qualifies as being a release. What’s working now?

  • A consistent way to decide how much land to give to the four crops (maize, peanuts, manioc and coffee). Circles are probably not the way forward but they are easy to code i.e. an agent asks n-of N nearest patches to become a field. Probably better is to draw out square fields from each of the outer corners of the world.
  • An flexible way to stop the simulation running and send a question to the output box that prompts for a response (pressing the A, B or C buttons), or simply a message. The possibilities here are endless e.g.
    • Do want to hire a tractor (the cost is $$$)
    • You’ve been asked to help out on a neighbors farm – do you accept?
    • You’ve come down with a bout of malaria. You won’t be able to farm for several days.
  • By throwing scenarios at the game player, we’ll be able to collect possible real-world responses e.g. if malaria is stopping the farming unit finishing the preparation of the fields (clearing elephant grass, tilling etc) then there could be a way to request help. (Many of the actions will have costs e.g. hiring a tractor will be a large household cost).
  • A more-or-less flexible way to have farmers leave for the fields and do work at a rate where it will be easy to configure the model to match the real world i.e. 1 patch in the model will be about 5 square meters, and I can give different agents (parents, children etc) different rates for doing different jobs e.g. weeding
  • I have put choosers for game levels (niveau) with the idea that I will through in more difficult conditions for the game player to deal with (more variability with respect to weather, cows and health)
  • I’ve also put in monitors that will show forecasts for rain, and market prices for each of the 4 crops. I just have the tricky tasks of finding out what how different rainfall patterns effect different crops e.g. what are the likely yields of the 4 crops in a dry, average and wet year? (Note, market prices might be inversely related to yield; I’ll stick to KISS for now.
  • The central pink/orange/terracotta area is a families plot of land. David suggested important farming happens here e.g. garden crops and perhaps animals
  • Lastly, the game is seeded with a family in this case i.e. 2 adults and 3 children. I understand there are many possibilities here so I’ve made it easy to change how many farmers there are (e.g. more than one wife)

So in summary. Plenty to do but this is starting to feel like I have code with the right structure to go forward. What’s more, when I port it to modelling4all, it will open up the game even more e.g. I will be able to expose variables hidden in procedures as text entry boxes inside micro-behaviours. Also, as I say above, in Cameroon I will hopefully be able to add new agent behaviours as we learn more about the details of the farming calendar.

Posted in Cameroon | 1 Comment

One Response to “Getting closer to having a simple game that will help us gather data”

  1. Kenneth Kahn says:

    Something went wrong with the sentence “I have put choosers for game levels (niveau) with the idea that I will through in more difficult conditions for the game player to deal with (more variability with respect to weather, cows and health)”

    Why not draw the fields etc? It’ll look better and no need for circles. Just make the pixel dimensions of the drawing be the same as the world dimensions.