Knowledge Raiding: Too easy/boring?

[Hearteater]

Moving the K-Raiding conversation from GUDare's AAR over here so it doesn't clutter things up.  Basically, are Knowledge Raids too easy, too boring, or otherwise in need of some changes in how the AI responds.  Check out the linked thread for various suggestions.

[TechSY730]

Too easy. Even at difficulty 8 and a mildly high AIP, my fleet can easily hold them off.


Copying my suggestion from there.

Perhaps a variant of the bursting mechanic for superterminal could be used. Instead of the strength varying with AIP reduction and AIP, it could increase with total knowledge collected.

In fact, the non-burst waves should scale with total knowledge collected, just like the non-burst waves of the superterminal scale with AIP reduction. The more you abuse it, the more harsh the AI gets with its response.

The important thing in that note is that whether bursting is added or not, the ferocity of the retaliation scales with how much knowledge has been gathered total, just like with superterminals, the ferocity of the retaliation scales with how much AIP has been reduced.

Not sure how to handle multiplayer though. Should it scale with knowledge gathered only among players that are K-raiding, or the sum among all players?

[keith.lamothe]

Currently I'm thinking something like this:

1) Make the spawns happen at a fixed interval (every 5 seconds, or 10 seconds, probably), but adjust the spawn size math so that the numbers are the same as they are now (5 ships/second on diff 10, 2 ships/second on diff 9, etc)

2) Have the game track the total amount of k that's been raided (not gained total, just gained through raiding) and multiply the size of each spawn by (amount_raided/3000).  Minimum multiplier of 1.0, though maybe 0.5 would encourage a single light k-raid in each game for variety's sake, dunno.

3) Instead of just always dumping the spawn on the local command station, make a d100 roll on the Wild Mage Table like so:
1-25 spawn normally
26-50 pick a neighboring planet with an AI command station as the spawn point and reroll on this table.  If no neighbors have AI command stations, just reroll, allowing for isolated planets being easier to k-raid.  Note that multiple roll/rerolls hitting this range for a single spawn could move it more than one planet away.
51-55 pick a random planet in the galaxy with an AI command station as the spawn point and reroll on this table.  So even complete isolation doesn't guaruntee "crowd control".
56-65 multiply spawn size by 1.5, reroll
66-70 multiply spawn size by 2.0, reroll
71-75 multiply spawn size by 0.1, and make spawn be composed entirely of zombie guardians (if those are normally seedable in the game), spawn
76-80 multiply spawn size by 0.1, and make spawn be composed entirely of raid starships (if those are normally seedable in the game), spawn
81-99 dunno, but I'll think of something to break this range up into
100 set a flag that makes the game immediately repeat the spawn process after done with this one, reroll

As a sanity-check, if it gets a "reroll" result and has already rolled 3 times, it treats it as a 1 and spawns normally with whatever modifiers are there.  Probably increase the roll-limit by 1 per 1000 knowledge already raided or something like that (rather, max_rolls = max(knowledge_raided/1000,3) ).


Anyway, kind of a loony idea, but it wouldn't be that hard to implement so I figured I'd see what y'all think of it

[TechSY730]

Currently I'm thinking something like this:

1) Make the spawns happen at a fixed interval (every 5 seconds, or 10 seconds, probably), but adjust the spawn size math so that the numbers are the same as they are now (5 ships/second on diff 10, 2 ships/second on diff 9, etc)

2) Have the game track the total amount of k that's been raided (not gained total, just gained through raiding) and multiply the size of each spawn by (amount_raided/3000).  Minimum multiplier of 1.0, though maybe 0.5 would encourage a single light k-raid in each game for variety's sake, dunno.

3) Instead of just always dumping the spawn on the local command station, make a d100 roll on the Wild Mage Table like so:
1-25 spawn normally
26-50 pick a neighboring planet with an AI command station as the spawn point and reroll on this table.  If no neighbors have AI command stations, just reroll, allowing for isolated planets being easier to k-raid.  Note that multiple roll/rerolls hitting this range for a single spawn could move it more than one planet away.
51-55 pick a random planet in the galaxy with an AI command station as the spawn point and reroll on this table.  So even complete isolation doesn't guaruntee "crowd control".
56-65 multiply spawn size by 1.5, reroll
66-70 multiply spawn size by 2.0, reroll
71-75 multiply spawn size by 0.1, and make spawn be composed entirely of zombie guardians (if those are normally seedable in the game), spawn
76-80 multiply spawn size by 0.1, and make spawn be composed entirely of raid starships (if those are normally seedable in the game), spawn
81-99 dunno, but I'll think of something to break this range up into
100 set a flag that makes the game immediately repeat the spawn process after done with this one, reroll

As a sanity-check, if it gets a "reroll" result and has already rolled 3 times, it treats it as a 1 and spawns normally with whatever modifiers are there.  Probably increase the roll-limit by 1 per 1000 knowledge already raided or something like that (rather, max_rolls = max(knowledge_raided/1000,3) ).


Anyway, kind of a loony idea, but it wouldn't be that hard to implement so I figured I'd see what y'all think of it

1 and 2. Love it, that is what I am looking for.

3. Meh. It seems weird for knowledge raiding to do this, as shouldn't their priority be to get rid of K-raider? But then again, many of the things on that list would aid that (especially the raid starship ones, which can just go straight for your Mk. III science lab under the FF)
If 3 were to be implemented, probabilities would need to be tuned some. (IMHO, spawn normaly should have the highest chance, so either eliminate the other planet options, or change it to something like 1-35 spawn normally, 36-50 pick neighboring planet...)

[keith.lamothe]

Part of the what makes the current model both relatively easy and relatively boring is that you basically always know what it's going to do, and where it's going to do it.  Going wildly random isn't really necessary, it just seemed entertaining, but I think some degree of "out of the box"-ness will be necessary to keep it from simply being a predictable transaction of "If I build this, the K will come"

[TechSY730]

If you do go for this wacky "roll to choose modifiers" scheme, then would a similar one be introduce for the superterminals? (Maybe with or without the other planet stuff)
The "multiply wave size" and the "try again immediately after spawn instead of after N seconds" modifier model the bursts pretty well, and add new possibilities on top of it.

[Wanderer]

I like 1+3 personally.

On a side note, in regards to my current game, you might want to test a non-Maw defense screen against MK III zombies.  Once the maws fill up it's still not a horrible challenge but the first half of the k-raid is a non-issue. 

What I don't want to see is low AIP games completely hammered by daring to do two or three K-Raids.  I don't mind a sloped increase, I'd hate to see it go from 'that was easy' to 'holy mother of moses' in 2 worlds.  An exponential increase might be valid but it depends on where Keith decides he wants the line drawn as to how to map that curve.

Remember, I'm only on 9.3.  This is only going to get uglier...   

[keith.lamothe]

If you do go for this wacky "roll to choose modifiers" scheme, then would a similar one be introduce for the superterminals? (Maybe with or without the other planet stuff)
The "multiply wave size" and the "try again immediately after spawn instead of after N seconds" modifier model the bursts pretty well, and add new possibilities on top of it.

Yea, could do something there, but the other ideas about "the AI hacks you back" sounded better there.  So it randomly picks one of your planets with energy reactors, randomly picks a type of energy reactor on that planet, and then scraps a random energy reactor of that type on that planet (woe be unto you if the dice roll really the wrong way and it gets your ZPG).  Or randomly paralyzes (EMP's) one or more of your forcefields, etc.

[chemical_art]

Some variety would be great, but at the same time I actually think the level of difficulty is fine.

It SHOULD be easy if you park your whole fleet and use some defenses. This still comes at oppertunity cost:

1. Defenses are used
A. Cost of using the defenses on a oneshot goal.
B. Time to build them up
C. Weakened defenses elsewhere

2. You fleet is teethered to the planet
A. Cost of repairing said ships
B. Lack of offenses, planet clearing, hybrid smashing
C. Weakened defenses elsewhere


Basically, what I am saying that is each choice as cost. If it "seems" easy it often involves great investment. And that is good! And variety is good, too! What I am trying to say is that don't make them hard across the board. Of course, it may have something to do with my 8+ games.


EDIT: So what would work is actually make the threat of counter attack to you a random effect. That highlights and exploits the risk of the blob over the k-raid

[TechSY730]

Yea, could do something there, but the other ideas about "the AI hacks you back" sounded better there.  So it randomly picks one of your planets with energy reactors, randomly picks a type of energy reactor on that planet, and then scraps a random energy reactor of that type on that planet (woe be unto you if the dice roll really the wrong way and it gets your ZPG).  Or randomly paralyzes (EMP's) one or more of your forcefields, etc.

I don't know about you, but I would never use a superterminal if there was even a chance that they could bypass my defenses and scrap an irreplaceable structure like a ZPG.

[TechSY730]

Some variety would be great, but at the same time I actually think the level of difficulty is fine.

It SHOULD be easy if you park your whole fleet and use some defenses. This still comes at oppertunity cost:

1. Defenses are used
A. Cost of using the defenses on a oneshot goal.
B. Time to build them up
C. Weakened defenses elsewhere

2. You fleet is teethered to the planet
A. Cost of repairing said ships
B. Lack of offenses, planet clearing, hybrid smashing
C. Weakened defenses elsewhere


Basically, what I am saying that is each choice as cost. If it "seems" easy it often involves great investment. And that is good! And variety is good, too! What I am trying to say is that don't make them hard across the board. Of course, it may have something to do with my 8+ games.

Good point, it does not need to be too much harder (unless you have been K-raiding like a maniac)

EDIT: So what would work is actually make the threat of counter attack to you a random effect. That highlights and exploits the risk of the blob over the k-raid

The wacky roll scheme the Kieth gave would accomplish that. Sometimes, the retaliation force would spawn on an adjacent planet, or possibly even a random AI planet. If you invested your best defenses into the K-raid and one of those triggered, you would have a situation on your hands.

[keith.lamothe]

What I don't want to see is low AIP games completely hammered by daring to do two or three K-Raids.  I don't mind a sloped increase, I'd hate to see it go from 'that was easy' to 'holy mother of moses' in 2 worlds.

Yea, the idea is that:

1) You're getting part of the gain from a normal planet capture (K).
2) You're not paying the usual hurts-for-the-rest-of-the-game cost (AIP).
3) Instead you're paying an up-front-hurts-right-now cost (ships-in-the-face).

What I'd like is for that replacement cost to start low enough that in most scenarios where AIP is a problem, that replacement cost is easier to bear than the AIP.  Then at 2 planets it's still not as bad but it is more rough, at 3 it starts to really be a mental toss-up of whether you want to deal with the k-raid-shenanigans instead of 20 AIP, and at 4+ you're really starting to wonder if AIP is such a bad thing

This way the main purpose of the mechanic (letting someone dig themselves out of a hole) remains: those first couple planets won't be that bad.  But someone who's _really_ deep in the hole, or repeatedly gets in the hole, has to either face the consequences of more AIP or a lot of k-raid-backlash that gets increasingly nasty and unpredictable.  And someone who's not in a hole but is just keeping AIP down is very much allowed to do that, but there's more of a tradeoff to consider as you do more and more of it.  Also, the actual tactical situation should be more entertaining with something like this, but dunno.

Another thing we could look at is reducing the time required to actually pull off a k-raid, and just intensify the spawns proportionately.  Though I guess you can already do that by building multiple stations.

Remember, I'm only on 9.3.  This is only going to get uglier...   

That's true   Though you really embarassed 9, largely because your AIP stayed low, except for that part with the Core-CPA/Core-Raid where the AI got some of its own back.

[keith.lamothe]

On difficulty: yea, k-raiding is not too easy in general.  It's just boring.  At least, it's not as interesting as the rest of the game, it seems.  My main concern is making sure you're not having to make decisions like "well, this is boring, but it's the optimal way to play the game..." .  That was the impetus behind the previous overhaul of k-raiding too, though that manifested more as a nerf-hammer on the tactic and I don't want to do that.  I just want to make it more interesting (possibly in the sense of "may you live in interesting times").

But I am also concerned about the sheer amount of K that can be raided (and thus AIP bypassed) without the game really even giving you a run for your money.  If you're getting 9000-12000 this way, fine, go forth and prosper and I hope you have fun both doing it and with what the K buys you.  But if you're getting 18000, 24000, etc... that's a huuuuuuge advantage, and that last k-raid is probably easier than the first since you have all the extra toys, particularly if AIP is no higher this time than it was in the earlier attempts because you haven't taken any planets and/or have killed enough DC's, etc.

[TechSY730]

I'm for speeding up the K-raiding pace (aka. increasing the knowledge/second that the Mk. III science lab generates), but increasing the ferocity of each reaction spawn accordingly.

Even if none of these other suggestions are done, I would strongly push for this. Just waiting around at 1 (or was it 2) knowledge per second for 3000 knowledge is agonizing. (that's 50 minutes for 1/s, 25 minutes for 2/s)

[keith.lamothe]

Even if none of these other suggestions are done, I would strongly push for this. Just waiting around at 1 (or was it 2) knowledge per second for 3000 knowledge is agonizing. (that's 50 minutes for 1/s, 25 minutes for 2/s)

The game-time cost is important because you're getting waves during that time, the next cpa is getting closer, the AI is reinforcing, etc.

The wall-clock-time cost is regrettable, though.  Perhaps if the AI provided some entertainment during the wait