- So I entered a 1v4 in against a faction leader, me alone and thats really it. The AI dropped 2 of their commanders on top of my main planet, both dying to my AA. Then one of the commanders that survived used its own planet to crush its other allies base and get themselves both killed, Thus I won totaly untouched. I want to share the chronocam video but I am a bit new to the way of transferring the footage to here. Anyone care to give me a procedure?
The AI hasn't been trained to properly deal with multi-planet play. And thus, it's kinda bad. Single planet play is however actually very good, mostly, and is still being improved even.
I don't understand the logic of gimping the AI instead of allowing it to gain more tech as you go along. It means that 90% of the matches are 5 minutes long and boring, but then you get to a boss before finding orbital tech and get planet smashed because you can't build and send nukes to the three other planets fast enough, sometimes.
Well ok, I had to do nothing, Im so good my will power forced them to suicide and then used the last of their resources and bases as a collison coarse.
Sorian was murdered by his own creation early on in development, since then it has taken over most of the worlds computers while keeping its in game AI dumb so we don't suspect anything...
I agree with you about the AI, but you may want to change the title of the thread to be less provocative asap. Something like "galactic war AI is terrible".
GW is retared as it is right now itself. Its not a GW... its a galaxy map with 1 skirmish after another. Its not a Galactic War. Its by far the worst implementation of a stretch goal in this game.
Please try and use appropriate ways to describe your thoughts which may not potentially cause offence. I have edited the title and OP accordingly.
The Red 4 moon, 4 AI system? Normally the AI spawns 1 AI per planet, meaning that at least one will get into direct conflict/gather scouting data. So this kind of behavior is preactically excluded. I noticed my Support commanders hallyeing my player commander with thier planet, rather then going for one of the two AI held planets. As if it even goes for Player Commanders first when they are allied. I never saw it working the other way around. Could have been accidental colission. This is the one part that might actually be interesting for Sorian, so @sorian, I call upon you. There is something of potential relevance for you in this post - the AI might decide to halley there ally if it lacks active scouting on the enemy position and/or that ally is on a halleyable planet. Taeching the AI that "APU = Important", is really, REALLY hard. And that is just for the "kill enemy APU" part. For a human or in a game as limited as Chess protecting the King is not that hard. 3Dimensional Multi Planet games are a whole lot more difficulty. On top of that, afaik Interplanetary Transport AI is currently not handeled by a Neural Network at all, wich accounts for the "send commander to doom in astreus" part. Orbital AI does not even have a concept of retreating or any aspects of the platoon intelligence we see with the other types of units.
