Illusion of Danger/Difficulty?

[ScorpionWasp]

I have been playing for a dozen or so hours and... there's something that's bothering me about this game. I may be wrong about it, I don't know if this intuition is correct but... it doesn't seem like it's possible to lose AT ALL, THEORETICALLY, if you know how to cheese the AI.

Like, as long as you don't take any unnecessary risks, ever - and you're never forced to - it's IMPOSSIBLE for the AI to check-mate you. They keep checking all the time, sure, but the "mate" part never comes. There's ALWAYS a move you can make that'll throw them off (closing a door, jumping behind cover), as long as you always position your agents very deliberately.

It's almost like they need your "permission" (by messing up REALLY bad) to defeat you.

Am I right about this, or am I missing something? Can the AI overwhelm you if you always play optimally?

[Geothermal1159]

Sure same thing with tic-tac-toe or, pac-man or, most any other electronic game that has AI which people have figured out how to cheese the AI into being unable to be a threat of any kind.

There is a reason why it is called "cheesing the AI".

[ScorpionWasp]

It feels different to something like Pac Man though. Sure, there might be a non-trivial, unintentional way to ensure a 100% win rate in procedural Pac Man if you dig REALLY deep on how the AI "thinks", but in Invisible Inc, it's almost like that's part of the rules. It's like you're intended to never lose.

The ghosts in Pac Man are ostensibly trying to herd and corner you, and the rules of the game allow them to herd and corner you, but it seems like the rules of Invisible Inc never allow the corps to finish the deed unless you give them "permission" by doing something outrageously poor like leaving your agents wide open.

[Benkyo]

Well, what difficulty are you playing on? How far have you progressed?

Just as an example, there are specialist guards with 360 degree awareness and 180 degree vision on harder maps. Or guards that throw stun grenades where they think your agent might be. Does that affect your theory any? Playing without visible sight cones also makes things a lot more interesting, IMO.

Certainly, the game is designed to be winnable if you play well, regardless of RNG, but I consider that a feature, not a bug.

[Kootness]

Not sure how long you've played, but the higher security measures in the late game will stretch you. As such, it might be advantageous to risk a little more to maximize rewards in the early game- splitting your team, sticking it through higher alert levels, etc.

[Hekateras]

As per the description on the campaign start screen, the developers consider "Expert" to be the *base* difficulty. Keep that in mind. Easier difficulties are meant to teach you the mechanics without an immediate TPK every time you mess up. I also assume you're playing with the DLC on because that changes things considerably.

That aside, you're not completely wrong. Good knowledge of the guard AI goes a VERY long way in mastering the game. It's why it's possible to beat the game without any (non-story) items whatsoever, and why user-made mods that make the game more difficult lean heavily towards making the guard AI more sophisticated, whether it be by having guards be able to alert each other or by making their field of vision more realistic and punishing. So congrats on your realisation. If you get bored with the gameplay on higher difficulties, I recommend you check out those mods, they should fix things for you. I made a collection of these, feel free to check it out after you've beaten the game on Expert:
https://steamcommunity.com/sharedfiles/filedetails/?id=1926128047

Although:

Originally posted by ScorpionWasp:

Like, as long as you don't take any unnecessary risks, ever - and you're never forced to -

Technically true, but there is an omnipresent pressure to complete as many objectives and loot as much of the mission as you can. If you leave each mission without cracking any of the safes or ditch the main objective, you're going to start feeling the lack of upgrades really soon. You say that kind of flippantly, like it's no big deal, but knowing when to fold them and making that decision is a pretty big part of learning to be a better player. :P Also, I don't want to spoil things for you in case you haven't gotten that far in the campaign yet, but with the DLC enabled, you're going to find it pretty hard to avoid taking risks later in the campaign, because after a certain point you have to loot a specific item off each mission or it'll be game over. Unforgiving map RNG or unlucky scouting early on can very quickly make procuring that essential item very risky.

[ScorpionWasp]

I'm playing on Expert with two modifications:

1 - Iron Man
2 - No red square marks before seeing the source of the red squares, because frankly that feature feels too much like cheating.

I did lose my first attempt pretty early because my grasp on the mechanics was lacking, but I'm on my second playthrough and just about to challenge the final mission... and it just doesn't feel "fair" to my AI opponent. Like the entire game is a formality that realistically can only go one way.

The fact guards are forced to end their turn the moment they spot you, combined with the fact you can escape by simple sidestepping behind cover/out of line of sight, or closing a door, or cloaking, or having an ally mess them, AND the fact guards are entirely reactionary lemmings that don't plan or retain any memories beyond the very next thing they intend to do, nor communicate with each other, AND the fact you can literally read their minds by spending one AP...

Can you even lose this thing? I've evaded guards sent to investigate the dead-end room I was holed in, simply by hiding behind some furniture in the corner. It feels like a farce.

[Hekateras]

Very well-observed about the guards ending the turn the moment they spot you. Not enough people realise how powerful a tactic is anything that forces guards not to expend all their AP. Sometimes letting yourself be spotted on purpose to prevent guards from advancing can be incredibly useful.

Most of the time people here complain about the game being too hard so this is a nice change of pace. :P

You should join us on the Discord, you'll be in good company with the other tryhards and we can help you get more out of the game. https://discordapp.com/invite/aQrXEse
Also, as I did mention already, there are mods that add things like guard communication. There is also an unofficial higher difficulty mode called E++ which may interest you. You have 0 starting credits, 0 starting PWR, all the modifiers are ticked even more against you and it's a total hoot.

[ScorpionWasp]

Thank you, I have joined it.

I have thought a bit more about the issue, so let me conceptualize what I believe the fundamental problem with Invisible Inc is:

There are 3 types of squares in the game, from the viewpoint of the player:

1 - Squares that are 100% safe. The enemy never gains vision of these under normal circumstances and you could leave your agents there til kingdom come. Let's call these Type 1 Squares. (yeah, I know some new guards come straight for you when they warp in, but you know what I mean, it's still trivial to vacate them temporarily/distract the guard/stun the guard, etc)

2 - "Safe for all practical purposes" Squares. The enemy could gain vision of these squares, but should that happen, you have a trivial... Contingency Plan *flicks eyebrows* to deal with it. And then it's just business as usual. Let's call these Type 2 squares.

3 - Squares that are unsafe. The enemy could see you and you wouldn't have a way to deal with that.

There's only one rule you have to abide in order to pretty much guarantee victory in Invisible Inc (regardless of difficulty as far as I could tell): NEVER end your turn in category 3 Squares. Forget the increasing security count, the "threats" it introduces are IRRELEVANT in the grand scheme of things. Take as long as you like, loot everything if you must.

If a mod or game change is to fix Invisible Inc, it has to do something about this fundamental problem. It has to make it so no strategy - and in particular no TRIVIAL strategy - allows a player to be 100% safe, ever.

There, now I think I have turned my vague discomfort into something concrete.

[Hekateras]

Hm, I was with you until you called it a "problem with Invisible Inc", to be honest. Certainly, they could have created a different game with more complex cover mechanics and a less simplified, but more realistic approach to whether the enemies can see you, but I don't agree with the idea that it would necessarily have been a better game or that the game they made is inherently worse off for it. You may notice your experiences are far from universal, many people struggle with getting through the game on Experienced because even as simplified as they are, the mechanics are complex enough that they take a while to master for most people. I just finished a modded E+ playthrough and while ultimately it wasn't a difficult run past the first few missions, it remained a tense one towards the end and the game was very good at making me like my victory wasn't guaranteed and I could slip up at any moment, if not with campaign-ending consequences, then at least result in something that would make things a LOT harder.

I also slightly disagree with your assessment because it takes resources (AP) to get intel on *which* tiles are in what category. It's very common for door and cover placement to combine in such a way that tiles are safer from one direction than the other, meaning you have to decide if you care about the risk from that other direction, whether there are guards there, etc. At other times you have more agents you need to place than you have quality cover. And even with your best intentions sometimes RNG will stick it to you and a guard will run up to investigate right up to you and hole you in. Sure, you can distract them with another agent or a program or whatever, but that takes the right planning and setup and moving all your "pieces" strategically, and that's far from trivial.

I think you'll enjoy the mods, though, it is seriously right up your alley. Advanced Guard Protocol and Programs Extended (in particular the 'hard mode' setting for the former and the 'extended alarms' for the latter) change the exact mechanic that's bothering you here so much. Not directly, but by way of dramatically expanding guard vision cones. So while the mechanics of cover stay the same, the number of safe squares is reduced vastly because now essentially the entire guard force has the peripheral vision field of Spec Ops. Programs Extended even (optionally) adds an enemy AI that actively works against you, much like Incognita.

Programs Extended's "extended alarms" feature also fixes the pesky problem you were referencing before: that after alarm level 6, it's as bad as it's gonna get and not all that bad, really. The existing alarm levels are overhauled and the new alarm levels introduced (up to 8) are a lot more brutal to the point of making both combat and full stealth impossible if you overstay your welcome. I also made a mod called Permadeath which makes agents bleed out in a few turns if shot. A recently released mod called Advanced Corporate Tactics will also do things like have guards be alerted by looted safes, teleport away downed agents, and other fun stuff.

[Kira]

I'm playing vanilla (mostly Expert Plus / Endless Plus, with Contingency Plan) and even after dozens of games I don't win 100%. Especially in the first two maps, but even later on, you can get very difficult random setups. Chokepoints / long corridors can be very bad, you can get stuck between multiple guards. Initially you have very little resources, it's difficult to get started especially with random agents/programs. I just finished an ironman game where no detention centers / exec terminals spawned, so I had only my two starting units until the end - it was quite hard! Once I had three (3) Validates in a row thanks to Faust... And so forth. :)

[ScorpionWasp]

I think the thing is, Invisible Inc presents itself as a Rogue-Like, but at heart it's a Procedural Puzzler. Outside of initial generation, pretty much everything is 99.9% deterministic. I've been referring to the guards as "The AI", but that's a misnomer. The guards have no "free will", they aren't trying to outsmart you. They're deterministic cogs in a machine, to be manipulated much like the pieces in Tetris, or Space Chem, to achieve a desired outcome.

It makes the opposite journey as a Rogue-Like. In a Rogue-Like, where you are NOW is never safe. You have to move and do so as to try and avert death. "Move, or in 3 minutes an invincible, one hit kill ghost will start chasing you" - says Spelunky. And if you think that's an idle threat, that thing WILL kill you, dozens of times. "Move, or you'll be attacked by a ridiculous well armed warship. Over and over." - says FTL. "Move, or Winter will come and you will starve and/or freeze" - says Don't Starve".

In Invisble Inc, where you are NOW is usually perfectly safe. You could stay there and maintain homeostasis basically indefinitely. It's by advancing carelessly that you place yourself in danger. I reckon Impatience is the one thing that kills most players, because as long as you're methodical and thorough, there's always a solution.

[Kira]

When you're well equipped you can win even if you reach Alarm 6. But the challenge is getting there. Try getting to Alarm 6 on a map with a chokepoint with two agents with little more than their starting gear, and not get randomly spotted by any of the 9 guards running around.

[Hekateras]

Technically II aims for more of a "roguelite" feel. You're right about the determinism, that's also very deliberate and part of the appeal. Fun fact, in the early alpha stages of the game, units had hit points and attacks had hit percentages (there are still remnants of this in the code), much like in XCOM, but they opted against this design choice later on. As a result, it's very low on RNG except for some specific things: map generation, some campaign parameters, some search/patrol behaviour, and some daemon spawning.

Originally posted by ScorpionWasp:

[...] as long as you're methodical and thorough, there's always a solution.

I'm genuinely confused why you seem to be presenting this as a design flaw.

Besides, idleness in II will kill you just as surely as in Don't Starve because the map does not stay static. FTM scanners will kick in and draw attention to your position. Alarm level changes or new guard spawns will make guard patrols change and the location you're in may no longer be safe. At the latest, hitting alarm level 5 guarantees that all guards are now alerted and moving randomly.

I am honestly curious how your opinion of the game will develop as you play more. By your hour count I'm guessing you've completed the main campaign maybe 3-4 times? I wonder if your perception will be different once you start experimenting with different loadouts more.

[ScorpionWasp]

Originally posted by Hekateras:

I'm genuinely confused why you seem to be presenting this as a design flaw.

I believe it's down to genre expectations. In a Puzzler that's EXACTLY the sort of outcome I'd be looking for. But the appeal of roguelikes is more... I suppose you could call it "masochistic", but anyway. You don't start another roguelike play session to win (necessarily). You start it to experience an emergent, personal tale in an uncaring universe. A tale that could go ANY way, and usually goes terribly wrong because entropy. It's not supposed to be a sanitized experience, with you as the center, heavily and immersion-breakingly slanted so that you win. Because when that is the case, victory tastes hollow.

Spelunky is in my mind the golden standard for Rogue-likes, and my win rate in that game is something like 1 out of 300. That one win has meaning. Every defeat has meaning.

Originally posted by Hekateras:

I am honestly curious how your opinion of the game will develop as you play more. By your hour count I'm guessing you've completed the main campaign maybe 3-4 times? I wonder if your perception will be different once you start experimenting with different loadouts more.

Lol, 3-4 times? I guess I DO take my time making turns. I have only played two campaigns thus far, both on the same "expert with adjustments" difficulty. The first one ended shortly because I was still in the dark about mechanic nuances. The second one is ongoing, I have just completed the first Contingency Plan mission with the power cell thingie (I had read somewhere the mission before that was the final mission, but I guess that was before the expansion).

But you never get the feeling of "Wow, that was close!", "We had to make many sacrifices to even get out of there alive!", "That was a heroic and brilliant effort. I still don't believe it worked!" It's always, "Oh jesus, they spotted me again and "trouble" is coming, lesse... should I Ping so the idiots immediately forget everything and go back to being lemmings? Maybe cloak and run so they react to the noise... AND forget everything? Maybe close a door over there so they check. AND FORGET EVERYTHING? Seriously, so many highly effective (and ludicrous) options I don't even know what to do."

[Benkyo]

Even though the skill ceiling on many real-time roguelites is much higher than in turn-based roguelikes (at least for me, with my poor reflexes and a lifetime of playing turn-based roguelikes to draw on), the end result should be the same: you get to the point where your win is pretty much guaranteed as long as you play perfectly.

It seems to me that you are just saying you think you've reached that point earlier than you would have liked in Invisible Inc., but I'm not sure why you are making such a song and dance about it before exploring more difficulty options, or even had a win!