---[Epilogue Part 1: Wooo! \o/]---Excited rambling go!
TL;DR: AI War is AWESOME.The game was a roller coaster. A number of times I thought I was boned but decided to give it a shot anyway rather than concede, wanting to see how it would play out, especially after managing to break two Mk.4 planets right out the gate. And it paid off, revealing room for overcoming seemingly-impossible scenarios where most other games I've encountered would have simply told the player "you lose now" because of a blunt numerical strength check. It makes for quite an adrenaline rush and encourages the player to stick it out and try whatever crazy plans they can come up with rather than simply look at the numbers and quit. This kind of "fuzzy space" is all too rare in games lately, so those that have it are beautiful jewels.
On more than one occasion I found myself in a state of hyperfocus while flicking rapidly between battles on half a dozen planets trying to make sure the line held while also controlling my fleet in some apocalyptic battle, splitting the high speed bombers and raiders to rip apart high value targets while keeping the core intact and the Neinzul Enclaves covered and simultaneously manually managing a flock of Leeches and Parasites for maximum wealth redistribution, all the while flicking my glance at the attack counters every 5-10 seconds as I hear distant gunfire to try to keep a feel for what's happening elsewhere and detect any fluctuations that hint that I need to swap planets NOW again.
The insanity of that run-on sentence is the best way I can think of to try to express the sheer flood of data and activity when all hell has broken loose in this game and you're looking 5 moves ahead and trying to act accordingly in the madness.
The euphoria when you pull it off, and suddenly reverse a situation that looked like it was going to end in checkmate in 30-90 minutes at most, is exhilarating, to say the least.
Last but not least, the AI itself is a pleasure to duel. I never once felt like it was remotely unfair or did anything out of bounds (no magical unit arse-pulls or impossible build time cheese here -- other RTS designers, take note!), and found I could frequently glean clues as to how it was reading the board and its plans by watching its behavior. Unit movement is the language the AI speaks, to indulge in bit of poeticizing.
Both in this full game and in the couple of partial test games before it I used to get the hang of the interface and quickly get a feel for some settings, I was surprised and delighted at the number of times it deftly avoided a classic RTS AI pitfall or took advantage of a situation in a logical, evil way. I'll restrain myself from rambling on here about a subject that particularly fascinates me. Once I have more experience studying it, perhaps I'll write an analysis. Very curious to see what tricks it has up its sleeve, I have a feeling I've only scratched the surface. Nerd bliss!
Thank you for an exciting ride, Arcen! I look forward to sinking way too many hours into exploring this game and its intriguing AI further and maybe someday taking a shot at that honest 10/10 challenge. The voice of Aggressive Stupidity has already sweet-talked me into skipping straight to difficulty 8 for my second game and seeing how it goes.
(So far -- look at all these neat new guardians, I wonder what they do- OW!)---[Epilogue Part 2: Self-Analysis and Random Thoughts]---Despite the narrative drama in my writeup, this actually went smoother in some ways than I had expected, especially for the AIP I racked up and my level of inexperience. I had initially started up the game confident I'd see the "You Lose!" popup around 4-8 hours in, especially after I discovered the auto AIP setting I had picked was... thoroughly not recommended. This was redoubled when I saw my starting position had boxed me in with Mk.IV and Mk.III planets.
The AI never successfully set foot on my homeworld after I set up my chokepoints, nor did it ever really squash anything I hadn't already classed as expendable. Granted, this is probably because I watched its movements like a hawk and preempted a number of attacks it was in the process of staging that would have been fairly huge if allowed to complete, especially if they synced up with a wave or CPA, which I suspected was what the AI was intending to do. I also took full advantage of those Mk.V turret controllers so that nothing was ultra-fragile.
The early ZPG was a huge stroke of luck, especially as I didn't have Zenith Traders enabled, so if I hadn't found that one the only other one in the galaxy was on Black's homeworld. It had a definite impact on the fleet build I chose. If I hadn't captured it, I would have had to have shifted some K into unlocking the higher marks of Economic command stations to try to fuel all those starships. That would then have been a bit more of a pain to defend, but not necessarily fatal.
I go back and forth somewhat on it, but I probably shouldn't have captured that cluster of planets recorded in the "Operation Hindsight" post. The +80AIP for capturing those had a definite impact on how the rest of the game went, especially with how Preemption fed on it. Giving Preemption less to work with would have prevented or at least delayed the eventually uncontrollable threat population growth, which would have meant fewer warheads used to bomb it back into line, which would have meant less AIP...etc. The start of the fairly slow moving, but detectable AIP spiral was easily traceable to that point. On the other hand, I needed the K and the resources... The captured unlocks were a mixed bag. The Sniper turret controller was godly, and the Decoy Drones were noticeably useful. The others in that cluster weren't worth the AIP hit under the circumstances.
What might have been a better move would have been to have hacked the Missile turret controller on Boulokai and left that planet to the AI, freeing up 20AIP. That one became a thorn in my side to defend with an unfavorable wormhole arrangement, and the AI's habit of spawning Warp Guardians up in that northeast corner made gate raiding not helpful. I probably would have been better off not taking Oololon, as my plan of being able to shift defenses off of Ops ended up being less doable than I had hoped, which would have been another 20AIP clawed back. Amphitrite was useful for interfering with AI movement into my territory so I probably would have still taken it, but hacked the Sniper controller so it would have been less concerning when targeted. Barnard's Star had the Decoys and a decent chunk of resources so that probably would have been taken too. The other 2 back there should have been neutered and probably K-hacked to still get the research, especially as the ARS back there wasn't useful. All the other planets I took were necessary and I don't regret, as they had either a CSG, vital capturable, or were needed for security or homeworld attack staging.
If I had gone with that alternative plan, I suspect I would have ended the game at least 150 AIP lower, between the planet AIP, warhead AIP, and auto AIP from taking time to capture them or cull additional threat. The end result would have been unavoidably high regardless, due to the aggressive auto AIP, but it would have been not quite so nuts. On the other hand, I *did* get away with it, and it *was* fun.
And yes, I used way too many warheads, I know. But they are just so much fun... And the high auto AIP changed the calculus in using them pretty significantly from the norm, as time refleeting and repairing, or slowly gnawing away targets with guerilla raids, was far more costly to me than in low auto AIP games. If a Mk.I Lightning would save me 10 minutes, it was a net gain of 1 AIP in the end. Conclusion? Let the bombs fly! I used a few early on that were higher mark than I needed, but I hadn't gotten a feel yet for how much force was needed against what, so that was simply a natural cost of learning. By the end I had a much better sense of how to use a sledgehammer surgically.
Hacking. I should have hacked a bit more, some of which I covered above, though by the end I was up to Moderate response from the AI. Sabotage hacking is pure joy, I did so much of it that by the end, it would have cost me 100 HAP to fry another target. I regret none of it. Huge boon in taking the first AI homeworld. Didn't do much of the other hack types as they involved either ship types I didn't care if the AI had (corruption) or cost more than I was comfortable with and took too long (factory). I did do some fab hacking, and should have done more. I only downloaded one design, the Zevastator, which paid off nicely. My swarm of Parasites removed the need to download anything else -- why buy the cow, when you can get the (fleet) milk for free? ARS hacking I would have liked to have done, I missed one or two ships I would have liked that way, but forgot to peek first and also that is really, really pricey. And slow. With the ARS's mostly ending up on Mk.III-IV worlds for me, it really didn't feel worth the risk, especially when I could Parasite-steal much of the other options from the AI. I kind of wonder if ARS redirecting would benefit from some cost/time tweaking to make it more appealing as it's one of the riskiest hacks.
Superterminal hack was spicy. There was a minor savescum I glossed over in my AAR as the first shot had a truly absurd number of wild rolls (I counted over 6, three of which happened in the first 10-15 seconds!) and I misclicked my warheads. Second go was fine, I thought the response was reasonable for the conservative amount of AIP I clawed back. I wonder if there should be a cooldown between wild rolls that decreases the longer you hack, so you don't have quite so much of a variation right at the very beginning of the hack. No issue with it being able to spike wildly and that unpredictably after maybe about 15 AIP net reduction, as it would signal nicely that if you want to be safe, you need to pull the plug very soon, and if you go longer, you are clearly asking to play with fire.
Homeworlds. No shame in admitting that the gargantuan difficulty spike in going from midgame to the first homeworld attack caught me off guard, especially as I avoided going through an initial strength check on a core world. My first impulse did have a bit of "uh what this is totally different from the difficulty of the rest of this match/this seems unfair due to that" to it, but I didn't permit myself to pay too much attention to that first impression as it was my first time trying it, therefore I didn't know what I was doing and didn't have an accurate picture, plus I like a challenge and that's a factor in why I was playing this game. After poking at it a bit and getting a grasp of how I was going to need to proceed I did ultimately change my mind and don't think they're out of line, if a newbie like me can crack them with higher than it should be AIP, they're definitely not too difficult.
I think that initial twitch "omg is this even the same game?!" reaction could be avoided though, assuming that it's not intended. If CSGs are enabled, it seems like an intuitive approach would be to have the AI reinforce the remaining ones more heavily each time you take one, doubly so for the CSG-A series since it's the most vital one to their security. If the final CSG ultimately ramped up to near-core world levels of pain, it would create a natural difficulty curve and provide a clear "exit interview" check as you prepare to leave midgame. Tying to the CSG option would theoretically not feel intrusive in the overall game design, as players who turn it on are specifically inviting a certain amount of structure and required strength-checking (from the CSGs that will inevitably end up on Mk.IV worlds). Players who don't turn CSGs on presumably don't want any of this and are probably fine with the sudden spike as they probably are familiar with it already.
A final few thoughts on the difficulty and settings I used before I end this ramble.
I'd definitely not hesitate to recommend a brand-new player go straight to 7/7, even without superweapons. The ramp in waves and CPAs was quite generous, even when piling on the AIP, and if you use Simple ships to start with the number of options are by no means overwhelming. There were definitely detectable breakpoints in the AIP scale with regards to the severity of its response, but nothing felt overly difficulty-spiky. Overall seemed pretty relaxed until somewhere between the 230-250 mark. I also notice that the AI seems to not use as wide a variety in Guardian types on this difficulty, compared to what I've seen of 8/8 so far. If you're new to the game, definitely come on in, the water's fine, don't worry about easing in from the shallows.
Auto AIP 1 per 5 minutes. This had a major effect on the game. I did the math and it contributed exactly 181 AIP to my final tally, or 34.8% of my total AIP. To put it another way, even having captured 20 out of 60 planets (counting both AI homeworlds) *and* having used a nuke, I would have ended at 338 AIP instead if 519. Could probably safely subtract an additional 50 as the nuke would likely have been unnecessary (I'm sure I could have avoided it in the actual game, but it looked like fun and I suspected I could get away with taking that shortcut). That would have been downright sedate! It may have been somewhat of an accident, as I was following an outdated tooltip, but in the end I was happy with the result. I liked the time pressure and the constant ominous update from the doomsday clock flickering in the corner of my screen reminding me to do or die. I have a hunch I will habitually play with aggressive auto AIP. I may have to tone it back on higher difficulty settings, but we'll find out when we get there. >D
Shark B on intensity 2 is completely invisible, unless you accidentally trigger it by scrapping a command station on a border world of yours with no defenders hanging around. (That was pretty comical when I found out that scrapping will chum the water for the shark. I'm not sure that should trigger it, as long as there are no AI ships on the planet at the time so you can't use self-destructing a station under attack to cheat Shark, but it was funny.) I think I saw Shark B contributing something like 8 ships to CPAs from what had spawned, looked at my planets and then nope'd off to the threatfleet. I like the idea that the AI will aggressively capitalize on bad things happening to the player so I will probably always play with this one turned on, just bumped up to 4 or so. I'll have to compare it to Shark A sometime.
Preemption is nasty, I'm surprised I don't see this one get talked about more often. Even on intensity 2 it was very noticeable. You get a fairly long grace period before it really starts to bite, but when AIP gets over about 150 you can catch it working. Over 200 and it becomes an increasingly big deal. AIP over 350/time over 10 hours and it was maintaining a constant threat population of 1500+ ships no matter how fast I killed. All plots are definitely not created equal when it comes to difficulty on a given intensity. I liked it, will probably make a habit of usually playing with a few points in this one on at least one AI. (This would be brutal for both to have, especially if you paired it with the threat makes warp gates plot.)
But I think I'll stop there and save anything else for its own threads -- I have an 8/8 match to return to!
