so what are fair advanced costs?

[ollj]

so what are fair advanced costs to optimize evolution speed & diversity?

primarily to maximize evolution and efectivity of whatever evolutes.
secondarily to keep all kinds of populations alive without one taking over the world.
tertiarily in a small world with as less cancer and bad performance as possible.

Of course I want my bots to pay with energy for junk code by increasing DNA-upkeep (but still want to keep up to +10% junk code without certain death).
Of course I want to speed up evolution by making bots pay for everything that slows down evolution.
But what are the boundaries that work best?

[Numsgil]

Well, the second point, keeping one species from dominating, is quite difficult to do without some clever geographical seperation.  Darwinbots is too simple for speciation.  All non-plants compete with all other non-plants.  It's just an emergent property of the system.

Performance and evolution are in direct odds.  You need high populations for evolution.  You need low populations for performance.  I usually aim for about 2 or 3 cycles a second.  Feel free to play to figure out how many animals that means for you.

Really, you don't have to worry too much about cancerous bots.  Cancerous animals almost always die off.  Cancerous veggies can be controlled with a population limit.

Default F1 settings obviously give you a good starting value for everything.  The most direct way to create more strenous settings is to multiply all the costs fields by some value.

Getting a feel for how much 1 nrg is in relation to bot action takes a bit of practice.  it's just something you need to develop a feel for.

Bottom line, experiment.  No one's really done alot of comprehensive experimentation in this area (it's still rather new).  I don't think you can get a 'wrong' nrg setting.  The bots will either slowly become extinict or rise to the challenge and evolve to meet the challenge of any settings you throw at them.  I don't know that anyone knows the emergent properties of Darwinbots enough to really know what the effects of increasing or decreasing any of the values would be.

[Endy]

I normally find a population that varies between from about 120-60 bots, works best in balancing speed with evolution. Still kind of a trick to keeping them all from an extinction event. Maybe something with more bots would be better, I just get annoyed seeing the sim crawl along.

I have to second Nums on the need to experiment with costs. In general the bots seem to care about direct actions the most, presumably changing the costs of storing and shooting would affect them the most.

Making everything free might also prove as interesting as increasing the costs, since this would allow the bots better chances of surviving.

I kind of like cancerous veggies myself. The toughest of these propigate themselves nicely, without the unrealistic repop.

Cancerous animals require almost a constant food supply or something blocking their "reproducing area"; normally it's only a short lived phenomenon(only in systems where food is freely provided, like our ownselves :wacko: can they survive long periods).