AI War Beta 7.023-4 "So A Carrier And A MkIII Lightning Warhead Walk Into A Bar

[keith.lamothe]

@Toranth:

I appreciate the testing and feedback.  I'm sure it needs adjustment, but bear in mind:

- alt-response goes up in proportion to the number of champions (it treats 0 as 1 for this); playing 8 champions AND high alt-response is really asking for it

- on diff 10 it should kill you

- it is much worse than the normal response, but that's partly because the normal response is actually really not doing its job (playing-with-champs is basically "easy mode" if you really utilize them).  The champ-exos will be coming to normal-response soon, and possibly more.

- when I adapted the normal nemesis logic I made them be able to happen earlier in the game; it may be that they need to ramp-up from 0 at 0 to full at 2 hours or so

OR, remove all the goodies completely from Alt-Progress, so that Alt impacts nothing but the XP gain rate.  Balance that so that Alt-Response intensity 6 or so is equal to the traditional response, and 10 is no more than a diff 10 Heroic AI.

I'm not sure what you mean.  Even just having only-part of the goodies available was pretty much unacceptable to a significant chunk of the people giving feedback.  That's why I tried to find a way to incorporate basically all of it

Anyway, what I'm thinking right now is:
- have the threatfleet-nemesis component of alt-response get a multiplier that starts at 0 at 0-time and scales up linearly to 1 at 2-hours-time-and-later.
- halve the rate at which threatfleet-nemesis units are spawned for alt-response
- have the effect of each extra champion on alt-response be +50% instead of +100% (so 8 would be +350% instead of +700%) since many of the benefits of higher alt-progress don't really scale up with the number of champs (though the main point of it does)

Reasonable for the next go-around?

[Toranth]

I appreciate the testing and feedback.  I'm sure it needs adjustment, but bear in mind:
- alt-response goes up in proportion to the number of champions (it treats 0 as 1 for this); playing 8 champions AND high alt-response is really asking for it
- on diff 10 it should kill you
- it is much worse than the normal response, but that's partly because the normal response is actually really not doing its job (playing-with-champs is basically "easy mode" if you really utilize them).  The champ-exos will be coming to normal-response soon, and possibly more.
- when I adapted the normal nemesis logic I made them be able to happen earlier in the game; it may be that they need to ramp-up from 0 at 0 to full at 2 hours or so

While I understand that Difficulty 10 should kill you, it's the "You will die in 45 minutes, no matter what you do" that bothers me.  There's not even that false fig-leaf of hope that Diff 10 normally gives you.
There's also the fact that because the goodies are unlocked with higher Alt-Progress, and Alt-Progress should be matched by Alt-Response, you can't get everything until both are at 9+.

OR, remove all the goodies completely from Alt-Progress, so that Alt impacts nothing but the XP gain rate.  Balance that so that Alt-Response intensity 6 or so is equal to the traditional response, and 10 is no more than a diff 10 Heroic AI.

I'm not sure what you mean.  Even just having only-part of the goodies available was pretty much unacceptable to a significant chunk of the people giving feedback.  That's why I tried to find a way to incorporate basically all of it

I meant disassociating goodie unlocks from the Alt-Progress setting.  Aka, if Alt-Progress is on, at any intensity, you get everthing.  The Alt-Progress setting itself only controls the rate the Knowledge coverts to XP in that case.  The extra starbases, etc, may come with hull size or modules unlock count or something else, but not be directly dependent on Alt-Progress intensity.



====
Re: Response

Right now the response is based on:
(Champ_# * ChampHullSize * AI_Difficulty_factor * Intensity_Factor) * (0.72 * 2) = #_Nemesis_per_Hour
AI_Difficulty_Factor for diff 7/8/9/10 is 1/2/3/6 if I remember correctly.
I'm not certain on Intensity_Factor, but at Intensity 1, Factor = 0.25, Intensity 4, Factor = 1, Intensity 5, Factor = 1.5, Intensity 9, Factor = 5, and Intensity 10, Factor = 6.  No threatfleet spawns for Intensity 4 or lower.

For 1 Champion (or zero) frigate at diff 7 and intensity 9 (minimum to get everything), this results in 1 * 1 * 1 * 5 * (0.72 * 2) = 5 * 1.42 = 7 (actually 6, and halfway to the next spawn of 2).
Every hour or two, you'd need to find and destroy all the Threatfleet Nemesis.  If you don't, in 10 hours, you'd be facing 70-ish Nemesis, which is enough to destroy most non-FS homeworlds.  It also puts a severe crimp in attempting to hold any isolated or remore systems.  20-30 Nemesis frigates can wipe out most systems that have only turrets.

The second item is the Exowave.  I tried hard to get one now, and finally managed through even more cheating.

The first double Exowave launches at 5:50 into the game, regardless of difficulty or intensity; second at 10:00.  Looks like the charge rate is not based on difficulty or intensity - How's that work?
The number of Nemesis in the spawn is based on the exo-budget purchase price of 2480.  First exo is 26,000 * intensity_factor points / 2. 
On Intensity 4, this is 5 Nemesis frigates from each AI player.
On Intensity 9, this is 26 frigates from each AI player.
On intensity 10, this is 31 frigates from EACH AI player.
Each subsequent exo looks to be 25% larger.
Each additional Human Champion or hull size increases this Exo size, same as threatfleet spawns above.  So double for 2 champs, etc.  8 Champions, Intensity 9 = 416 total Nemesis Frigates = Game Over, no matter what.  8 Champions, Battleship hull?  Over 4000 Nemesis frigates for just the first Exo.

That means that on Intensity 9, with 1 frigate Champion, at 6 hours into the game, you will face 52 AI Nemesis frigates that will beeline to your Homeworld and attack your Home Command Center.

Anyway, what I'm thinking right now is:
- have the threatfleet-nemesis component of alt-response get a multiplier that starts at 0 at 0-time and scales up linearly to 1 at 2-hours-time-and-later.
- halve the rate at which threatfleet-nemesis units are spawned for alt-response
- have the effect of each extra champion on alt-response be +50% instead of +100% (so 8 would be +350% instead of +700%) since many of the benefits of higher alt-progress don't really scale up with the number of champs (though the main point of it does)

Reasonable for the next go-around?

For adjustments, instead of making it purely time-based, what about making it wait until after the first unlock before starting the response?  So, the first ARS, FacIV, or 4 Fabs gives your Champion and upgrade, and also the AI a countering upgrade so to speak.

Right now, the Alt-Response is also per-AI player, as opposed to being AI 1 only, so basically everything is doubled from what it first appears to be.  I'd suggest picking one of the two players randomly, and apply the response to that player each time.

For the threatfleet response, halving the rate means 1 Nemesis every 16 minutes at Diff 7, Intensity 9, 1 Champion.  4 per hour, roughly.  12/hr @ Diff 9, Int 9, 1 Champ.  That's workable, if there is a cap on the total number of Threatfleet/Special Forces Nemesis out there.  Maybe use:
(Champ_# * ChampHullSize * Intensity_Factor) = Nemesis cap?  Like the multiplier above, but without the AI Diff factor.
That'd max out at 8 * 4 * 6 = 192.  Roughly 200 Nemesis frigates is enough to kill almost anything - so maybe half that?

The multi-champion response tone down would be, hmm, still a total of 45 or so Nemesis frigates in the Exowave for a Diff 7, 1 Champion, Intensity 9 game.  That's likely to be Death for anything but your Homeworld or chokepoint.  If you put a cap on the Exo size, like I suggest on the Threatfleet, that might keep it to manageable limits.  Instead of 45, you'd get 1 * 1 * 6 = 6 Nemesis Frigates in that first Exowave against a Diff 7, 1 Frigate Champion, Intensity 9 player, and 24 when it'd been upgraded to a Battleship.


Part of my dislike for the way this is going, at least the Threatfleet part, is that nome of this is a counter to the Champions.  Instead, it's increased AI response through the normal channels.  The Threatfleet will use those Nemesis to crush isolated systems when you fleet and Champions aren't there.
I feel like I've said this before...  Yup, there was a thread from this time last year, where I expressed the same concerns.  I'm guessing, after reviewing that thread, that the current Alt-implementation came from there... but it still has the real problem of matching tool to counter.
Maybe instead of spawning normal Threatfleet units, it spawns a unit that behaves almost like Special Forces, doing nothing but hunting the Human Champions through AI space?  Add in some Armored Warheads with that behavior - watch the Human player be surprised at the sudden BOOM that wipes out his Champions.  Have units spawn that just camp the Nebula entrances.  Make a unit that has the ability to drain Champion Shadow Power, and have it join the hunt.  Stuff like that, that targets the Champions without being overwhelming.

I kinda like the Exos, as long as they stay small enough to handle with reasonable defenses.

[Fleet Unity]

When you say the “The champ-exos will be coming to normal-response soon” does this mean that the players of the nebula missions will have no choice if they want this or not, or do you mean it will be a option that is selectable in the lobby or scale with AI Difficulty like higher response on difficulty 8, 9, and 10?

[ZaneWolfe]

Questions. If 4/10 is the baseline for most non-AI-Difficulty settings, should not Alt-Response 4/10 be THE EXACT SAME as what the normal champion response should be/is going to be? In theory, should not going Alt-Progress 10/10 give you the same goodies as nebula, with the same estimated growth rate, and 4/10 Alt-Response should give the exact same response as the nebula route? Why is it necessary to have  Alt-Response be the exact same as Alt-Progress? In all honesty 4/10 Alt-Response should be the baseline, where we balance things, and everything above it should be for those who want a bigger challenge. If nebula champs are going to be getting nemesis exos, then 4/10 Alt-Response should have those. Or at least then 5/10 should have them without being any stronger than 4/10, nemesis exos being the only thing that adds. I can understand needing to have Alt-Progress having so much variance, where 10/10 is where it matches nebula champs. Given all that is involved, of course it needs to have a much broader spectrum. Some people want just the champ, nothing more, some want all the toys. But Alt-Response should be balanced like a more typical setting, where 4/10 is the balance point and everything above it is where it gets crazy.


On a somewhat related note, if 9/10 for Alt-Progress is supposed to be where it matches nebula, then perhaps 10/10 should be where it matches the one advantage nebula have currently, duplicate scenarios to offer higher caps of the ally starships. Scale it to map size, but only at 10/10 Alt-Progress. At 9/10 it should give the same number as you would get if you did just enough to get a Battleship hull. 10/10 it scales to map size increases the caps of ally ss. Sure in theory a 120 map going the nebula route could give you 6 extra of the final scenario, giving you a metric ton of extra ally starships. In practice, I've rarely gotten more than 1-2 of the final one. So based on a 120 map, 6 extra nebula, average of 2 per nebula, 9 total caps, 5 factions, so at 10/10, on a 120 map, you should be able to get 3 extra caps worth of each ally starships.

[Fleet Unity]

For the Normal Nebula Champion Ships Exo’s it would be nice if it was an option to have them if you want an extra challenge or to be able to turn it off if you did not want it. I like challenges and balance as well but I think it would be nice to be able to just use the champions sometimes without the AI sending exo waves at you all the time or the nemesis frigate counters at you. Even if you had this option to just have the champion ships with no exo’s or nemesis frigates would be nice even if it reduced your score or disabled achievements for that campaign game. Because Exo waves can be hard for new players just learning the game and if there is no option to turn it off, some players may not like the champion ships as much if you cannot choose how you want to play them. I know this has not been added into the game yet as of version 7.024 for the Nebula Mission Champions, all I guess I am saying is it would be nice to have a choice if you want the exo waves for champion ships or not even if you got a reduced game score or disabled achievements.

[keith.lamothe]

While I understand that Difficulty 10 should kill you, it's the "You will die in 45 minutes, no matter what you do" that bothers me.  There's not even that false fig-leaf of hope that Diff 10 normally gives you.

Sure, and I think the adjustments I mentioned would avoid that degree of severity.  But we'll see.

There's also the fact that because the goodies are unlocked with higher Alt-Progress, and Alt-Progress should be matched by Alt-Response, you can't get everything until both are at 9+.

Yea, I guess I'm just cruising for confusion when I try to use the scale differently for this one group of mechanics.  Basically I want 9 to be "fundamentally changes the game" (as in, gives you everything) but not "dominates the game" like Golems 9 might with its massively inflated exos.  Though I guess the overall impact here was actually more dramatic than Golems 9 might be.

I meant disassociating goodie unlocks from the Alt-Progress setting.  Aka, if Alt-Progress is on, at any intensity, you get everthing.  The Alt-Progress setting itself only controls the rate the Knowledge coverts to XP in that case.  The extra starbases, etc, may come with hull size or modules unlock count or something else, but not be directly dependent on Alt-Progress intensity.

Hmm, that would work for folks who want the champion mechanic to fundamentally alter the game, but not for those who only want basically a respawning hero unit that can get pretty buff but otherwise the game is basically the same.

Though I'm not sure I've actually heard from any of the latter, so perhaps I'm trying to cater to the empty-set audience

The first double Exowave launches at 5:50 into the game, regardless of difficulty or intensity; second at 10:00.  Looks like the charge rate is not based on difficulty or intensity - How's that work?

The timing varies with AIP, the power varies with all the factors that the TF nemesis part varies with (which doesn't include AIP).  Basically the same pattern as golem/spirecraft/etc exos.

Over 4000 Nemesis frigates for just the first Exo.

No problem, right?

For adjustments, instead of making it purely time-based, what about making it wait until after the first unlock before starting the response?  So, the first ARS, FacIV, or 4 Fabs gives your Champion and upgrade, and also the AI a countering upgrade so to speak.

One thing I'm strongly tempted to do is actually remove the concept of getting the unlock sets via capturing/hacking, as I quickly found in testing that it's really annoying when all said capturables start a long way away.

Instead I'd rather they just all come from leveling up, i.e. from xp.  And since xp rate varies with intensity, so would unlock rate.  That way I could actually break up the large "unlock sets" and instead of "for the first unlock set you get 1 new hull type, 1 new module type, and 1 splinter faction if intensity >= 9" I could do something like:
- at level 2, you get 1 new module type
- at level 4, you get 1 new hull type
- at level 6, if intensity is high enough, you get 1 splinter faction
- at level 8, you get 1 new module type
- at level 10, you get destroyer hull size

and so on.

Not sure if that's overall better.  Feedback hasn't really gotten to the point of evaluating the reward mechanism, I think, because it's too busy being eaten by swarms of nemesis frigates.

Right now, the Alt-Response is also per-AI player, as opposed to being AI 1 only, so basically everything is doubled from what it first appears to be.  I'd suggest picking one of the two players randomly, and apply the response to that player each time.

I could, or we can just halve the rate and have it do the double-spawns.  But perhaps the random picking gives more potential variety of experience.

Part of my dislike for the way this is going, at least the Threatfleet part, is that nome of this is a counter to the Champions.  Instead, it's increased AI response through the normal channels.  The Threatfleet will use those Nemesis to crush isolated systems when you fleet and Champions aren't there.
I feel like I've said this before...  Yup, there was a thread from this time last year, where I expressed the same concerns.  I'm guessing, after reviewing that thread, that the current Alt-implementation came from there... but it still has the real problem of matching tool to counter.

Part of the problem is I just don't see why the counter needs to directly address the tool.  When we tried that with Golems (back when bringing a golem to an AI world would cause an immediate multiplied-reinforcement, and the AI would get wave multipliers against any planet with a golem on it, etc) Golems didn't see a lot of use because it just felt like you were on a treadmill: you got a cool toy, but the AI would get an immediate counter whenever you tried to actually use the thing.

So Golems and Spirecraft now don't have any direct counter mechanic.  Sure, exos sometimes go for golems, but even that's been a source of complaints.

It seems like it would make even less sense for the AI's counter units to go straight after the human champions.  What are they going to do?  Kill it?  What does that accomplish?
 
I guess it could send its newly spawned nemesis units on a beeline for the player champion.  That way they wouldn't revert to threat behavior until the champion was dead (as the beeline logic terminates with unit destruction, even if a "replacement" for that unit appears promptly)... but then it's just a matter of parking your champ behind your chokepoint and all the nemesis units die (one or two at a time) on the chokepoint.

Maybe instead of spawning normal Threatfleet units, it spawns a unit that behaves almost like Special Forces, doing nothing but hunting the Human Champions through AI space?  Add in some Armored Warheads with that behavior - watch the Human player be surprised at the sudden BOOM that wipes out his Champions.  Have units spawn that just camp the Nebula entrances.  Make a unit that has the ability to drain Champion Shadow Power, and have it join the hunt.  Stuff like that, that targets the Champions without being overwhelming.

What I don't understand is what that would accomplish, other than make it nearly impossible to use the champ offensively?  What's the fun in "I think I'll take my champ to raid the-" "AI: NOPE"?

Or do you think there's enough room for player sneakiness and with-fleet attacks and so on to still be able to pull off those kinds of operations?

[Fleet Unity]

That was a big post so I did not want to quote it all but in the part were "Hmm, that would work for folks who want the champion mechanic to fundamentally alter the game, but not for those who only want basically a respawning hero unit that can get pretty buff but otherwise the game is basically the same." I do think for some games that would be nice, I like to play on difficulty 7 and no higher I think it is challenging enough, for me. Anyway I do think it would be fun to just have the champion ships without a direct counter to them in some games. I know this is considered too easy by some but I do still find the game challenging maybe I do not understand the point of the post or how to play with them, but all I say is that not all players are what you call "Hardcore" I think that is the term and some would just like to use the champions with out all of the extra response like exo's, ect  (I know alt champ lets you but I mean for those who like the nebula missions) you know like you can select with the Golems or Spirecraft all I am is asking is there be an option to have the extra response if you want it or not like with Golems Easy, Medium, ect. Like I said I may be misunderstanding the post about what the new changes may be. So yes I am one who thinks sometimes it would be nice to have "basically a respawning hero unit that can get pretty buff but otherwise the game is basically the same" maybe not in all games but just to have the option would be nice

[Peter Ebbesen]

For what it is worth, I can chime in with nemesis champion exo waves - as currently constituted - being really rough on new players.

I'm about as new as they come, playing on 7/7 as I understand this is recommended for veteran strategy games players getting their feet wet, and I played my first game a week ago with champion missions. After abandoning that game when I'd learned the basics after 6-7h, I started a new game with alt-leveling of champions, since I hadn't found the champion mission minigames all that interesting but rather detracting from the overall strategic game, but at the same time liked having a single champion ship around to support the fleet and engage in the occasional suicide mission.

Thus I selected 10/10 alt-progress to be sure I unlocked everything, so I'd be able to learn how all the different modules worked, and 6/10 nemesis response, because I did want some sort of counter to the increased power I'd gain, and this seemed the obvious one. Since the tooltip advocated using at least the same level of nemesis as alt-progress, and I chose substantially lower, I expected this to provide an interesting but not dominating experience as I learned the game, or at least one where the strength I gained (one champion hull, champion-bases at home base, extra ships from champion-bases, extra modular fortresses) would have a significantly greater impact on the game than the exo-waves (10/10 vs 6/10). Boy, was I ever wrong. 

I'm a turtling player by heart, but I had seriously underestimated how nasty such an exo-champion wave could be, and the first one - and it wasn't all that large, punched through my defenses. So that was 2nd attempt at the game ended, but at least it was an awesome way to go!

Now, a first defeat is pretty much guaranteed from something, so whether it was one thing or another, it really didn't show much with regards to whether the champion exo was too strong or not compared to other game mechanics - it just showed that I wasn't adequately prepared.

I'm currently 20h and a bit into my third attempt, with the same settings as for the second, and while nothing else the AI throws at me on 7/7 is remotely dangerous in a strategic sense, even at the current 523+ AIP I am running (destructive testing, got to love it; 1083 AIP total, 560 AIP reduction), the ever larger exo champion waves the AI throws at me are truly scary. The static defenses on my homeworld that easily chew through 8-9k normal tier-2 waves without a sweat and with my forcefields untouched wouldn't stand a chance without the presence of most of my navy.

The game has basically turned into a crawl towards victory while learning new mechanics and discovering new tricks, with me aborting whatever I'm doing once champion exo-progress reaches about 90% or so in favour of pulling back the fleet and setting up the newest layers of static defenses or redeploying some fortresses from the various fronts to beef up the homeworld defenses.

I am sure that this would be considerably easier for a veteran player with a better notion of what counters a wave of scores of shadow frigates, but at my current level of playing it is basically an "everything and the kitchen sink" situation, with very little learning beyond "once again, it was enough, but some of the frigates actually reached the forceshields before dying, so I'll probably need even more defenses next time".

Everything on 7/7 with The Tank/Vanilla AIs, 80 planets, Zenith Traders, medium Golems and medium Botnet, alt champ progress 10/10, alt champ nemesis 6/10 a basic setup I'd not expect to be all that uncommon for new players, as it puts the extra risks and rewards firmly under player control.

I still don't really know if the champion exos are too strong - after all, I am surviving them without firmly grasping all relevant mechanics - but I can't help feeling that if they currently are at an appropriate level, the tooltip for alt champ progress (§4) advocating the player choosing at least the same level of champion exos as champion alt progress should be changed. That is currently pretty bad advice.  >D

[keith.lamothe]

For what it is worth, I can chime in with nemesis champion exo waves - as currently constituted - being really rough on new players.

Yep, they're pretty nasty right now   Definitely more balance changes coming on them in the next version.

[Fleet Unity]

I have been reading the forms and I noticed some players seem to like the salvage mechanic and some seem to not like it. I know I like it and the extra challenge it adds with the AI reprisal waves and the extra metal you can get. If there was an option to turn it on or off then players could choose if they wanted the extra income and challenge, or if they turned it off they would not get any salvage or AI reprisal waves throw at them. Thoughts?

[chemical_art]

I have been reading the forms and I noticed some players seem to like the salvage mechanic and some seem to not like it. I know I like it and the extra challenge it adds with the AI reprisal waves and the extra metal you can get. If there was an option to turn it on or off then players could choose if they wanted the extra income and challenge, or if they turned it off they would not get any salvage or AI reprisal waves throw at them. Thoughts?

Against, it would split the community.

Either we keep it, and work to improve it the best we can, or we abandon it and try something else.

Half-...working it will only leave everyone wanting something better.

[Fleet Unity]

Never thought of it that way I see your point.

[keith.lamothe]

The salvage mechanic has only been in the game since 7.011, and received one of the better receptions I ever recall a new feature getting   It would be extremely premature to yank it or make it optional, as it's still in the initial balancing stages.

[Toranth]

Part of my dislike for the way this is going, at least the Threatfleet part, is that nome of this is a counter to the Champions.  Instead, it's increased AI response through the normal channels.  The Threatfleet will use those Nemesis to crush isolated systems when you fleet and Champions aren't there.
I feel like I've said this before...  Yup, there was a thread from this time last year, where I expressed the same concerns.  I'm guessing, after reviewing that thread, that the current Alt-implementation came from there... but it still has the real problem of matching tool to counter.

Part of the problem is I just don't see why the counter needs to directly address the tool.  When we tried that with Golems (back when bringing a golem to an AI world would cause an immediate multiplied-reinforcement, and the AI would get wave multipliers against any planet with a golem on it, etc) Golems didn't see a lot of use because it just felt like you were on a treadmill: you got a cool toy, but the AI would get an immediate counter whenever you tried to actually use the thing.

So Golems and Spirecraft now don't have any direct counter mechanic.  Sure, exos sometimes go for golems, but even that's been a source of complaints.

I wasn't playing back when the Golem-counter you are describing was in place, but from what I remember of earlier discussions it increased reinforcements in any AI system when the Golem was there, right?  That's a direct counter, yes, but if the rate of reinforcement was as high or higher than the rate at which the Golem could kill stuff, then I understand why it'd be unpopular (and inefficient) to use Golems.  Same with the 'increased waves to planets with X' that used to exist.  Unless whatever it was you got was so important/powerful that the increase didn't matter, it wouldn't be worth it.

On the other hand, Exowaves are a "Show up, smack the player around, and depart (be destroyed)" sort of event.  They're infrequent (relatively).  They're based on AIP and Time.  This means inherently capped and somewhat controllable by the player.  You're correct they are not a direct counter, though. 

I guess it could send its newly spawned nemesis units on a beeline for the player champion.  That way they wouldn't revert to threat behavior until the champion was dead (as the beeline logic terminates with unit destruction, even if a "replacement" for that unit appears promptly)... but then it's just a matter of parking your champ behind your chokepoint and all the nemesis units die (one or two at a time) on the chokepoint.

Maybe instead of spawning normal Threatfleet units, it spawns a unit that behaves almost like Special Forces, doing nothing but hunting the Human Champions through AI space?  Add in some Armored Warheads with that behavior - watch the Human player be surprised at the sudden BOOM that wipes out his Champions.  Have units spawn that just camp the Nebula entrances.  Make a unit that has the ability to drain Champion Shadow Power, and have it join the hunt.  Stuff like that, that targets the Champions without being overwhelming.

What I don't understand is what that would accomplish, other than make it nearly impossible to use the champ offensively?  What's the fun in "I think I'll take my champ to raid the-" "AI: NOPE"?

Or do you think there's enough room for player sneakiness and with-fleet attacks and so on to still be able to pull off those kinds of operations?

A matter of frequency and scale, I think.  For example, if the Champion-seeking Warheads only launched after spending an hour or two's worth of anti-Champion charge, it'd be a way for the AI to toss a spike into the plans of the human player on an unpredictable, but not frequent, basis.  Caps help this, since the opposition can rise to 'difficult' but not to 'impossible'.
And that's probably the biggest problem - Nemesis Threatfleet are immensely powerful.  Without caps, it can quickly rise to the point where no non-HW/non-Chokepoint can possibly survive, and at that point it's also almost impossible to destroy the Threatfleet.  They're like Hybrids, but more powerful and without any facilities to destroy to weaken them.

It seems like it would make even less sense for the AI's counter units to go straight after the human champions.  What are they going to do?  Kill it?  What does that accomplish?

That's... a really good point, actually.  When there is little-to-no cost for losing a Champion, it does make it rather ineffective to threaten to kill one.
The "Wait until 200 Shadow Power to respawn" helped prevent the HW-defense spamming of Champions that was silly-powerful, but it isn't enough to actual cost the player anything...
Maybe switch from 'respawn' to replacing a dead Champion with a "Broken Champion" unit - Permacloaked and indestructable, blind, and without weapons, with half normal speed.  You'd need to bring it back to Human territory to be repaired and restored to the normal unit again.  Self-regen on friendly systems, or repair at cost for faster recovery?  I dunno.

The more I think about it, the more it seems this is the most important point - How do you counter something that is disposable?

[Fleet Unity]

The salvage mechanic has only been in the game since 7.011, and received one of the better receptions I ever recall a new feature getting   It would be extremely premature to yank it or make it optional, as it's still in the initial balancing stages.

I know I really do enjoy it, now that I got the hang of it