/tg/Station 13

So, let's face it. SS13 is played in a way, way, way more powergame manner here in TYOOL 2022 compared to when all the rules about not pre-emptively prepping for malf AI were written and designed.

Those rules were written for an era where the crew was not roundstart arming for war ops tier threats every shift and security wasn't allowed to antag check random people via shotgunning them into crit, checking their bags for antag items, and leaving them in medbay lobby dying if they find none. The fact of the matter is that the bar for acceptable powergaming has risen drastically since the rules against powergaming a malf AI were put into place over a decade ago, and they're frankly antiquated compared to modern 2022 spaceman playstyles that are deemed acceptable by the community at large.

Why should we keep these rules against powergaming as and against the AI in very specific ways(AI can't bolt a laundry list of specific doors and rooms, crew can't cut AI's cameras FNR) when we've loosened up the rest of the game and have no other restrictions in place on powergaming anymore?

We used to ban for shit like excessively prepping for antags, but we don't anymore and despite rule 12's implementation the administration(not just the headmin team) has made it very, very clear they do not intend to enforce it against people powergaming the gameplay mechanics to such a degree in every other situation, so why should we ban AIs for excessively prepping against potential crew antags and why should we ban crew for excessively prepping against potential malf AIs?

Manuel is obviously notwithstanding because they're on a different ruleset and enforcement paradigm that actually does handle powergaming, so if I see any arguments about Manuel in here I'm going to report your posts and cry to the forum jannies to delete them.

simply because its boring

Its like the warden adding turrets to the armory roundstart. or the captain carding the ai roundstart. When antags are stifled. You get extended, and not many people want a 2 hour shift where nothing goes wrong. (Unless they have the job content/grindset for it like atmos or xenobio)

AI is piss easy to destroy and its only gotten easier. The only way to survive as a malf AI is to be hidden.

CPTANT wrote: ↑Sun Aug 07, 2022 9:19 pm AI is piss easy to destroy and its only gotten easier. The only way to survive as a malf AI is to be hidden.

Pretty much. Most AI chambers have very clearly designed weak points and as long as a person can deconstruct R-Walls and isn't dumb enough to sit there and get tased you're shit out of luck as an AI, whether MALF or not.

Kitfox wrote: ↑Sun Aug 07, 2022 11:33 pm

CPTANT wrote: ↑Sun Aug 07, 2022 9:19 pm AI is piss easy to destroy and its only gotten easier. The only way to survive as a malf AI is to be hidden.

Pretty much. Most AI chambers have very clearly designed weak points and as long as a person can deconstruct R-Walls and isn't dumb enough to sit there and get tased you're shit out of luck as an AI, whether MALF or not.

Doesnt help that theres a camera console in RD office that lets you view ai core to check for blue apcs when you even get a small sniff of something fishy going on with the ai.

iamgoofball wrote: ↑Sun Aug 07, 2022 7:17 pm Those rules were written for an era where the crew was not roundstart arming for war ops tier threats every shift and security wasn't allowed to antag check random people via shotgunning them into crit, checking their bags for antag items, and leaving them in medbay lobby dying if they find none.

...

Why should we keep these rules against powergaming as and against the AI in very specific ways(AI can't bolt a laundry list of specific doors and rooms, crew can't cut AI's cameras FNR) when we've loosened up the rest of the game and have no other restrictions in place on powergaming anymore?

We used to ban for shit like excessively prepping for antags, but we don't anymore and despite rule 12's implementation the administration(not just the headmin team) has made it very, very clear they do not intend to enforce it against people powergaming the gameplay mechanics to such a degree in every other situation, so why should we ban AIs for excessively prepping against potential crew antags and why should we ban crew for excessively prepping against potential malf AIs?

On one hand, yeah, those old restrictions (can't metagame X, don't bolt Y) are a list of specific things that used to be serial problems for AIs because you were, honestly, kind of stupid if you didn't do them from the play-to-win view, but making everything dangerous or funny hard to get at was obnoxious and unfun. Similarly, carbons probably should have comparable anti-powergame enforced through rule 12.

It's not like people didn't powergame in the era that og silipol was written; it was pretty regular that assistants would rush gloves, toolbelts, medkits, the works. Is it worse now? Maybe? I don't know. I think if you broke into sec equipment as an assistant without a specific plan, then validhunted with it, two rounds in a row would be enough for a "but why" bwoink. These days, you see someone do it (or break in to start fights with sec) six rounds in a row and eat a ban on the seventh.

On the other hand

You're making some very bold and confident statements for someone with 6 hours connected in the last month

and I'm not confident in the truth of several of your assertions about how things are; even if they are true, we shouldn't loosen those restrictions because they make the game materially worse and shut out malf AIs from existing with things like roundstart card checks.

iamgoofball wrote: ↑Sun Aug 07, 2022 7:17 pm Those rules were written for an era where the crew was not roundstart arming for war ops tier threats every shift and security wasn't allowed to antag check random people via shotgunning them into crit, checking their bags for antag items, and leaving them in medbay lobby dying if they find none. The fact of the matter is that the bar for acceptable powergaming has risen drastically since the rules against powergaming a malf AI were put into place over a decade ago, and they're frankly antiquated compared to modern 2022 spaceman playstyles that are deemed acceptable by the community at large.

...

We used to ban for shit like excessively prepping for antags, but we don't anymore and despite rule 12's implementation the administration(not just the headmin team) has made it very, very clear they do not intend to enforce it against people powergaming the gameplay mechanics to such a degree in every other situation, so why should we ban AIs for excessively prepping against potential crew antags and why should we ban crew for excessively prepping against potential malf AIs?

i say this a non-confrontationally and non-sarcastically as i can but do you actually think the game would be better if the players were allowed to powergame against malf ais or are you just annoyed at the admin team again

nianjiilical wrote: ↑Mon Aug 08, 2022 1:16 am i say this a non-confrontationally and non-sarcastically as i can but do you actually think the game would be better if the players were allowed to powergame against malf ais or are you just annoyed at the admin team again

I think that malf AI as a gamemode is held back from being redesigned to be actually good until we remove the training wheels and safety padding that is the AI powergaming rules in relation to malf AI

the fact that malf AI only works as an antag if there's a laundry list of rules saying "you can/cannot do X Y and Z because what if its malf AI and you fuck them over immediately by doing so" means malf AI needs fixed on a design level, but that can't happen until all the training wheels are removed

Pandarsenic wrote: ↑Mon Aug 08, 2022 12:02 am You're making some very bold and confident statements for someone with 6 hours connected in the last month

The shotgun thing with security is something I witnessed in those 6 hours and every time I've logged onto play the video game this year I've been met with absolutely absurd levels of powergaming that, in 2014-2015, would of been slapped down immediately by the admins but is not because our rules and administration methodology has changed.

iamgoofball wrote: ↑Mon Aug 08, 2022 6:51 am I think that malf AI as a gamemode is held back from being redesigned to be actually good until we remove the training wheels and safety padding that is the AI powergaming rules in relation to malf AI

the fact that malf AI only works as an antag if there's a laundry list of rules saying "you can/cannot do X Y and Z because what if its malf AI and you fuck them over immediately by doing so" means malf AI needs fixed on a design level, but that can't happen until all the training wheels are removed

why? what indication have you received for this?

i haven't seen a malf AI win in ages, can we pull the stats to verify my claims

that's not my point, my point is i see absolutely no indication allowing ais to power game harder would mean anyone would redesign malf ai. seems very much like putting the cart before the horse. i agree with nian's implication above that, as is, i see absolutely no reason to think changes in this area will improve anything right now

iamgoofball wrote: ↑Mon Aug 08, 2022 6:51 amthe fact that malf AI only works as an antag if there's a laundry list of rules saying "you can/cannot do X Y and Z because what if its malf AI and you fuck them over immediately by doing so" means malf AI needs fixed on a design level, but that can't happen until all the training wheels are removed

I feel like you can very much generalise this statement to say

The game only works if there's a laundry list of rules saying "you can / can not do X Y and Z".

Like mindshielding the entire crew at shift start. You can do it, mechanically speaking. It breaks the revs game mode. Sometimes this happens off the back of fake revs calls/behaviour and I'm clear with people, be it a revs round or not, that this behaviour has major round antag ruining consequences potentially and that I'm not very happy with them (TM).

If your suggestion is that every game mode that requires admin oversight needs reworking then revs likely needs reworking. As probably does a whole ton of other stuff.

Why's malf any different?

Seems like you want there to be a full code solution that requires no additional rules or oversight, and while that's a nice goal (and the only way this game could ever scale multiple orders of magnitude in a single server base), it's also not where we are, and not where I think we want to be, because it would destroy the very soul of the game in doing so.

iamgoofball wrote: ↑Mon Aug 08, 2022 7:06 am i haven't seen a malf AI win in ages, can we pull the stats to verify my claims

I won two weeks ago as a Malf ai while deltaing, twice actually and they were both on Sybil. Mind you I rarely play ai and even more rarely play on Sybil. Clearly there is more to the dynamic.

also should be noted that the story is more important than who "wins" in the uncaring eyes of code

Mothblocks wrote: ↑Mon Aug 08, 2022 4:12 pm also should be noted that the story is more important than who "wins" in the uncaring eyes of code

Hulks killing the AI by breaking 2 reinforced walls is not that great of a story in my opinion.

tl;dr - I'd really like Malf AI to be less of a hide and seek gamemode because "AI won at delta by having an engieborg RCD it behind a wall" is stupid, but this proposal isn't going to fix that

AI used to lose hide and seek too until the thermal mutation was whacked.

Could not disagree more. Powergaming if anything has gone down since rule 12. Players for the first time ever on TG have been banned and/or noted for playing in gamey sorts of ways. Security isn't ever allowed to just shotgun someone to crit to randomsearch them then leave them dying, what?

What does it even mean for an AI to "prep against crew antags"? Crew is allowed to get whatever gear they like generally but the line is drawn at preemptively hindering antag rounds: i.e., mindshielding everyone roundstart, hiding the ablative trenchcoat roundstart, hoarding all the station pets in armory, cutting AI air alarm access in atmos roundstart (recently someone was banned for this), etc.

iamgoofball wrote: ↑Mon Aug 08, 2022 7:06 am i haven't seen a malf AI win in ages, can we pull the stats to verify my claims

It heavily depends on the malf player and circumstances. I rolled midround malf the other day, kept it under wraps until I had fabricated an advantage, and the only reason I didn't greentext was due to being eaten by a dragon. (And I still had fun even if I lost.)

Randomly cutting AI cams shouldn't be against the rules out of fear of screwing over a malf AI (who actually has a tool to restore cut cams), it should be against the rules because it's unfun for AI in general. It's the same reason randomly stabbing someone's eyes out is against the rules.

I personally wouldn't be against dropping the "no repiping atmos to deny plasma floods" rule, though. Most of the time an AI has an engiborg that can undo such changes, and otherwise the AI might just need to be a bit creative with their mass harm.

zxaber wrote: ↑Thu Aug 11, 2022 9:09 am I personally wouldn't be against dropping the "no repiping atmos to deny plasma floods" rule, though. Most of the time an AI has an engiborg that can undo such changes, and otherwise the AI might just need to be a bit creative with their mass harm.

Yeah to be honest this is a silly rule. Atmospheric technicians not allowed to make atmospherics have a safer setup?

It’s also wonky because they’re allowed to redesign atmos how they like, making it effectively AI-plasmaflood-proof, just not with the explicit intention of making it such.

There's all sorts of levels to the plasmaflood-proofing situation, the clear cut bad case is I feel more where the 'rule' is aimed - Round start atmosian, walks into distro, removes one pump, leaves the department to tide and fight never to return to atmos again.

And sure, there's a whole range between that level and "oh i didn't know, i was just burning all of these highly flammable gasses in the middle of the department(!) and needed that pipe", and somewhere there's a grey area of hidden agenda. But at the same time its just a silly pump over a small chance of pre-empting an unlikely malf AI, so I guess most people will be sufficiently discouraged. You'd have to be a metagamer and a hardcore atmos to pull that off regularly.

But I think the rule covering the openly bad first example is fine. For everything else there is context. Also never had a ticket about this point either way I don't think.

You're totally fine to incidentally make plasma flooding impossible (rerouting all the plasma to a burn chamber, all the plasma and N2O to a Bz chamber, etc.) as long as you're not doing it purely to leave the AI unable to use one of its most important tools.

Shellton(Mario) wrote: ↑Sun Aug 07, 2022 11:51 pm

Kitfox wrote: ↑Sun Aug 07, 2022 11:33 pm

CPTANT wrote: ↑Sun Aug 07, 2022 9:19 pm AI is piss easy to destroy and its only gotten easier. The only way to survive as a malf AI is to be hidden.

Pretty much. Most AI chambers have very clearly designed weak points and as long as a person can deconstruct R-Walls and isn't dumb enough to sit there and get tased you're shit out of luck as an AI, whether MALF or not.

Doesnt help that theres a camera console in RD office that lets you view ai core to check for blue apcs when you even get a small sniff of something fishy going on with the ai.

The camera console in the HoS office can do the same

SkeletalElite wrote: ↑Tue Aug 16, 2022 8:04 am

Shellton(Mario) wrote: ↑Sun Aug 07, 2022 11:51 pm

Kitfox wrote: ↑Sun Aug 07, 2022 11:33 pm

CPTANT wrote: ↑Sun Aug 07, 2022 9:19 pm AI is piss easy to destroy and its only gotten easier. The only way to survive as a malf AI is to be hidden.

Pretty much. Most AI chambers have very clearly designed weak points and as long as a person can deconstruct R-Walls and isn't dumb enough to sit there and get tased you're shit out of luck as an AI, whether MALF or not.

Doesnt help that theres a camera console in RD office that lets you view ai core to check for blue apcs when you even get a small sniff of something fishy going on with the ai.

The camera console in the HoS office can do the same

I cannot put into words how irrelevant this is in practice.

fairly sure APCs dont show their display when viewed through a camera console but that mightve changed idk

Pandarsenic wrote: ↑Sat Aug 13, 2022 4:59 pm You're totally fine to incidentally make plasma flooding impossible (rerouting all the plasma to a burn chamber, all the plasma and N2O to a Bz chamber, etc.) as long as you're not doing it purely to leave the AI unable to use one of its most important tools.

I understand what the rule is, but I’m not sure I understand why. The atmospheric technician’s job is to make sure the air on the station stays breathable, and taking measures to prevent a single Ctrl-click from leaking an irreparable plasma inferno onto the station, seems to fall squarely within that goal.

If they rig it up so that only the atmospheric technician has control (or understanding!) over the flow of the atmosphere, that makes sense intuitively.

Is the concern just that the malf AI trying to husk everyone won’t be able to do so without adjusting pipes? It just seems odd. By the same logic, secure tech storage and other objective areas shouldn’t be bolted so that traitors can access them easier, and chaplains shouldn’t be allowed to spray holywater around without a clear sign of cult first.

Please just give the AI a roundstart shell so we don't have all this nonsense and the AI can actually do things instead of being completely powerless for the simplest of preparation.

sinfulbliss wrote: ↑Thu Aug 18, 2022 10:43 am By the same logic, secure tech storage and other objective areas shouldn’t be bolted so that traitors can access them easier, and chaplains shouldn’t be allowed to spray holywater around without a clear sign of cult first.

Yes, that is true. (and not that un reasonable)

You have discovered the point

tjatpbnj wrote: ↑Fri Aug 19, 2022 4:31 am

sinfulbliss wrote: ↑Thu Aug 18, 2022 10:43 am By the same logic, secure tech storage and other objective areas shouldn’t be bolted so that traitors can access them easier, and chaplains shouldn’t be allowed to spray holywater around without a clear sign of cult first.

Yes, that is true. (and not that un reasonable)

Difference is there isn't a station job dedicated to making sure the items in secure tech storage don't leak out.

sinfulbliss wrote: ↑Sat Aug 20, 2022 4:02 am

tjatpbnj wrote: ↑Fri Aug 19, 2022 4:31 am

sinfulbliss wrote: ↑Thu Aug 18, 2022 10:43 am By the same logic, secure tech storage and other objective areas shouldn’t be bolted so that traitors can access them easier, and chaplains shouldn’t be allowed to spray holywater around without a clear sign of cult first.

Yes, that is true. (and not that un reasonable)

Difference is there isn't a station job dedicated to making sure the items in secure tech storage don't leak out.

clearly you arent shitseccing hard enough

There's too much powergaming so we should allow even more powergaming seems somewhat self defeating

You are mostly grappling with the outcome of taking a pretty shitty gamemode and just inserting it into another one without much thought instead of trying to fix it.

iamgoofball wrote: ↑Mon Aug 08, 2022 7:06 am i haven't seen a malf AI win in ages, can we pull the stats to verify my claims

I've seen multiple malf wins on Sybil in just the last week.

VexingRaven wrote: ↑Thu Sep 08, 2022 2:52 pmI've seen multiple malf wins on Sybil in just the last week.

You have probably seen more malf wins this week than goof has played rounds with a malf AI, successful or failed, this year

Malf can win using hide and seek or if they get lucky with having enough borgs, stealth isn't particularly interesting though.

CPTANT wrote: ↑Thu Sep 08, 2022 9:23 pm Malf can win using hide and seek or if they get lucky with having enough borgs, stealth isn't particularly interesting though.

Why is stealth not interesting? I think it's an interesting goal conflict trying to balance doing what the crew wants with doing what the AI needs to do. The closer to winning the AI gets the harder it becomes to hide the hacked APCs and also the more tempting it gets to start disobeying the asimov laws.

VexingRaven wrote: ↑Sun Sep 11, 2022 9:21 pm

CPTANT wrote: ↑Thu Sep 08, 2022 9:23 pm Malf can win using hide and seek or if they get lucky with having enough borgs, stealth isn't particularly interesting though.

Why is stealth not interesting? I think it's an interesting goal conflict trying to balance doing what the crew wants with doing what the AI needs to do. The closer to winning the AI gets the harder it becomes to hide the hacked APCs and also the more tempting it gets to start disobeying the asimov laws.

I think they mean literally running off with the core and physically hiding it, like putting it in the little 1x1 empty space in the wall behind the Captain's office, then sealing it behind fully, which nobody - nobody - will find, unless it gets bombed open or an ENORMOUS gamer checks the powernet for suspiciously high draw (or they ghetto powersink the station with a dead body/monkey on a conveyor belt into a shocked grille to starve out the AI)

I have seen one traitor AI reveal itself early to the station, played peaceful, killed off other antagonists for the crew, then go full malf and successfully nuke the station.

Archie700 wrote: ↑Mon Sep 12, 2022 2:51 am I have seen one traitor AI reveal itself early to the station, played peaceful, killed off other antagonists for the crew, then go full malf and successfully nuke the station.

Pretty sure I was in that round, deadchat was salty to the max.

Pandarsenic wrote: ↑Mon Sep 12, 2022 12:48 am

VexingRaven wrote: ↑Sun Sep 11, 2022 9:21 pm

CPTANT wrote: ↑Thu Sep 08, 2022 9:23 pm Malf can win using hide and seek or if they get lucky with having enough borgs, stealth isn't particularly interesting though.

Why is stealth not interesting? I think it's an interesting goal conflict trying to balance doing what the crew wants with doing what the AI needs to do. The closer to winning the AI gets the harder it becomes to hide the hacked APCs and also the more tempting it gets to start disobeying the asimov laws.

I think they mean literally running off with the core and physically hiding it, like putting it in the little 1x1 empty space in the wall behind the Captain's office, then sealing it behind fully, which nobody - nobody - will find, unless it gets bombed open or an ENORMOUS gamer checks the powernet for suspiciously high draw (or they ghetto powersink the station with a dead body/monkey on a conveyor belt into a shocked grille to starve out the AI)

Disk pinpointer targets ai core once it goes Delta and I recall that there is two on station i think. One cap and one hos

We have no intention of walking back restrictions against powergaming by taking action to pre-emptively cut off options for malfunctioning AIs without any in round reason to believe those actions necessary or justified.

Spook: Deny (No.)
Rave: Deny
San: Deny