General > Announcements

Darwinbots 2.44.05 and DarwinbotsIM 1.1 Released

<< < (2/3) > >>

Testlund:
May I request another fix for DB?  :D I'm a bit annoyed with the costs....again...still...

It would be nice if the shooting costs had the same accuracy as the moving costs, like 0.01 instead of rounding it to 1 or 0. As it is now shooting costs are too aggressive. I don't think producing enzymes for dissolving food for digestion should be completely free.

ikke:

--- Quote from: Testlund on August 10, 2010, 05:57:12 AM ---May I request another fix for DB?  :D I'm a bit annoyed with the costs....again...still...

It would be nice if the shooting costs had the same accuracy as the moving costs, like 0.01 instead of rounding it to 1 or 0. As it is now shooting costs are too aggressive. I don't think producing enzymes for dissolving food for digestion should be completely free.

--- End quote ---
Are you annoyed with cost resolution (int vs real) or with their relative weight? You seem to be saying a bit of both..

I once did an experiment in an attempt to find cost settings to merge F@ and F3. I ran the f2 and f3 versions of gimmick (f2 modified to have some more naive replenishment strategies) at different cost settings. I did this manually, as I am unaware of an option to run competitions at non F1 cost. I had to increase cost of f2 specifics by a factor of 10 to achieve some sort of equilibrium. Note this was done manually so I iterated between f2 always wins and f3 always wins until I found some  sort of middle ground around 10 for F2 specifics. I probably posted this somewhere here...

Testlund:
What I mean is that shooting cost rounds to whole numbers and not counting in decimals. This means that a bot that shoots constantly will be dead already after 32000 cycles at the lowest cost setting, if it starts with 32000 nrg. Movement costs are not rounded, so you can set it so a bot can swim around for days before it dies without eating, if you want.

What I would like to try to achieve is some conditional logic. Don't produce enzymes if there is nothing to eat.

Shasta:
Shot formation cost has been fixed to use a decimal value.

Panda:
When did you do that?

Navigation

[0] Message Index

[#] Next page

[*] Previous page