Messages

16

Released Mods / Re: [SOURCEMOD] Brutal-OXCE 7.7.4

« on: November 18, 2023, 01:51:04 am »

@Xilmi,

I think the essential feature of Brutal AI mod that would make a gameplay interesting is the ability to configure the ability to play by stealth, while at the same time allowing the enemies to blaster you early.

So far, the observed behavior on the latest versions, when playing custom XCF-based mod, has been very plausible.  With aiTargetMode: 3, aiAggression: 2, autoAggression: 3, cheatOnMovement: false, the stealth play is satisfactory, while the enemy does not seem to be dumb (e.g. the enemy does not attempt a zergling rush, and does not refrain from trying to actually do damage every turn, relentlessly).  It is still possible to take Durathread Factory or a Cult Base with only two agents this way, albeit such encounter is a challenge in general.  The enemy is also actually using launchers in that mode, which makes a huge, huge difference when encountering early legates or mummies in XCF.

So far, the only downside is relatively slow performance, especially graphical one, when soldiers are on the upper levels.  I don't understand why that is the case: the openxcom level of graphics should pose no challenge for any hardware accelerated setup, even if coded straightforwardly and inefficiently.

I also noticed that the engine is not taking advantage of multi-core setups.

In terms of features, I think the most pertinent ones at this point would be the ones that allow certain types of explosion (e.g. "spectral smoke", which might as well be a regular smoke with a special flag set) to propagate through the walls.

I wonder, what set of config flags you find to be most optimal when playing your sessions?

17

OXCE Support / Blast radius calculations in OXCE

« on: November 18, 2023, 01:34:45 am »

I would like to start using dynamic radius more widely in a mod that I am working on.  I would like to inquire, how is the dynamic blast radius calculated as a function of damage power?  Would it be possible to post a table of this somewhere on the wiki?  The table is going to be much more useful than a direct formula.

The forum search reveals little w.r.t. this issue.

18

The X-Com Files / Re: The X-Com Files - 3.0: Beyond Human

« on: November 18, 2023, 01:32:39 am »

Oh, great! This mission was a hell earlier, considering how early it spawns, and now with unspecified amount of reinforcements...  :'(The only mission where Dagon's Staff was useful...

Dagon's Staff is pretty good on underwater missions as well.  You could even take a city with it (if your agents are high in psiStrength).  This obviates the need to research sonic weapons by that point (which may or may not be possible by then).

19

The X-Com Files / Re: [submod]Submod list for XCF

« on: October 23, 2023, 08:24:54 am »

No, 'cheatTurn' is a ruleset variable.


I think someone has been streaming XCF+BAI for a while now (and giving Xilmi feedback based on that), plus the other reports that occasionally surface here. I'm not sure it's fully 'supported', nor will ever be, but it seems to be compatible.


I imagine if you come up with some concrete proposals for the artillery commendations and, better yet, some artwork for them, Solarius will consider them. If they're not doable under the current commendation framework, you can always give it a shot in the OXCE subforum, too.

Ok.  All in all, it seems that the best approach would be to run BAI instead of OXCE for my long XCF sessions.

I have some developments with the ordnance variants.  If there's a way to post them on e.g. Github, or a similar public development repository, I'll consider contributing my rulesets and artwork.

Concerning the additional artillery commendations, I think that most relevant commendations should probably fall into the realm of total carnage and large weapons master.  Such commendations should be earnable by both artillery characters and also by BFG characters.

Perhaps, someone causing massive carnage while wielding a weapon that has explosive area effect and weighs more than e.g. 22 or 24, would be a source of commendation.

20

The X-Com Files / Re: The X-Com Files - 3.1: Lights in The Sky

« on: October 23, 2023, 08:19:44 am »

My point exactly. I'm a filthy savescummer, of course , but I occasionally do no-save missions as well. And I can't see how aggressive movement could be done without casualties, unless the RNG gods favour you, the AI grabs the idiot ball with both hands, or you're just clowning on some scorpions or cultists who can't hurt you through all that armour.

A fair approach would be to setup autosave to a value between 1 and 10, and to rollback strictly to autosaved points.  Essentially, this reflects the play performed by an oracle of various strengths.  By taking this approach, a player may focus on tactics proper, the strategy of decision-making, instead of minutiae.

21

Released Mods / Re: [SOURCEMOD] Brutal-AI 2.1.1

« on: October 22, 2023, 10:28:42 am »

I'm no longer actively pursuing OXCE-integration. If they want to have it, they can feel free to include it though. And that goes for everyone who wants it in their fork.

...

@Xilmi, What is your perspective on the mod compatibility with XCF?  Would it be reasonable to expect the Brutal-AI's behavior -- aside that from inside the tactical missions -- differing from the mainline OXCE when XCF is loaded as the master mod?

22

The X-Com Files / Re: [submod]Submod list for XCF

« on: October 21, 2023, 02:30:34 pm »

'Turn 20' is when aliens spot your troops via mind bullets and come out of their UFO to kick ass be killed like lemmings. The setting that controls that is 'cheatTurn'.

Thank you.  Is this setting available through the menu?

I don't think it's possible to make vanilla AI appreciably more aggressive. You could increase the 'aggression' and 'intelligence' values for individual units, I guess. I haven't noticed that making too much of a difference, though.

Well, melee enemies can actually be made dangerous by changing them to leeroys, like Zombies already are. Personally, I quite like leeroying cryptids or Chryssalids, because now they're dangerous and the mission is much less likely to turn into a bughunt. But that's only a small subset of all enemies. There's even a submod of mine for that, though it's not very up to date.


Big maps and showing more layers or more terrain have always been a performance headache, BAI or no BAI. XCF changing the vision range to 40 doesn't help there, either. BAI bugs should go into the relevant thread, though, since only Xilmi can resolve them properly.

Thank you for sharing this insight.

I wonder, if BAI is meant to be compatible with mainline OXCE such that it could run XCF in an essentially compatible way?

I think that the visibility setting 3 (that is, one short of omniscience) is a good spot for the AI to go.  The play is brutal enough to be fun, and yet could realistically be beaten by the player without casualties at all times, with good tactics.

It's always fun to take advantage of the area effect ordnance.  I wish there were more commendations for good artillery work. :evil:

23

The X-Com Files / Re: The X-Com Files - 3.1: Lights in The Sky

« on: October 21, 2023, 02:24:45 pm »

What do these 'fragmentation' shells of yours do, cutting damage instead of concussive?

Correct.  The charge is equivalent to a rocket of corresponding type, minus the incendiary effect.

Trying to peek so that you don't see the turret, only a tile nearby, is also dependent on the map generator collaborating with you, since one step near a corner corresponds to quite an angle in visibility.

This approach works very consistently in practice.  The only issue with terrain generation is when the transport is too close to the UFO.  Still, it's not impossible to win even in such cases, even against deployments that include Chryssalids.

Ultimately, the Skymarshall also still suffers from the 'open ramp' issue, even if the tail blocks most of it. A few Sectopods or Cyberdisks staring down on your equipment pile on turn 1 is a big problem there.

I actually moved the equipment into the middle of the vessel by editing a file in the mod.

Ideally, the backside would be doored as well, but in practice the situation is very manageable, especially when a couple of heavy troopers with rockets are near the exist (so, they only need to make a step, and still have time to aim).

Also, this probably only works due to the overwhelming power of shrapnel rockets, since otherwise you'd be getting serious degradation in damage done and thus armour penetration. And these effin' turrets are quite heavily armoured.

The effect is a consequence of shield resistance.  The resistance against cutting damage is uniformly at 125%, same as plasma, across all shield types.  Also, cutting damage works nicely against enemy armor resistances.  This seems pretty consistent with the real world, apparently.

The ship you're talking about is probably the Troop Ship. A Cruiser looks somewhat like a dildo and has two small and two large turrets.

Indeed!  Thank you for clarification on this one.  There is another vessel as well, a large oblong ship, called Arbiter, with many turrets, but with the same design feature: walls near turrets.

Isn't that what a Mind Missile is?

No.  The difference between 3 and 9 or unlimited waypoints is significant in practice.  The idea is to have a slightly correctable missile, like the one in the CoD, etc..  The rationale being that such missiles are still reasonably inexpensive to be available as a standard offering, if only for spec ops troops.

Anyway, top-attack munitions weren't as widespread in 1997-1999, though the Javelin did go into service in 1996 and the Bofors already in 1988.

The basic idea I wanted to illuminate when referencing the anti-tank missiles has been that the technology for missiles that could be steered in flight is already available.  The 3-waypoint missile simulates such approach nicely.

Now, given that the waypoints could be relatively arbitrary in this game, we should think of such in-game missiles as an extension of the current AT missiles technology, into a special product geared to elite special operation units.

So, the 3-waypoint rockets are a bit of acrobats, but they don't even remotely approach the magical effects of a genuine blaster bomb.  Especially, when you set the waypoints for the latter to infinity.

Ultimately, the 3-waypoint rockets are an optimization, and could be replaced by mortars and regular rockets, whether with frag shells or without.

Although I can't see how you'd fire a Javelin from a Dragon launcher or similar. It's a new weapon system, not a munition upgrade.

The in-game "Rocket Launcher" is a distinct product most definitely.  Still, the capabilities at the level of 3-waypoint guidance are more in the human-tech realm than even xcom-tech.

In my experience, grenades and Blasters (and cultist explosives) are the number one cause of deaths I could not have prevented by playing a bit more cautiously. YMMV.

It's possible to have a zero casualty playthrough, with very conservative reloading policy.  I tend to play very aggressively in the first turns, in order to create a suitable beachead.  It works much better than defense.

All in all, it's possible to overcome tactically the challenges of grenades and blasters.  The key here is access to ordnance.  Also, I would like to note that auto-mortars feel almost magical in their demolitions capacity.  A trooper could carry 7 rounds of mortar charges.  A fire team consists of 4 troopers.  That is usually more than enough to demolish enemy.

Rocketeers are there to provide cover for mortars.  They tend to be idle once the mortars start working (and with auto-mortars, it's possible to fire every turn).

Prior to promotion 3, some elements of such tactics could be simulated by the use of dynamite and HE packs (sourced as trophies).

Perhaps. Rockets and mortars have the distinct disadvantage of firing very slowly, not working at close range unless you have serious power armour, and needing an ammo supply line in any extended engagement. I suppose with good fire discipline and the OP shrapnel shells, it's not as much of a problem.

With mortars, it is necessary to setup a good spot to fire from.  In general, you do have four shots of rockets to help with that even when heavily surrounded by cyberdiscs.  The slightly maneuverable rocketry helps to bring casualty potential to nearly zero in such circumstances.

24

The X-Com Files / Re: The X-Com Files - 3.1: Lights in The Sky

« on: October 21, 2023, 12:24:19 pm »

The Skymarshall indeed provides better cover for indirect fires, no doubt.

How is it better for direct-fire, though, like for the aforementioned shrapnel rockets? If you have LoS, so does the turret. I mean, you could perhaps hide a bit of your agent's body at the time of firing, but that would be highly dependent on the turret and the craft aligning perfectly to give you that possibility.

The Lightning is nice because it can either provide high ground when you don't have to worry about return fire (smoke, non-static enemies, sniping and the like) or cover from arcing weapons when shooting from below the saucer, which even the Skymarshall doesn't really have an equivalent to.

If you double down on mortars and grenade launchers, you indeed get a unique advantage out of the Skymarshall. Pop-up fire seems better from a Lightning, though, especially as you don't have the doorway jams when the enemy is coming from several directions. That is, I can see the superiority of the Skymarshall against static enemies that always react and can be safely indirect-fired to death; but I don't see it if you are not exclusively facing those and using that tactic.

And you're somewhat at the mercy of the map generator, since if your craft is facing the UFO or other enemy concentration nose- or tail-first, the advantage is gone. Probably pretty rare, though.

With Skymarshal, it is possible to setup two teams of 6 rocketeers at each edge and fend off any assaults.

I also have made mods to mortar rounds set to enable fragmentation rounds (which actually corresponds to a real military practice across the world) and also have experimented with setting up 3 waypoints on the rockets.

Note that those are essentially optimizations.  With mortars, I could just expend 1-2 extra HE rounds and still destroy the turret.  With strictly direct fire rockets, the trick is to explore the geometry both of Skymarshall and of a UFO.  What needs to be done is to peek just a little bit from the cover, then target a part of structure near enough the turret.  The terror ship and cruiser (the one with lots of terrorists and four turrets; do you recall if it is indeed called a cruiser, or has a different name?) have such properties: their turrets are not set at the flat open surface, but rather near walls.

By the way, if you edit-in the frag charge rockets, you could add them as special x-com rockets with limited maneuverability.  That corresponds to current portable anti-tank weapon capabilities across all advanced militaries in the world, so it's a very reasonable addition.  I would even argue that such rockets should be sellable, and maybe even shootable from a regular (= not advanced) rocket launcher.

I appreciate you sharing your observations about arcing shots.  This is a nice property, albeit none of the mid-game enemies are a serious threat in that category.  Still, the enemies are pretty good at throwing grenades, so there's still extra advantage on the side of Skymarshall, with its narrow doors design.  In general, it suffices to keep a certain area around the craft clear to prevent any serious dangers, and that is very doable with rockets.

25

The X-Com Files / Re: [submod]Submod list for XCF

« on: October 21, 2023, 12:11:17 pm »

Yeah, performance is always a trade-off in making a good AI.
I suppose you could change 'turn 20' to 'turn 1', or give all enemies unlimited psi-vision.
Don't think so. XCF enemies already blaster and grenade you as fast as they can.

Thank you for sharing these insights.

What do you specifically refer to when talking about changing "turn 20" to "turn 1"?

Speaking more generally, how could I configure the mainline OXCE to have as aggressive XCF enemies as possible?  Ideally, that should not involve making any changes to the XCF itself.

I think, there's a bug in Brutal AI, since I had been having severe slowdowns on upper layers, as compared to lower ones.

26

The X-Com Files / Re: The X-Com Files - 3.1: Lights in The Sky

« on: October 21, 2023, 08:38:40 am »

XCF Lightning doesn't have a ramp, but rather "elevators". You can easily test the craft designs with a "new battle". The design also prevents enemy grenade launcher reaction fire (if you go down instead of up), but obviously not straight fire. I could very well see it used with a tactic, go down, shoot a snapshot, come back up to safety. You'll still have to deal with enemy reaction fire but you could just take it or drop a smoke. This design would be particularly useful in missions which are so full of enemies that you might have trouble dealing with all of them close by no matter what you do. You could also deploy units on elevator squares so that chryssalids or any melee-only units couldn't get at you.

With kitsune you can also jump down, take a snapshot and run back up. However, due to the ramp the craft is more exposed, and the enemies can come and throw a grenade in or shoot with a rocket launcher. So it's not full safety.

The same applies to Osprey and all the sky* craft. It's open-ended and at a certain angle enemies can shoot inside or throw a grenade inside. Trying to stay in the cover is especially risky if the landing area is in the bottom parts of the map. Obviously if the landing area is in the top, it is much more difficult for enemies to get to the angle that they could hurt you.

The benefit of skymarshall is that it is on ground level. So you can get back in to cover with far fewer TUs and with osprey and other sky*. I'd suppose this is particularly useful with the front exits because the doors there will actually protect you from attacks during enemy turn (while the open-ended main entrance does not). But if you end up in the top end of the map, "run back inside" tactic will work from the main exit as well.

I have never bothered with other than early Osprey/Skyraider and Kitsune. But I suppose I could see a use case for both Lightning and to some degree Skymarshall.

The problem with the hatch leading to the top is that even if a trooper shoots ordnance (rocket, grenade, or a mortar round) from the roof, there's a high risk of reaction fire.  In case of turreted crafts, the reaction fire might as well come from the turrets themselves.  The hit-and-run tactics opportunities are thus limited even with such designs.  The Lightning works out tactically as an improved Mudranger of sorts.

With the Skymarshall craft, a different approach is possible, where a portion of the craft could serve as a cover.  A cheap option versus turrets might as well be TT explosives, or mortars.  Hit-and-run at a distance, from very difficult to reach angles becomes a possibility.

I managed early Arbiter takeovers when employing such tactics consistently.  Most of my troops used shrapnel rockets, and a handful was using small launchers.  Result: a lot of dead sectopods and about half of ethereals (it's mid-1999!) were captured alive, including commander.  I highly doubt that such a result would've been even remotely possible had I used any other craft, including Avenger.

Yes, with blaster bombs (and my modded-in blaster stun bombs), the aforementioned fight would've been really easy, even without psi.  However, without these capabilities, the Skymarshall seems to be the only craft design that is able to provide such a tremendous tactical advantage.

That is the reason I particularly appreciate the Skymarshall (which I referred to by as "alloy skyranger" in the recent posts to this thread).

27

The X-Com Files / Re: The X-Com Files - 3.1: Lights in The Sky

« on: October 21, 2023, 08:29:07 am »

This is the OG Lightning, which is indeed not particularly better than the Kitsune or similar. I think XCF uses some version of the "Tactical Lightning", which looks like this.

Thank you for sharing this.

This is indeed a rather different craft than the one displayed in the video.  It is more similar tactically to the Mudranger and Dropship.  However, an important caveat goes with crafts that expose troopers on the roof: they are much more vulnerable to turret fire than the troopers shooting from behind the cover of the craft itself from the ground level, especially if the latter sort of troops shoots an arcing or a guided projectile.

28

The X-Com Files / Re: The X-Com Files - 3.0: Beyond Human

« on: October 21, 2023, 04:38:43 am »

I never tire of shilling for the Lightning, which is essentially a giant turret that flies around the globe.

I never used Lightning craft in a playthrough.  I have referenced this

video to check it out.  I see that it is hardly different from Kitsune, tactically speaking.  How do you use it on difficult missions?

29

The X-Com Files / Re: [submod]Submod list for XCF

« on: October 21, 2023, 04:33:05 am »

The only real incompatibility so far has been the multi-craft hangar feature. Easy to fix manually, and has been WIP in the main fork for quite a while now, with several people offering their solutions. Maybe it'll become a feature sometime in the not too distant future, who knows?

A whole bunch of Brutal AI changes have actually been rejected by OXCE devs.

And, in any case, Brutal AI is a pretty radical departure from vanilla in a direction most of the big mods were never intended to deal with. So it's naturally going to have issues, but from another POV many of these issues are precisely the point of having a 'brutal' AI in the first place.

Thank you for sharing this.  I tried Brutal AI build earlier this year, but it turns out that there have been some weird performance issues in certain places occassionally.

I wonder, if it's possible to configure the mainline version of OXCE to have similar properties to Brutal AI?  Maybe, something like omniscience?  While the omniscience on the side of AI may balance strongly in favor of AI, it would provide a good basic challenge.

Also, is it possible to configure the mainline OXCE to have the aliens use rockets and grenades much more actively and proactively?  That would introduce a tough landing zone cleanup, but it is also a fact that tactics (rocketeers shooting from the alloy skyranger, and the inventory stash moved to the middle) do exist to overcome even such challenges.

Natasha Morozova Join... mod does have some rather strange features, but these are about as incompatible with XCF itself as they are with BAI, i.e. none of this works particularly well without constant tweaking to accommodate for the changes to the parent mod/AI fork.

I tried to fix the mod at one point, but gave up, since overhauling logic to streamline it would've meant essentially an mod rewrite.  I think that multiplying entities, especially soldier types, in an incompatible manner, and also incompatible with future upgrades in XCF (consider e.g. the new cybernetic upgrades), is an unsustainable approach.

30

The X-Com Files / Re: The X-Com Files - 3.0: Beyond Human

« on: October 20, 2023, 08:07:57 pm »

We all play this game kind of differently. For me, i like to sneak my guys out on Terror missions. Set up strategic placement vs. hiding in the craft or coming out for one shots.
This of course is totally dependent on the terrain but it's funny how we all play this game so differently. We ask for advice, we take some parts of that advice and incorporate it into our own game plan.

For example, I only build one base until after all cults are finished. then only one more in the future.
Some people like to create 4 or 5 bases. I think it's all about what we like and what has been successful for us in our previous missions, game plays.

I noticed that when playing on superhuman, with aggressive enough of an AI (whether Brutal AI or not), the enemy attempts to storm the incoming vessel upon arrival.  This does get challenging with Chryssalids and dropship being placed too close to the UFO.  Still, even such missions are doable with a clever setup of mines.
 Note that on terror missions a relatively universal solution exists -- use shrapnel rockets extensively.  That does work even against early muton terror ships (actually, that might have been a cruiser; I refer to a nasty ship with a lot of terror units and 4 laser turrets).

The bottom line is that the current configuration of the game is very much playable even on superhuman difficulty level.  I did a challenge to myself to never skip a mission, except for a small finite number.  It turned out that this challenge is really doable in practice.  Early assaults against larger vessels are quite viable.  The key is to survive initial onslaught.  The solution is to use ordnance very extensively.  In fact, you might as well arm everyone with rockets and mortars in order to survive first 6-10 turns.  After which, rifles and cannons could be equipped.

A curious fact: I never found the robotic units useful prior to Enforcer level.