Poll: Should AI homeworlds be buffed? If so, How?

[TechSY730]

I wonder if something more sophisticated for the homeworld "perma-alert" logic could be in order.
Maybe, if one homeworld loses at least one core guard post (or brutal pick structure), that homeworld (but not the other) is perma-alerted. While one homeworld is under attack and alerted, the other homeworld is alerted (but not perma-alerted, it will go back to previous logic once the other homeworld stops being alerted.) Once one homeworld is destroyed, only then does the other homeworld become perma-alerted, as you have clearly shown yourself enough of a threat to take out a homeworld, and the remaining AI should be very "concerned" for the well being of its homeworld. (Note, once perma-alerted, nothing can take it out of alerted state short of destruction).

This suggestion is intended as a replacement for the "perma-alert both homeworld after first attack" option, not as a replacement for the rest of the options.

[chemical_art]

The problem is that it still doesn't change the problems:

-If one one homeworld goes on alert after the first post is destroyed, if you then retreat, you are encouraged to bait the ai even more then already...
-If the second ai homeworld is alerted simultaneously as the first one is, again, you bait it till you clear it out when you eventually attack it.
-If the other homeworld is alerted after you destroy the first one, you will do what is already done due to the massive aip and neuter the first homeworld, then neuter the second one, then pop both almost simultaneously.

To conclude, it doesn't actually change any current tactics, it just makes certain tactics more necessary.

[TechSY730]

The problem is that it still doesn't change the problems:

-If one one homeworld goes on alert after the first post is destroyed, if you then retreat, you are encouraged to bait the ai even more then already...
-If the second ai homeworld is alerted simultaneously as the first one is, again, you bait it till you clear it out when you eventually attack it.
-If the other homeworld is alerted after you destroy the first one, you will do what is already done due to the massive aip and neuter the first homeworld, then neuter the second one, then pop both almost simultaneously.

To conclude, it doesn't actually change any current tactics, it just makes certain tactics more necessary.

I'm trying to think of AIP and alert as the lore states, how much the AI thinks of you as a threat (AIP for overall, alert for where it thinks it is most in danger).
I admit though, certain concessions sometimes will have to be made for the sake of keeping the game enjoyable.

However, it seems that the issue of "baiting" the AI is still a big problem. Not so much that you can bait, but rather how easy it is to do so and how consistently the AI falls for it.

I do have a couple of ideas about this.

-Make it harder to free stuff (maybe like don't free ships just because they are fired upon, what they are guarding is fired upon, or they fired upon something close. Instead, make things freed when, say, what they are gaurding hits <=50%, or only free only N ships, where N is proportional to the ferocity of what provoked. N can be zero if it was only a small attack that was probably intended to bait or was just a stray ship or something)
-Allow some ships to be on permanent defense, and will never be freed until what they are guarding (or the command station) is destroyed.
-When considering when to enter a planet, have it consider not only the planet they are entering, but also planets adjacent to where they are entering, like a weighted sum based on a geometric progression. More technically, have the effective firepower be sum(sum((R^n) * firepower(planet), for each planet n hops away (with 0 hops being the actual planet to be entered)), for each n from [0, total_hops]) (where total_hops is how many hops out the AI will consider, and R is how much reduction in firepower the AI will consider each further hop to be, for some R such that 0<R<1).

The first and second of these ideas are already on mantis. I may want to post the third one there as well.

Actually, this problem of how easy it is to bait the AI is a big enough issue that it really deserves its own thread...



As for the encouraging a double-kill, that doesn't seem so bad in my opinion. It is trickier to pull off, and with the proper buffs, neither homeworld would be easy to take out even if you have never alerted them before, especially because you have only half your force on each homeworld.
However, the perma-alerting the other homeworld is less important in my idea than the other things, so if it would be too much of a hassle for the player, I wouldn't cry if it wasn't implemented.

[Toranth]

However, it seems that the issue of "baiting" the AI is still a big problem. Not so much that you can bait, but rather how easy it is to do so and how consistently the AI falls for it.

If you're concerned about converting things to threat, then baiting the threat, why not do something like this instead:

1)  Ships become freed to fight the player (attacked, over system strength total, guard destroyed, whatever)
2)  Battle ends (humans retreat)
3)  Freed ships run through a list of systems that aren't currently at max capacity, then randomly pick an object in that system to guard.
4)  Freed ships go guard something else, rather than remain threat.

Only ships above the max reinforce level for the universe would be "threat".  This would also make it MUCH harder to clear and keep cleared AI systems.  If you ordered the re-guard list by importance (HW, core, specials, other) then you'd also pretty much guarantee that the HW and core systems were always "topped off", so to speak.

The problem I see with things like "Perma guard" are that they allow the human to maneuver freely around the system, ignoring AI ships.  Even attacking from beyond counter-attack range, according to some of the ideas suggested.

[TechSY730]

However, it seems that the issue of "baiting" the AI is still a big problem. Not so much that you can bait, but rather how easy it is to do so and how consistently the AI falls for it.

If you're concerned about converting things to threat, then baiting the threat, why not do something like this instead:

1)  Ships become freed to fight the player (attacked, over system strength total, guard destroyed, whatever)
2)  Battle ends (humans retreat)
3)  Freed ships run through a list of systems that aren't currently at max capacity, then randomly pick an object in that system to guard.
4)  Freed ships go guard something else, rather than remain threat.

Only ships above the max reinforce level for the universe would be "threat".  This would also make it MUCH harder to clear and keep cleared AI systems.  If you ordered the re-guard list by importance (HW, core, specials, other) then you'd also pretty much guarantee that the HW and core systems were always "topped off", so to speak.

I like this idea. Although I don't think "re-gaurding" should always happen before max reinforce total for that planet (or galaxy, or homeworlds, or whatever) is hit, I still like having a pretty good chance of it happening. (The chance could be dynamic too, like higher chance if important worlds are poorly defended). However, some of the freed ships should still use current behaviour, or else you will have things like a much easier early game (where AI worlds are least defended, and freed ships wouldn't attack under this boolean cutoff, and miss out where offensive threat shines the most, before you really can establish your defenses)

The problem I see with things like "Perma guard" are that they allow the human to maneuver freely around the system, ignoring AI ships.  Even attacking from beyond counter-attack range, according to some of the ideas suggested.

You'll note in the issue I posted that "perma-guard" does not mean "perma-tethered". Perma-guard ships can still travel across the system like normal guarding ships can, possibly even in cases where normal guarding ships wouldn't. However, unlike normal guarding ships, they do not turn into freed threat when doing so. When there is no more hostile presence, they will go back to their normal guarding duties.
Just wanted to point out I did address this concern.

[Faulty Logic]

I see some votes for "they shouldn't be!"

My argument(s) for why they should be:

In the base game: Once you have a planet within transport distance of a Homeworld, you can hunker down, and send a bunch of sequential bombers-in-transports waves, destroying the core guard posts one by one. The first such wave does trigger the brutal post(s), but once you deal with the fallout, it is easy to just wear the AI down. You can also reduce a HW to just its AI command station for only 16 AIP. This means that you have effectively cheated the 100 AIP increase without attacking the HWs simultaneously.

If you have cloaking raiders, it is even easier.
Cloaker starships up this to ridiculous, especially given that warheads (and spirecraft rams) can be cloaked.

Superweapons are supposed to make the HWs easier, but in practice can make them trivial. Especially penetrators, as they let you bypass the 2 core forcefields+fortress III every HW has. Fallen Spire gets a pass because those ships are supposed to be incredibly powerful, and the AI gets a defence fund to compensate.

The main problem is that routine defence of your empire requires more military force than attacking a homeworld. This leads to regular sense of anticlimax in most cases.

The core worlds: they are supposed to defend the HWs, but in practice are almost always ignorable: send your fleet past them, and they suddenly don't care that you are wreaking havoc on their HW. When the HW is below its ship cap, the core fleets should be reassigned to the HW. Thus it is possible to bait the AI if the players so choose, but is costly and difficult.

I am aware that the last HW buff came recently, and that special forces now defend it. However, I have played six games since the HW buff, and every one of them except one was anticlimactic. The one, meanwhile, was only exciting because of a simultaneous AI attack, rather than the difficulty of my own attack.

Another problem is that your first attack has a chance to release all the AI defenders from its HW, allowing you to deal with them and the nasty HW structures as separate problems. This happens every time if you nuke the HW (this also puts the big, nasty fortress III out of supply).

My ideal version of the HW buff:

MOAR Ships/Guardians: This is a simple, boring increase on its own, but stacks with other suggestions to make attacking the HWs (the entire goal of the whole game) a challenge rather than a letdown.

Anti-cloaking measures: either roving decloaker patrols, or frequent planetwide tachyon pulses. This prevents cheese like having a penetrator waiting to kill the AI command stations, a cloaked warhead group destroying a bunch of posts with no risk, and stuff like sending wave after wave of cheap cloaked stuff to destroy posts one by one.

Extra Brutal Pick: As an option for both, or added to the second HW when the first is destroyed.

Bring Core Fleets to defend: The core worlds are meant to defend the HWs after all, and this would open up new tactical options, and force players to operate in the core worlds, where they currently almost never have to go.

Extra Core Guard posts: Another simple fix to bump up the difficulty without new mechanics.

Provoke Galactic Response: the most important. This makes the wave-after-wave tactics much harder to pull off. Though we want multiple waves, we don't want a grind, which is often the best tactic currently. Ideally HWs should go down in 2-3 waves. This also makes sense thematically: the AI is concerned. It responds.

Perma-Alert if attacked: You have demonstrated that you have power sufficient to challenge the AI. It is not amused, and becomes concerned for its safety. This also prevents players from lightly entering HWs.

Buff Core Posts more: they are still pretty insignificant compared to the Ion cannons, mkV fleet, fortresses, OMDs, ect.

Finally, my pet mantis suggestion: AI defence fleets. These make timing the HW attacks important, and I feel they would improve the entire game, not just the endgame, by supporting the "AI attention is fatal" theme.
http://www.arcengames.com/mantisbt/view.php?id=9137

In conclusion, I want the HW attacks to be engaging, multi-step tactical puzzles, which they, in their current state, are not.

[LaughingThesaurus]

^ So a homeworld should be a worthy final boss to a lengthy campaign, in summary?

[Faulty Logic]

Exactly.

[LaughingThesaurus]

I can get behind that, and I voted for a few things that I support as well. Just stuff I feel would be interesting, like provoking a galaxy-wide response and core world ships coming to the homeworld's aid.

[TechSY730]

I voted once, but I retracted my vote because I am undecided about the perma-alert thing. I think that perma-alert should come into play somewhere, but I feel that the conditions as proposed in the option are a little too much.

[Coppermantis]

Anti-cloaking measures: either roving decloaker patrols, or frequent planetwide tachyon pulses. This prevents cheese like having a penetrator waiting to kill the AI command stations, a cloaked warhead group destroying a bunch of posts with no risk, and stuff like sending wave after wave of cheap cloaked stuff to destroy posts one by one.

This definitely is a good measure, but wouldn't be enough on its own. Preventing Cheese is a step in the right direction though.

Bring Core Fleets to defend: The core worlds are meant to defend the HWs after all, and this would open up new tactical options, and force players to operate in the core worlds, where they currently almost never have to go.

This. In fact, I would say it should also extend to Special forces and idle threat. After all, this is the HOME OF THE AI. The AI should panic if it is threatened and call in everything it can to defend, since this is essentially the AI itself you're threatening.

Provoke Galactic Response: the most important. This makes the wave-after-wave tactics much harder to pull off. Though we want multiple waves, we don't want a grind, which is often the best tactic currently. Ideally HWs should go down in 2-3 waves. This also makes sense thematically: the AI is concerned. It responds.

Agreed, Just like the above. With the Ant analogy that AI War makes, if Ants started actively trying to kill you, you'd want to eliminate those anthills ASAP.

[TechSY730]

Somewhat relevant, and similar to many of the suggestions:
7496: Allow AI to use idle freed ships and special forces ships to aid defense.
It's been implemented for special forces, but it would be nice to see a variant of this logic used for freed ships under certain circumstances (like the AI homeworlds being attacked )

[chemical_art]

With the mark cap size standardizing, ai homeworlds should get around twice as many reinforcements. (?)

[keith.lamothe]

With the mark cap size standardizing, ai homeworlds should get around twice as many reinforcements. (?)

Yea, something like that.

With the way some of you guys play that just means twice as much grist for your turrets, but it's something

[chemical_art]

Good, well then that's a pretty big buff to ai homeworlds.