AI War Beta 6.043 "Controllers Out Of Control" Released!

[keith.lamothe]

We don't have to keep the new one but I don't think Chris is going to be happy about "2 or 3 players in the release thread didn't like it" as a reason for "sorry, we can't use that image you just paid that contractor $200 for" 

That would be remarkably close to the sunk cost fallacy.

Only if we first grant the "2 or 3 players not liking something means it needs to come out" fallacy

I think there's more than 2 or 3, but if that's all it was, etc.

[keith.lamothe]

As for Core turret controllers... sad panda. They still force you to do with them what they are meant to support (Aka more distributed defense to buff said defense ).

My advice to you, specifically, is to forget any irreplaceables in the game exist, including core turret controllers, unless you happen to be playing against two Chivalric AIs (which I designed with people like you in mind, have you tried those?  They don't protect from minor faction ships, but that can be mitigated).  From our conversations I'm convinced that you will never, ever be ok with irreplaceables that cannot be placed behind a chokepoint

Not sure how new re-balance will affect that.. they might be able to defend themselves if you get enough of them

Consider for a moment:

1) Turrets are balanced at roughly 3x the effectiveness of fleet ships, cap-vs-cap.
2) There will now be 8 core turret controllers on every map.
3) Each core turret controller enables you to place a full mkV cap on any in-supply planet you wish (assuming your resources support it).

Tell me this: if 120 mk*caps of per-planet firepower is not enough to defend however many of the controllers fall outside your normal perimeter... then what would be enough?

From another perspective: given that 8 planets on the map will have a controller, how many do you think you can enclose behind your normal perimeter?  Surely even 1 would substantially help with distributed defense elsewhere.  At the very least, it's likely that you won't need to spend K on that turret type.

did energy costs remain from mk V?

Edit: oops, meant to reply to this but left a gap instead: the energy costs for the mkVs were, and still are, the same as the energy costs for mkI.

Is there any plan for actual per planet turrets any time soon?

These are those, really, because changing the existing turret types to be per-planet is just a no-go for a very large part of the community, unless the turrets were otherwise untouched, which would just mean chokepoint-level defenses everywhere with no exposed bits (i.e. core turret controllers) for the AI to attack.

[Faulty Logic]

Is there any plan for actual per planet turrets any time soon?

These are those, really, because changing the existing turret types to be per-planet is just a no-go for a very large part of the community, unless the turrets were otherwise untouched, which would just mean chokepoint-level defenses everywhere with no exposed bits (i.e. core turret controllers) for the AI to attack.

I think the easiest/best way to add to distributed defense would be by extending the mini-fortress line.

You would have the riot-control minifort, doing 1.5-2x the ED than the riot mkI at range, with a per-planet cap of 1, and

the not-so-mini-fort with stats roughly equal to half a fortress (maybe a bit cheaper in mce, though), with a per-planet cap of 1.


Additionally, per planet turrets could be unlocked via the new hacking mechanic, when (if) that happens.

[Kahuna]

Additionally, per planet turrets could be unlocked via the new hacking mechanic, when (if) that happens.

Hope not.

[Tridus]

I don't think distributed defense will ever be as good as chokepoints. Fundamentally, distributed defense is best against distributed offense. If the AI had to launch attacks by building ships on each one of its planets at once and forming them up, you could make a really effective distributed defense by giving the AI targets all over the place (and doubly so if you could block off different sections of AI territory from each other).

But the AI doesn't to do that. It warps in stuff in one big attack group and uses a focused fire offense. So long as the tactic is to try to find a vulnerable spot and exploit it, the best thing to do is to leave as few places to attack as possible and fortify them.

[Faulty Logic]

I don't think distributed defense will ever be as good as chokepoints.

That isn't the goal. The goal is to make distributed defense not as obviously-inferior.

[Tridus]

If I lose irreplaceable stuff that was behind a distributed defense model that I wouldn't have lost if it was behind a chokepoint, than it's *always* obviously inferior.  If I'm not at risk of losing stuff, then I'm probably playing on too low a difficulty. And if the turret controllers become so good that you can distribute defenses to the point that you can defend everywhere effectively at once, then the AI doesn't seem like much of a threat at all (on top of obsoleting the normal turrets).

Chokepoints are better because of basic strategy - forcing the enemy to attack where you want them to is good for you. Short of eithre changing how the AI attacks or making turret controllers absurdly OP, there's not much that anybody can do to stop the benefit of chokepoints as a strategy. After all, human armies have used them to great advantage too.

(I guess this is my roundabout way of saying that I'm not a big fan of the new changes. Going out of my way to capture a controller that I can't use in the spot I want it most unless I stop using the other turrets instead just feels convoluted and weird.)

[chemical_art]

If I lose irreplaceable stuff that was behind a distributed defense model that I wouldn't have lost if it was behind a chokepoint, than it's *always* obviously inferior.

The conundrum of chokepoints.

I feel the current dual AI feature is flawed because it only stacks the current AI "good" features, and not the bad at all.

For example, if I choose the first AIs to be "random" but the second AIs to be "chivalric" then the AI didn't attack those irreplacables, that would be a lot more valuable then the current "Duals AIs just make them more harder, rather then more flexible" logic.

[keith.lamothe]

I feel the current dual AI feature is flawed because it only stacks the current AI "good" features, and not the bad at all.

For example, if I choose the first AIs to be "random" but the second AIs to be "chivalric" then the AI didn't attack those irreplacables, that would be a lot more valuable then the current "Duals AIs just make them more harder, rather then more flexible" logic.

I feel the current information-collection-and-reading-comprehension module installed in chemcial_art is experiencing odd intermittent failures, as demonstrated by the tooltip Chivalric has had since (iirc) its addition to the game:

This AI refrains from targeting irreplaceable structures (like Advanced Factories, Fabricators, etc) other than your Home Command Station, and instead gets a bonus to the size of its Cross Planet Attacks.

When present in an AI with a secondary AI Type, this type's "do not target irreplaceable structures" characteristic is dominant despite being a disadvantage.

[chemical_art]

I feel the current dual AI feature is flawed because it only stacks the current AI "good" features, and not the bad at all.

For example, if I choose the first AIs to be "random" but the second AIs to be "chivalric" then the AI didn't attack those irreplacables, that would be a lot more valuable then the current "Duals AIs just make them more harder, rather then more flexible" logic.

I feel the current information-collection-and-reading-comprehension module installed in chemcial_art is experiencing odd intermittent failures, as demonstrated by the tooltip Chivalric has had since (iirc) its addition to the game:

This AI refrains from targeting irreplaceable structures (like Advanced Factories, Fabricators, etc) other than your Home Command Station, and instead gets a bonus to the size of its Cross Planet Attacks.

When present in an AI with a secondary AI Type, this type's "do not target irreplaceable structures" characteristic is dominant despite being a disadvantage.

 

You got me. I admit it.

[keith.lamothe]

You got me. I admit it.

Just wanted to make sure you (and others) knew what Chivalric was there for.  It's specifically for certain players I had in mind (like you and orzelek) who are sick and tired of either not being able to use irreplaceables or inevitably losing them (in their particular preferred level of difficulty/chaos, at least).

And the Dual-type thing is largely so that you can play with Chivalric on but also get the variety of the other types you might pick.


Bouncer is another one kinda like this: sick of wormhole guard posts?  Set both to at least partial-Bouncer.  You get shot up a lot more the first time going through any door, but it can be a nice change of pace.

[Kahuna]

it only stacks the current AI "good" features, and not the bad at all.

Well this is true but is it a bad thing? ..I don't know. At the moment I don't care. I like the system so far.
Mad Bomber has 2.00 & 0.70 Wave & Reinforce Multipliers
Turtle has  (1.00) & 1.00 Wave & Reinforce Multipliers

Combine those and you get an AI with 2.00 & 1.00 Wave & Reinforce Multipliers.
=
"does extremely brutal attacks on the players" + "Planets are all very-to-incredibly heavily defended"

[orzelek]

My main problem with chokepoints is... that I find them a bit dull.
You get to chose that one and the only built up anything you can and thats it.
Then you start flying around doing stuff and laugh at AI trying to attack your fortress.

There are features that prevent that - beachheads and the warp me there stuff but they are not preventing you from having a major choke - they make only a slight disadvantage.

So I'm using chokepoints as needed and trying to stir some ideas here trying to make other approaches more viable.

[keith.lamothe]

New one

[keith.lamothe]

FYI folks, you weren't crazy when you thought the new title screen had low-res issues: when we approved the image from the contractor we found that it was sized smaller than we requested, so as the final step we asked for the proper size.  Turns out they achieved this simply by scaling up the image they had, when we thought they'd been working at a much larger scale and scaling down to match.

Still figuring out this whole contractor thing, you see

Anyway, the new title screen is still available in 6.044 (released earlier today) but it defaults to off.  So if you really like it you can still have it.