Yes the AI is dumber than the Rock, and needs a lot of tweaking. However, what we probably do not see, is the AI is pretty smart, and avoids cones of fire from Over Watchers. 
Actually, I think the panda AI has gotten quite good. The factions, Pure, and Forsaken are still pretty dumb.
AI knowing exactly the Overwatch cone is not a measure of it being smartā¦ it just has perfect information and cheats around itā¦
1 Like
Yes, AI can calculate LOS perfectly and now your overwatchā¦ For that reason, mostly they can avoid my overwatch somehow and yeah itās a anti-tactical cheat.
Animation delay is very retarded too because your soldiers at overwatch are starting the animation with weapon down. Why does a soldier wait enemy with weapon down? Itās very easy to solve. I hope devs handle itā¦
3 Likes
All games do stuff similar to that to make up for the fact that players are better than the AI.
yes, i know. but that doesnt make AI āSmartāā¦ its typical cheating AIā¦ as in every other game
And the fact that AI knows the exact position of OW cone means, that it can avoid your OW most of the time it would be successfulā¦ which just increases the numbers of fails during OW shots, that just increase player frustration with itā¦
So, having a cheating AI in this particular case is not a good ideaā¦
1 Like
It is a valid point. But the player can also see the overwatch cone from AI, at least the direction.
No, I donāt accept that AI can know my Overwatch because itās erasing all the purpose of it. I donāt want to see theirs too because itās not logical.
then AI should also only know the direction, not the exact spots it will activate itā¦
1 Like
Yes that would be fair.
But people shouldnāt call AI stupid, not in this case, if they have more or want more information as the AI.
1 Like
AI is not smart, is not stupidā¦ AI is just a bunch of scripts that check for certain conditionsā¦ nothing moreā¦
Iāve seen AI threading in between overwatch cones with a tile to tile precision, they see exact tiles covered by overwatch not just direction.
Aı is more stupid you think and itās the coders fault. The biggest issue is AI is acting about LoS much more then covers. The problem here, we do not play FraxCom so you can just step aside and you can see that so called hide enemy and shoot to death.
At many times, AI choose out of LoS rather then getting a solid cover and died miserable.
Addition to that, AI does not plan any sneak attackā¦
Imho I donāt understand why these issues still exist. Like shooting at obstacles in overwatch should be easily avoidable as itās known on which exact tiles AI is going to move. Playing animations at the same time, or when itās āidealā for them to happen doesnāt sound like rocket science either. Like I understand that soldiers are not shooting at the same time so that they donāt waster their overwatch on enemy that is going to be killed by another soldier. But this ties back into ability of running game logic forward without drawing it on screen. Firaxis did it well, this is why you can move multiple soldier at the same time, PP is missing this.
2 Likes
In my expierience this is mostly only true for Arthrons. Tritons often use solid cover and Sirens are mostly already moving from solid cover to solid cover, popping out of cover to MC and then back into cover so that you canāt hit them from your current positions.
I think the devs try to implement it so, that it should look for the players that Arthrons are stupid, Tritons a bit smarter and Sirens then even a bit more.
If so, then chapeau. Only bad that the players now complains about stupid AI behaviour ā¦
Tritons especially with sniper rifles try that all the time. They even go on greater distance to get out of your vision range entirely if they have enough movement to do so.
Or I misunderstood what you mean exactly ā¦
Just to make it clear: I donāt want to fully defend what the developers did about the AI. There are still many things that they could improve or could have done better right away. Always ā¦
Absolutely, I can only agree. Something like this this could change many of the bad things that happens only because of the mostly pure separate sequential order.
Some things could well happen simultaneously, as long as the crucial things remain in the hands of the player.
As I suggested above, to shorten these OW sequences it could help a lot if their trigger/animation/movement would run independent on their own.
But generally, just let them shoot sequential or even simultaneously if it fits in the time frame of their movement to get ready to shoot. That doesnāt really matter, because you can handle sequences or even try to get them close to shoot simultaneous with the cones. But they should not wait for the whole animation to be done entirely one after another. Animation about the crucial parts can then be separated even in extreme slow motion to show the player whatās going on.
@sectoid_br had a suggestion how these animations for the players could be achieved some posts above. (Complicated to describe and read accordingly, but I think I understood what it could look like)
2 Likes
I mean ambushes, trying to go back of your soldiers. Sniper AI is good because āIf you can shoot, just shootā AI works fine with that rifle. Maybe the maps are very little tooā¦ There is not much at a small map. I like the AI at Citadelsā¦
Sirens are good as taking LoS because they do not think of real attacking mostly. They use abilities. You should be very near to a Siren to make her attack you.
I saw many Tritons just go to perfect cover and then one step right/left. So when it did the cover right, it left. This is because of the LoS. AI thinks he is at cover but still in LoS and choose to go out from the LoS rather then taking cover.
2 Likes
Agree, not at all, but I also think that too many of them are simply too small to get a real tactical battle.
First off, it is hard to discuss if we had no real picture of certain situations so I can only tell about my own experiences. And from that, I saw this also very often, but almost always because the tile they stood close to a āsolid coverā was in real no good cover because one or more of my soldiers had a good angle to shoot them at this tile (you know, tiles direct adjecent to āhigh coverā means nothing as long as you have an angle about 45Ā° to this tile, then it is equivalent to complete out in the open, they are flanked in this position). Then they move one tile back or diagonal to get better cover to my soldiers as they actually stand. Sure, often I only have to move a little to get again a perfect angle on them, but in their situation it seems to me that it was often the best they can do with the rest of their movement.
But I have also to admit that especially Tritons sometimes do suicidal runs into the lines of my soldiers only to die in the next turn because they spend all their AP to get there and do nothing. That seems to me not to be very clever, but who knows ,maybe the devs have thought this through but Iām simply not able to get what they want to achieve with such moves. Or they simply fail to script something more useful in such situations, who knows. 
Again, there are always things to improve and this counts of course also for the AI.
2 Likes
MadSkunky and @drages both make excellent points, and clarifications. 
These are the kinds of comments that help not only players. But the Devs too, when they see something that will improve their product ā¦ While keeping them employed 
That said. Itād be nice to have Procedurally Created Missions added to those set in stone only for the game plot. I canāt imagine it would take an enormous amount of resources and work to add that in.
1 Like
When maps are little, you always fight as a whole group. So the enemy tries to fight against all your soldiers. No enemy can survive against 7 soldiers with skills used in one turn.
For that reason snapshot goes for bullet sponge method. They want the enemies survive more but the solution is not that. As we got tons of skills to able to attack again and again, the damage potantiel of the squad for one turn is huge, even you kill a scylla in one turn, or 2 for sure.
At fraxcom this was not the problem as the damage per turn was so limited compared to pp, because you just got one attack per turn.
So the AI tries to damage you somehow and avoid all of your soldiers because he knows that he wonāt survive 1 more turn probably.
Result, bigger maps, big nerf for double attack or multi damage skills and delete all damage from random secondary weapon perks. Maybe more foot soldiers per missions for enemy but if they will survive more then one turn, we need better armor to survive too as I always said.
So those mean we nned good overhaul.
Still, enemy should not know my ovetwatch info and they should not attack to the weakest armor soldier even there is much easy target near it with higher armor. This makes using high armor absolute as nobody attacks them when we got already bad armor overall. There should be some solid armor mechanic too which just absorbers damage.
2 Likes
Agree 
Almost :
Rapid Fire, Chainshot, Reaper in combination with Blademaster and Bladestorm, stupidly OP especially on a Templar, Sniper with Serial, Lighning Hands, Faceoff, Fanfire, any combination with Death from Above or even on its own, pretty OP Psi Operatives if you take the time to train them up.
Probably some more that I forgot.
In WotC you are also able to single some of the missions pretty easily if you choose the right character (for instance Templar on Missions with lots of Lost). There are play through on youtube where people completed Legendary / Ironman campaigns with only 2 man Squads. Etc pp ā¦
It is not so that Firaxis did all better especially in the case of āmagicā things, the possibilities werenāt that much weaker, at least in my opinion.
Maybe PP is partially on top of that, but IMO not that much.
1 Like