Raid engine + Viral Shredders = Von Neumann Machines deluxe

[eRe4s3r]

As the title says, the ai is completely unable to understand that warping in "massive raids" against Van-Neuman Machines is a VERY dumb thing to do. Even if its 50 MK2 fighters, Viral Shredders are NOT weak against fighters - they just die easier, remember that to match this viral shredder fleet, the ai would have to warp in at least 1600 ! fighters all at once and hope that they kill about 25% in 1 minute in that minute these viral shredders have DUPLICATED each 1 time at least the ones surviving for that long.

Heres what happened here -  warped in with 170 ships (original built, with my own resources, i only did 60) - after 25 minutes (and dodging marauders), i have 826 - what you see here is my last "wave" which went from 555 to 826 - in just 5 minutes thanks to the adjacent raid engine (i needed them elsewhere for a while ,p).

Remember, the cap for these shredders is 170 - that means i currently have 5 times more viral shredders MK1 than you could build. This also means, that my viral shredders are about 50 times more powerful now (because now we enter a field where Van Neuman machines become swarms)

This swarm is now evolutionary in the forming phase, it doesn't take long where a equilibrium is reached (i am guessing around 2000 ships) (resources you can harvest vs drones you have vs how much 1 needs to replicate) where the cycles for reproduction become much longer (but that still doesn't mean you can ever hope to beat them)

What should happen here? Well don't nerf them for one. But give the AI info not to warp in raids against a planet where 826 viral shredders sit around.

What can the swarm do?

It can totally wtfpwn everything thats slow and has many HP (starships/sentinels/frigates)

[x4000]

The AI doesn't have scout intel into your planets, so it is not allowed to make decisions of that sort.

[Lancefighter]

The AI doesn't have scout intel into your planets, so it is not allowed to make decisions of that sort.

perhaps its time the AI tried to gain scout intel?

[eRe4s3r]

This isn't my planet, its an enemy planet with an adjacent raid engine My initial investment was 60 viral shredders, i would call this "massive ultra pwnage" because the ai doesn't understand that fighting against Van Neuman Machines in swarm state is suicide

In fact, with these sustained raids i actually can not loose the swarm anymore - because the raid engine raid comes from 1 wormhole (the one i am sitting around here in the screen)

[x4000]

The AI doesn't have scout intel into your planets, so it is not allowed to make decisions of that sort.

perhaps its time the AI tried to gain scout intel?

You're thinking of the AI as a player, when in fact it's a massive bureaucracy.  Besides, I'm not particularly worried about it having some waves going into a single planet and getting slaughtered.  If that happens -- great, more power to you.  That just means that there are more other locations with lesser defenses, and the AI will find those soon enough.  Or, if there are no places with vulnerable spots, then congratulations you have managed to turtle very well -- but as soon as you have to expand some more, which should be fairly imminent, that's likely to break your turtle power.  

Really, this hasn't been something that is too much of a challenge, whereas overengineering the AI is what leads to the crappy AI you see in so many games.  I've written in the past about why sometimes it is a good idea not to fix minor AI glitches in order to avoid much, much worse AI glitches.  Humans make mistakes, and the AIs must be allowed to make little mistakes, too -- trying to make the AI play a perfect game tends to lead to massive exploitability of the ruleset.

[x4000]

This isn't my planet, its an enemy planet with an adjacent raid engine My initial investment was 60 viral shredders, i would call this "massive ultra pwnage" because the ai doesn't understand that fighting against Van Neuman Machines in swarm state is suicide

In fact, with these sustained raids i actually can not loose the swarm anymore - because the raid engine raid comes from 1 wormhole (the one i am sitting around here in the screen)

Oh, now I see what you mean.  Well -- it's not a worry, the raid engine doesn't cost the AI anything to use, anyway, so it's not like it is diverting attention from elsewhere.  Interesting that you have turned it to your own advantage, but normally the raid engine is just meant to be a penalty for players not scouting well enough before they go raiding.

Very interesting situation you've got there, though, that's kind of cool actually -- using the strength of the AI against it is kinda neat, and it doesn't seem like it will lead to too horrible of an advantage for you.

[eRe4s3r]

No.. if this goes on, it'll only lead to horrible lag given that i can probably go beyond 2000 ships (just imagine how this one fares in a cpr of 1000 ships) ;p

This planet is a novum anyhow, its unusual dense with ships + marauders incoming as well, i am just waiting for the marauders to kill everything turret and building, and then i kill the marauders with my swarm as well... haha

Well but this tactic is exactly what Van Neuman Machines are for, i am sure the one who suggested them hoped for them to turn out like this. ^^

When the swarm is evened out i will split them up and put 400 of em around my raid endangered wormholes and go fishin with my *real* fleet (which i am still building up . while i have a little fun here ,p)

In many ways, this is like a breeding sector for them.. haha ^^

[x4000]

Hmmm... I don't recall... are viral shredders able to go beyond their ship cap?  I thought they were capped.

[Lancefighter]

Well... imean..


After like the 5th time it loses 200+ raiders (them things aint cheap!) it should start scouting.. perhaps maybe when it reaches mk2 raids it tries to scout?
what I'm saying is, its too easy at this point to merely stop up the wormhole full of defenders and wait for them to come into the meat grinder.

[x4000]

Well... imean..

After like the 5th time it loses 200+ raiders (them things aint cheap!) it should start scouting.. perhaps maybe when it reaches mk2 raids it tries to scout?

Trust me -- my way is better.  There's a reason why the AI in AI War is generally better than most any other RTS.  If you leave it open and evolutionary, it will generally provide a more interesting, varied game.  If you try to constrain it too closely, it starts getting predictable and trickable.  I know that's hard to believe, but there are nth order effects to any rule that have to be weighed fairly carefully.  The AI in AI War doesn't use historical data or scout data in any way -- it's like a grandmaster looking at a chess board with a fresh pair of eyes each time.  The fact that more RTS AI programmers think they should make fixes like this is why you end up with variants on the "gap in the wall" problem that I've written extensively about (see the articles section).  My approach is radically different, and occasionally you do end up with an issue like the one described here, but in general it does a much better job.  In a situation like that, who's so say that the 6th time wouldn't be massively successful?  Past failure is not predictive of future failure because the board state is constantly changing.  That's another reason why most RTS AI is so ridiculously sluggish to react, whereas the AI War AI -- the tactical AI, I mean -- reacts instantly.

As far as waves go, that's not really AI but scenario building in this case, anyway.  It's all semantics, and I'm sure we could have a very long discussion on this, but I've written page and pages in response to another player who had a similar train of thought about a different issue.  That sums it up well enough.  At any rate, I understand where you're coming from, and general AI programmers would hold that you're correct -- but, again, the reason my AI works better is because it is so off the beaten track.  I'm not trying to be arrogant here, but I get this sort of argument a lot, and there's not much new for me to add at this stage.

[eRe4s3r]

I do really hope you won't cap them, that would take all the fun out of them for the player, and the AI (170 is not enough to sustain a swarm)

But i do have a suggestion for you

Allow a swarm unification

- If player has 1500 Viral Shredders allow them to merge together to a super-massive Viral Shredder Golem , how about that ,)

I am close to 1500 ,p

Yes, that is 400 more shredders in 6 minutes, i redact my statement, the swarm equilibrium is at 4000 against MK2 ships (the ai has tanks = yummy)

Screen 2 This is how it looks up close

[x4000]

Hmm, I really can't do any new units of that scale (or any at all) for 3.0.  The lack of a cap seems like a major oversight to me -- perhaps the cap should be very high, like 1000 or so.  But having a true completely self-propagating unit that has no caps is just ripe for so many ways of abuse that I can't see any way to ever counter all of that in such a manner that wouldn't be game-breaking eventually.

Of course, on the other hand, these guys die pretty quickly, so if you bring them to another planet they might lose numbers and then not recover.  On the flip side, the very fact of you having so many of them makes you less likely to lose any more of them, thus making it so that "those who have, get more" in this case.  I'm curious what your thoughts are, using them in practice in this particular sort of scenario where you have such an upper hand -- I don't think I've ever had them in quite that way.

[eRe4s3r]

I will report back how useful they are, but i don't intend to use them all it gets kinda laggy when you have so many on FRD - besides they do cost energy (1330 cost 133300 energy) the only reason i can sustain them at the moment is that i have a Zenith power generator and still building up fleet

Here is the save, kadiloc is where the fun is at (they are self-propagating for now).

I think you should come up with a way to reduce numbers that is useful to the player, ie - give shredders a 50% resource backflow when scrapping them, instead of 10% - that way the player could harvest resources directly from the enemy (how is that for an awesome idea?)

I think the replicator cap should be at 1000 though - they are still massively powerful against starships but a large amount of defenses can wipe them out.

Starting with 1000 it became a bit laggy on that world If you let it run you could reach infinity.. technically at least

And speaking of that, maybe ramp up energy cost to 200 so its not as easy

Maybe its because i am on 200% + Handicap (yeah yeah, after restarting i want more action faster ,p)

[Lancefighter]

I do recall those pages from your blog.. its just still weird. A chess player knows exactly what is on every single square every single time. however..

the gap in the wall is /exactly/ what i am referring to. 100% of the time I try to leave a single warp gate open for the AI to attack from. This single warp gate is either stocked to the brim with turrets in greater numbers than the raid coming through, or blockaded by the best my fleet has to offer.  this 'gap' in my warp-gate defense leaves the AI with but a single ingress point. And by Jove it takes the bait every single time.

On the bright side, I will actually have a chance to play a singleplayer campaign sometime this week - So i will certainly be pushing myself this time around. No more wimpy little 7.0 and random moderate..

[x4000]

I will report back how useful they are, but i don't intend to use them all it gets kinda laggy when you have so many on FRD - besides they do cost energy (1330 cost 133300 energy) the only reason i can sustain them at the moment is that i have a Zenith power generator and still building up fleet

Oh yes -- the energy cost is an excellent limiter, now that I think about it.  Not perfect, but pretty good for keeping it from being truly an infinite series.

I think you should come up with a way to reduce numbers that is useful to the player, ie - give shredders a 50% resource backflow when scrapping them, instead of 10% - that way the player could harvest resources directly from the enemy (how is that for an awesome idea?)

Hmm, you know -- I really do like that!  Pairing that with a higher energy use by them might cause them to be even more useful for scrapping compared to huge stockpiling.