search

matthewhively / sots_mod

Organize and track all modding changes to Sword of the Stars (mine and 3rd party)
0 stars 0 forks source link

Hiver AI improvement #107

Closed matthewhively closed 4 years ago

matthewhively commented 4 years ago

Hiver should have pulsed fission as a VERY high priority always. Since it will increase their expansion rate by 20% and its pretty cheap.

matthewhively commented 4 years ago

I think this can safely be done blindly (with a simple test to prove that its working)

matthewhively commented 4 years ago

This sort of AI tweak is more involved than I'd like. Looks like we have to add lots of logic to both AItechPrio.csv and AItechMode.csv

Looks like I also have to deprioritize several weapon techs to make room for empire/engine techs. This will undoubtably effect more than just hiver... but hopefully beneficially.

matthewhively commented 4 years ago

Pulsed fission costs 5k, credits needed to invest approximately Liir ~= 350k credits Morrigi/Human ~= 435k credits Tarka ~= 500k credits Hiver ~= 530k credits Zuul ~= 600k credits

For Hiver, is it optimally cost effective to WAIT before building any ships until pulsed fission is researched? Distance 6 system travel time @ 0.5 = 12 turns travel time @ 0.6 = 10 turns

Distance 9 system travel time @ 0.5 = 18 turns travel time @ 0.6 = 15 turns

Distance 15 system travel time @ 0.5 = 30 turns travel time @ 0.6 = 25 turns

Only if it could be researched in less than 3 turns.

matthewhively commented 4 years ago

~OOOO OOO OOOHHHH!!! Lets try setting an explicit command section for gate ships. Make it a copy of std command (so that its cheap). This should fix the problem of putting deep scan on all gates.~

matthewhively commented 4 years ago

~Tech re-prioritization is just about dead until much later. I'll do what I can, but the effect will be limited.~

What I've ended up with is attaching reasonable per-requisites to the techs that they beeline for, and tweaked the priorities as much as possible otherwise. Problematic techs: waldo, battle computers, VRF, poly silicate, reflective coating, sensor jammers, shaped nuclear warhead

Cybernetic interfaces & waldo were almost equivalent in overall effect. So I took their combined cost and split it. So waldo is more expensive, and cybernetic is cheaper.

Made pulsed fission empire,engine and weapon. So that the AI can get it in ANY mode.

Lowered priority on FTL broadband and FTL econ. AI is getting those before they can reasonably be expected to have enough colonies to make trade routes, so they can wait on those a bit longer.

Lowered HIVER priority on colonization techs. If its gonna take them at least 15 turns to get anywhere, they have time to wait before getting those. (Best guess of a decent priority for the other races)