Ai easly tricked.

[Wingsofdomain]

So if some ai ships decide that they want to attack, they will be shown as "threat" in the ...uhm...hud. However those ai ships won't attack until their fire-power is higher than the adjecent human planet. So what do you do? You move out all your ships from your planet and send them right back. Win.

I think this was fixed during the light of the spire updates but either I was wrong or something changed again. Or the only change that was made was that they won't send in the ships IF the fire-power wnt from human planet to the AI planet where the "soon to be attacking" ships were. Dunno... :S But it's really boring since the only way that tactic would work is if you are unobservent and move some ships through your empire because of some other reason. The ai's tactic that is.

"Bah..pityfull humans...have to answer phone calls."

[keith.lamothe]

It has its drawbacks, but it has fewer drawbacks than the previous behaviors (stuff coming in by ones and twos, and then later bunching up but immediately retreating as soon as more force came back).

And it seems to be pretty effective against most players if the human is having a hard time keeping up with everything.  "Defusing" an AI accumulation takes time and ships and so on, and sometimes you just can't spare them.  Tends to lead to it all coming down on your head when you're not ready

[Wingsofdomain]

Tractor beams.

[keith.lamothe]

Try playing against a Thief AI and a Neinzul Nester AI and you'll learn about tractor beams

[Sunshine!]

Man, early game that Thief AI is the worst.  I would rather play against pretty much anything else (besides an "Everything" AI, maybe...)

Also, if you're having too easy of a time dealing with AI ship buildup, you're probably playing on too low of a difficulty level, or don't have enough other AI plot/minor faction nastiness enabled.  That's part of the beauty of AI War, if you ever think things are too easy there are pleeeeenty of ways to make your life harder in very interesting fashions.

[Red Spot]

Try 2 raider-AIs on a lvl above 9. See if you can get past the initial 10 minutes of the game. Than find out how utterly decimating they can be when they just race past your defences. 

[Bleek]

You can also lure the AI away from key spots and chokes by attacking in the unwanted regions i.e. destroy the base there (can be your own), don't build anything and you have a buffer. Works well for attacking their home if you surround it with empty/null planets. 

[Wingsofdomain]

nonono. You guys seem to missunderstand me. I mean that I use tracotr turrets so shen the ai luanches it attack at least 60 ships will be caught in it and the rest of the ships are already spread out on hte planet anyway. They are stil not powerfull to do any harm and my botnet golem just reclaims them. :S xD So no I am not having a problem with them, I just think its boring to use that tactic against the ai. Poor ai. Can't do anything else than what it's programed to do.

[Entrenched Homperson]

Actually, this AI can. That's the awesome nature of this AI. It operates on a very long and complicated rule set, and a decision is made once it's filtered through all these rules. While that my make you think, "Isn't that just programmed?", it is far more effective at creating unpredictable, natural feeling strategic behavior because all of the behavior is emergent. That means even one of our developers who know all the rules by heart can't tell you exactly what the AI is going to do. (But they can guess!)

This was just one of the simpler, high level decision making rules that became very easy to recognize. Other's aren't so easy, so aren't so exploitable. Which is why the behavior was changed a bunch in LoTS, so as to make that situation of standing fleet v standing fleet on either side of a wormhole more interesting and the rules that govern the AI's decisions harder to discern.

[Wingsofdomain]

Well the ai never learns that some of the fp on nearby planets can be moved. Can the ai move its forces to another planet to attack the human on a less defended planet?

[Entrenched Homperson]

No because of the nature of build ups. If you scout out behind alerted to planets you'll see that they are empty comparatively. This is because they aren't alerted and aren't building up. The ships are assigned to the planet and don't move around the way a player would ships. From the AI's perspective this is just assigning resources to a more demanding area. The AI has an infinite supply of ships, and an infinite supply of ships. It also has an infinite supply of ships. ^^ 

[Nalgas]

The AI has an infinite supply of ships, and an infinite supply of ships. It also has an infinite supply of ships.

That sounds like a lot of ships.  Which of those are countably and uncountably infinite?  I just imagined the AI attacking with an uncountably infinite wave and defending with martyrs and lightning warheads and stuff that can hit a countably infinite number of targets, and I think my CPU/RAM are in the corner crying now.  Maybe that's why the AI decided to kill us in the first place.  Heh.

[x4000]

Well the ai never learns that some of the fp on nearby planets can be moved. Can the ai move its forces to another planet to attack the human on a less defended planet?

If you move back in to attack, as you suggested in the original thread, then the AI will decide to retreat if it has substantial forces, and then it will indeed come around a different direction.  The only time this shouldn't work is if you incredibly outnumber the AI anyhow, and it's a mop-up on your part either way.

[Nalgas]

If you move back in to attack, as you suggested in the original thread, then the AI will decide to retreat if it has substantial forces, and then it will indeed come around a different direction.  The only time this shouldn't work is if you incredibly outnumber the AI anyhow

Or you have enough tractor beams, or EMP mines, or force fields over the wormhole(s), or a black hole machine, or [insert anything else that prevents them from leaving], or...

They do kind of leave themselves open to being baited into traps like that in some situations, and the way they mass up in preparation for that also leaves them open to being taken out with lightning warheads or martyrs before you even have to go through the trouble of pulling them into your deathtrap.  It's still definitely better than them trickling through harmlessly a couple at a time like they used to, because it gets pretty hairy (in a good way that didn't happen before) if you don't make a point of keeping up with it, but you can still do some kind of cheesy things to work around it sometimes.

[x4000]

Fair enough on the tractor beams, mines, etc -- now I see what you meant there.  But, the alternative in the past was them getting clobbered with that anyhow.  And... really, if the AI never falls for any of your traps, where's the fun in making traps?  The AI makes plenty of traps that players tend to fall for, too, so I think that balances out.

That might sound like a cop-out answer, but I'm actually very serious.  If you read about, say, programming games of Billiards, you'll see that one of the biggest challenges there is making the AI play imperfectly.  FPS games have the same problem -- in any FPS game, a true AI would shoot you square in the head on every shot as soon as you were remotely nearby.  There is no reason for it to miss, ever.

So the thing that FPS and Billiards AI programmers are tasked with is simulating how other humans play, which is a whole other form of AI all together.  And, in terms of AI War, there's a little bit of that going on, too -- if the AI is allowed to just count up all your tractors and such on the other side of the wormhole, and decide not to go through because of the presence of them, then that really gets into tricky territory for a lot of players.  Not to mention the complexity of even weighting that compared to other goals -- leading to problems such as being able to hold the AI infinitely at bay thanks to tractor beams.

There's also a very real chance that your plan could backfire -- an AI ship can slip through, or more might have built up than was clear, etc.  Or some of those that come through might retreat and come through somewhere else unnoticed.  So it's not like it's a Press To Win button, and the past alternative was just automatic victory through ongoing attrition of the AI trickle.

All THAT said, I'm not claiming that the current system is perfect by any stretch.  I anticipate extending this logic and making it significantly smarter.  But doing that is a tricky thing, because it's easy to create even bigger problems (see the tractors example) if you're not careful about it.  I don't yet have a better idea on how to handle this specific scenario, and my head's not really in full AI War AI mode at the moment anyhow, but this is probably on of the things that is higher up on my list to improve next time we're in a full-out AI War development phase. 

If folks want to discuss it in the meantime, of course, be my guest -- sometimes a simple rule addition or change can be found that has huge impact in an emergent system like this, so we could potentially change that even before the next big dev cycle.  But I have thought about this one a fair bit over the last 8-9 months, and I think that while there is a solution to this one, it's not so simple.  I always love to be wrong on that sort of assessment, though.