Poll: what most needs attention for 6.0? (Or: let's see if we can break polls!)

[TechSY730]

There's probably one on a mantis issue somewhere if I took the time to look but I'm trying to stay in the code: any saves out there with big threatballs lying around that could be doing something more useful?

If you just need to see a save with lots of threat, I can conjure up one using cheats ("Pull the lever, Kronk").
However, if you want to see what normal threat balls can look like with certain natural styles of play, yea, that would take longer.

[eRe4s3r]

I had a good laugh when looking at the currently leading items. They are both identical issues in a sense. And I don't think there *is* a real solution to either.

Code: [Select]

Mechanics: make threatballs do something else if they don't attack after a while

Code: [Select]

Balance: Making multiple-ingress-points (including sattelite worlds) not so much less viable than singe-chokepoint
The budding AI War gamer knows. The reason threatballs don't do anything is because you choke the Ai with a fortress world of doom, where nothing, ever, leaves that enters. And because intelligent warpgate channeling makes it so warp raids only come in from 1 side of the "wall of doom" they see the fortress wall.. and do nothing. You just need to mop them up. (Especially with raids to )

And to the "open the floodgates" I just say, let them come, and rain against the fortress world of doom. Har har ;P

[Hearteater]

The problem with threat balls is you can get them stuck against a fortress world of doom, and leave every other system basically undefended and the threatball never moves.  It has more than enough ships to break through into your systems if it just picks another front door.  But it never will.  My Neinzul AAR abused this heavily.

[keith.lamothe]

The problem with threat balls is you can get them stuck against a fortress world of doom, and leave every other system basically undefended and the threatball never moves.

This is about to change

[orzelek]

Are threat balls aware of hybrids and vice-versa?
Making them cooperate a bit would make both of them more dangerous.
And bunch of hybrids braking in with large amount of threat would be Fun

[keith.lamothe]

Are threat balls aware of hybrids and vice-versa?

They're both aware of how much total AI firepower is on their target planets, but they don't explicitly communicate or coordinate   I think it's better this way in terms of emergent logic, but I'll consider it.

[Draco18s]

I was wondering if we could get the threat-ball logic dropped down to difficulty 6+, to give the middle-area in difficulty more AI smarts.  Seems like all the better-logic has gone to the 7+ department.

[keith.lamothe]

I was wondering if we could get the threat-ball logic dropped down to difficulty 6+, to give the middle-area in difficulty more AI smarts.  Seems like all the better-logic has gone to the 7+ department.

I can if people want that, it's one character in the code, I just generally put the "smart" stuff on 7+ because 7 is supposed to be when the AI is using all of its "intelligence".

I'll go ahead and move it to 6+ and see if anyone ever actually complains about it being there

[Draco18s]

I was wondering if we could get the threat-ball logic dropped down to difficulty 6+, to give the middle-area in difficulty more AI smarts.  Seems like all the better-logic has gone to the 7+ department.

I can if people want that, it's one character in the code, I just generally put the "smart" stuff on 7+ because 7 is supposed to be when the AI is using all of its "intelligence".

I'll go ahead and move it to 6+ and see if anyone ever actually complains about it being there

I just suggest it so that the "smart stuff" gets a little distributed, rather than this cut-off threshold between the "dumb AI" and the "smart AI" but providing a bit of a gradient.

[eRe4s3r]

Are threat balls aware of hybrids and vice-versa?

They're both aware of how much total AI firepower is on their target planets, but they don't explicitly communicate or coordinate   I think it's better this way in terms of emergent logic, but I'll consider it.

bad temptation that is, we can always build a 2 staged fortress world as lure that way we can channel the threat to a specific location (ie, world) and crush it all the easier.

For the records.. on DIF7 i never seen any threat get through my fortress world designs (with a neutral world as buffer zone to confuse the AI). That includes 4000 MKIV ships and 2 golems at once. I may not be able to entirely stop the "rush" to the other wormhole but behind it is another staged defense. (And usually ships ,p)

Did I ever mention I play AI War just because of the defense aspect?

I think, there is no realistic way to make chokepoints not the total blockade that they can be (and thus, be a thing the AI can't handle). If you think about it, the reason the AI is (sometimes) strong is that it can respond to your attacks in escalating ways, but a fortress world is not an attack, it is a magnet for threat. It sucks the threat up, and eats it. Because the AI may be aware of fire-power, but it is not aware of defense placement. And it will always get lured in, and believe me, it never gets out alive. (I seem to be always lucky whith the trader visits ,p)

While the threat is sucked up by a fortress world, the threat is spread less across the rest of the galaxy, which makes a fortress world and a neutral buffer (with as many connections as possible) always an absolutely superior position. ^^

To fully understand the problem. You have to think like this. The AI is like a collection of water, strong forces of the player are gravity wells. Around gravity wells, the AI will accumulate. And this describes the basic premise of AI War, and the main AI flaw we can exploit unless the "math" is no longer in our favor. (10/10)

If you change in any way the AI responds to strong forces near it you create a hole in the wall. The AI is on alert (Accumulates near us) because it is vital. If it doesn't do that, it would be weak where we are very strong. If you shift where threat goes you also shift how players project force. Right now, I need to keep a certain ship and production force (which by the way is the far stronger "force" a player has) in my backyard for any kind of real threat that appears. But if you change threat to spread, I don't need that anymore.

Hence Why I think, there is no realistic way you can make single point of ingress worlds NOT a huge problem for the AI and mess up it's threat (alert/raid/cpa) responses. It's just the way AI War works.

[keith.lamothe]

Yea, making an invincible chokepoint on diff 7 isn't all that hard, if the map permits it.

There are any number of things I could do break it up, but I don't think you'd like them

[eRe4s3r]

Well, the point of them is not to be invincible, it is to suck up threat Threat that is not applied to worlds where it should be applied. Ever since you changed how the AI responds to Neutral zones the game has become much easier on DIF 7 imo ;p I literally only need to carve out certain neutral worlds to keep as many worlds (that are not MKIV or something..) as possible on alert. If everybody is on alert.. nobody is alert. ;P

The change to CPA's only even reinforced that, because the threat is spread on many worlds. Back before neutral worlds triggered alert you had to be close nearby to trigger alert. Which made it much harder. (because if you were nearby, it was because you wanted to attack ;p) and the only worlds where you wanted to attack for real, were MKIV or Homeworlds. If those switch to alert as single planet on alert you are in for a tough fight.

[keith.lamothe]

Well, the point of them is not to be invincible, it is to suck up threat Threat that is not applied to worlds where it should be applied. Ever since you changed how the AI responds to Neutral zones the game has become much easier on DIF 7 imo ;p I literally only need to carve out certain neutral worlds to keep as many worlds (that are not MKIV or something..) as possible on alert. If everybody is on alert.. nobody is alert. ;P

The change to CPA's only even reinforced that, because the threat is spread on many worlds. Back before neutral worlds triggered alert you had to be close nearby to trigger alert. Which made it much harder. (because if you were nearby, it was because you wanted to attack ;p) and the only worlds where you wanted to attack for real, were MKIV or Homeworlds. If those switch to alert as single planet on alert you are in for a tough fight.

Well, the threat changes in 5.090 might change things for you, we'll see

As for neutral alert, that's something of a two-edged sword in harder scenarios.  Either way, if you're taking a planet and not holding it you're not getting everything you paid AIP for, and if you can absorb that on a wide scale it's mainly a sign of the difficulty you're playing on.  Which is fine, just saying why that's not as big a deal on the harder levels.

[eRe4s3r]

You have a point that it's also about the settings chosen.. ^^ As I play without timed AIP increases which is a pretty big factor on how AIP development and fortress worlds work out.