Waladil

Reposting what I edited into an above post (it took me a lot longer than I thought to do this testing and a bunch of replies popped up while I worked) You guys can watch what happened with super-nerfed aliens. Okay. So clearly changing the XML files had SOME impact. But he ought to have had ONE TU. The first turn he turned around (4 TUs), and then crouched (4 TUs). Second turn he uncrouched then crouched again (8 TUs). So he got at least seven more TUs than he was supposed to. I should point out that this test game was in NORMAL mode, and I should further point out that I noticed in gameconfig.xml that the difficulty modifiers say alienAttribute<difficulty> not alienHealth<difficulty>. However, that doesn't effect this test because it was in normal (1.0 modifier should do nothing). A further test that I did with this mod: I decided to test alien sight range. I discovered that an alien would always face my nearest troop (If two troops were equidistant then the alien would alternate which he faced every turn). So I eliminated all but one alien, put one troop on each side of the last one, and started leapfrogging them out. This test gave me two results: One, the alien always stood up, TURNED 180, and then crouched, so we can up their minimum TUs to 12. Second, they have a sight range of (drumroll please) ...Results inconclusive. The alien remained responsive to my actions until I ran out of map to test on. The final range was 26 tiles, and the alien continued to alternate his facing every turn. He remained responsive to 26 tiles. He always maintained vigilance on the nearest enemy up to 26 tiles. That means we can clearly and definitively declare that the AI is using far more information than it ought to have, whether because they have a silly-long LOS or, my theory, the AI is SIMPLY IGNORING THE LOS SYSTEM ENTIRELY. P.S. The alien also remained responsive to my troops that had things blocking LOS, namely those huge desert pillars. They can still see through walls.

I'll start work on testing aliens with a mod now. Work would go faster if a Certain Someone (Chris, preferably), can answer this question: How often does Xenonauts read data from the XML files? When you load a save? When the game starts up? When you start a new mission? When you start a new game? Until that question is answered, I have to assume that Xenonauts must be quit out of, and a completely new game started with every iteration of test. So I'll go extreme. All Caesan and Sebilians non-combatants (this test wont run long enough to fight anything else) will be reduced to ONE TU, ONE sight range, and ONE HP. (Might as well test if their health is being cheated too and it's just harder to tell.) Pretty much if they ever shoot me then something's going wrong. Ditto if they ever survive a single bullet. EDIT: Well, you guys can watch what happened with super-nerfed aliens. Okay. So clearly changing the XML files had SOME impact. But he ought to have had ONE TU. The first turn he turned around (4 TUs), and then crouched (4 TUs). Second turn he uncrouched then crouched again (8 TUs). So he got at least seven more TUs than he was supposed to. I should point out that this test game was in NORMAL mode, and I should further point out that I noticed in gameconfig.xml that the difficulty modifiers say alienAttribute<difficulty> not alienHealth<difficulty>. However, that doesn't effect this test because it was in normal (1.0 modifier should do nothing). A further test that I did with this mod: I decided to test alien sight range. I discovered that an alien would always face my nearest troop (If two troops were equidistant then the alien would alternate which he faced every turn). So I eliminated all but one alien, put one troop on each side of the last one, and started leapfrogging them out. This test gave me two results: One, the alien always stood up, TURNED 180, and then crouched, so we can up their minimum TUs to 12. Second, they have a sight range of (drumroll please) ...Results inconclusive. The alien remained responsive to my actions until I ran out of map to test on. The final range was 26 tiles, and the alien continued to alternate his facing every turn. He remained responsive to 26 tiles. He always maintained vigilance on the nearest enemy up to 26 tiles. That means we can clearly and definitively declare that the AI is using far more information than it ought to have, whether because they have a silly-long LOS or, my theory, the AI is SIMPLY IGNORING THE LOS SYSTEM ENTIRELY. P.S. The alien also remained responsive to my troops that had things blocking LOS, namely those huge desert pillars. They can still see through walls.

Stellar, did you not see (and since I edited it like 4 times I completely understand you not seeing) my point about "aliens have more TUs than the files tell them to have?" I'm getting more and more convinced that the AI code is completely skipping normal game mechanics. The aliens aren't bothering with things like "TUs" or "sight range." The code tells them to take action against any enemies within 25 tiles, so they do that. The code says they ought to fire a burst when possible, so they do that even if it takes 15 more TUs than they have. I mean, let's say you're right and the LOS code is good (which is pretty obviously is not). That 50-TU Caesan must have fired a burst, then turned around and walked six tiles, then turned back around. That's 40 TUs in burst fire, 8 TUs in turning, and 24 TUs of walking. Grand total: 72 TUs minimum. So we can prove, conclusively, that the AI must have more TUs than the XML files tell them to have. I have done that both here, assuming no LOS bugs and above (when that same Caesan fired a 40-TU burst after being suppressed down to 25 TUs). Why would you not believe that they also have a longer LOS than the xml files say?

I'm gonna start a new game... with FRAPS recording. When (and I expect this to be "when" not "if" I get a good recording of aliens with ludicrous sight ranges I'll post a link on the forums. I'll try to get a save, too, but that's harder. EDIT: Yep. First mission, alien took a shot at me from 20+ tiles. Using... burst fire from a pistol? Anyway, here's the video I promised. You can see me at the end count tiles from the Caesan to my troops. He missed badly and I dunno if I counted tiles to the right troop but the nearest troops were about equidistant from him... and I counted OVER 23 tiles. If you watch the end you can count with my cursor. That video should also pretty clearly prove that it can't be based on squadsight, since I've got the whole area clear. Save is no good, since it's from the beginning of the level not that turn. EDIT 2: I should point out that the game's balance is currently way far out of whack. I've been fighting that ONE CAESAN for a couple turns now, and my troops cant hit worth a damn. Neither can he, though. Also, he got suppressed by my first rifle bullet that went past him. Also also, he's clearly not using the proper AP system. He's a Caesan non-combatant, aiprops.xml says he ought to have 50 APs. Well, I have ended every turn with him suppressed, and can usually return fire with a burst next turn. The plasma pistol's new burst fire costs 40 AP, so he's clearly starting each turn with at least 80 AP. (Unless the suppression system got changed and now you don't automatically lose 1/2 of your AP. I've not been suppressed yet to test that.) EDIT 3: Suppressed my own troops with flashbangs. They lost 1/2 their AP next turn from suppression all right. Suppression damage also seems borked: Like I mentioned above, I suppressed a Caesan with a single rifle round but a few turns later a flashbang at his feet did not suppress (even though the weapons_gc file indicates that a rifle bullet has a suppression value of 40 while a flashbang has a suppression value of 100). After I eventually killed him, I flashbanged my own troops to test. It took two flashbangs to suppress two of the five troops I had standing right next to each other. EDIT 4: Finished the first mission. Discovered a few new things. 1.) The aliens like to hide more, now. I was playing "find the last guy" at the end of that mission, and eventually found him in a building. Right after I found him I also learned that 2.) The "Secure UFO for 5 turns" objective is active now! So we can win without playing "find the last guy" if we don't want to. (Downside: You do lose 2 points per alien not killed, for a net loss of 4 points (if you had killed them) or 5 (if you had captured them). Dropping UFO secured value to 5 makes getting points from other things much more important!

There is a bug. I suspect (and this is just a layman's suspicion, NOT an informed statement of truth) that somewhere in the new AI code it has hardcoded sight ranges of 20 (or thereabouts), which is considerably longer than it is intended to be. The new AI code might also skip a few steps on testing whether or not the lines of sight can actually be drawn, hence the "firing from inside buildings" bug.

Mostly the difference between them having infinite ammo and them having finite ammo would be having to stop and reload. If they had finite ammo and a couple more cells, it would mostly mean them having to stop shooting (or just shoot less) perhaps one turn out of four -- not really a big deal.

...However, it doesn't work. Try setting the sight ranges to 1. See how that goes. (Hint: Nothing changes.)

Because, Lightzy, it's too easy. Yes, it's no fuss no muss -- but this game has definitely attracted a certain kind of clientele. Lots of people want a hard, unforgiving game. They want mistakes to be punished. Of course, I suspect that many of them don't actually want that, they just think they do. (Technically, I think they want to succeed at a punishing game -- whether they're actually good enough to do so is another question entirely.) But, my internet armchair psychology aside, what they are asking for is hard decisions and punishing results for making bad calls. My suggestion a couple posts above yours provides for potentially punishing results, but something softer than it is now while adding extra tactics and depth. So it wont punish you too much unless you REALLY botch -- like sending one ship alone and outgunned. (Also, your fighters aren't faster than a lot of enemy ships -- go toe to toe with an enemy Interceptor and try to escape.)

I've seen a similar version of this, but with completing a ground combat mission that I had to reload (either to savescum or just because I was going to come back to it later).

Fun-fact: Your LOS angles are not, actually 360 degrees at night. You can see _terrain_ in 360 degree angles, but you cannot see aliens in those same angles. Very deceptive, actually. (I've turned around and BOOM. There was a Reaper standing right behind my troop, in a spot I thought was clear because of the light.)

I think it's SUPPOSED to be "You've got troops inside the alien ship for 5 turns with NO ALIENS in it." To prevent the player from having to hunt down one mook hidden between some boxes on the third floor of an otherwise-empty warehouse. So you don't have to worry about accidentally winning mid-clear, nutbarz. Also there is an "aliens escaped" score penalty that currently can't actually be a penalty right now for just that scenario, of capturing an enemy ship without killing all the aliens.

Reflecting on this thread, and people's various realism v. gameplay and difficulty v. forgiveness arguments, I'd like to recommend the following idea. It's not fully finished, it's not fully realized, but it's an idea: For ALL aerial vehicles (human and alien, dropship and interceptor), have them remain on the air combat map for a short time (2-5 seconds?) after being shot down. During this time they can be targetted and fired upon, and all missiles will continue to track and hit them. As their health plummets into the negatives, things are changed on the geoscape. For the human vessels, they start taking more and more damage, before ultimately being completely destroyed. So there's, say, a 5 day extra repair time if they're shot down (representing salvage efforts and the extra repairs from impact), and up to 5 MORE days if the aliens took it down to the very edge of death. At the worst, that interceptor could be out of the game for about two weeks. (5 days for being shot down, 5 days for post-shootdown repairs, 4 days for regular repairs.) Dropships that start taking extra damage will be more and more likely to have deaths among the troops: If they're just tagged and mostly control the crash, most of your soldiers will live with minor injuries. If the dropship got hit hard, well the frame is salvageable, but the troops... no-one survived, sir. Alien vessels also have this overdamage math. No longer is it random whether an alien vessel took "severe" or "minor" damage. It's now calculated based on how much overdamage it takes. Moreover, I'd recommend adding a "destroyed" level where the crash is unrecoverable (no crash assault mission at all). So a smart player can calibrate missions to his/her own readiness: If they need the money and have the capable troops, they want to just barely damage enemy birds and get the most loot from a minor-damage crash. Or they can hit it hard and have an easier run on the ground, good for training the noobies. Or, if they don't consider it "worth their time" (for whatever reason), they can blow that sumbitch to smithereens. Note on the AI: The computer AI would target nearby active human interceptors before trying to overdamage the ones crashing. So covering your weak birds would help them be salvageable, even if you have to spend extra time repairing them. The auto-controls on the human interceptors would automatically switch to the next active enemy, and wouldn't try to overdamage enemy ships at all, you'd have to manually order that. (Possibly add an extra UI button/keyboard command to continue firing on selected target past initial shootdown.) Final note: For the "destroyed" enemy birds, including fighters, add a small amount of alloys (+alenium?) as long as they're taken down over land, from scavenging the debris. There'd be bound to be something useful there, even if it's not much. (Like, 2 alien alloys per fighter. Maybe even just 1.) So this would add another level of realism, where planes can go down without being completely destroyed but might actually be destroyed, AND add a level of tactics to the human side, both offensively and defensively. Seems to me like it might be able to satisfy all sides? Feel free to critique it, y'all. (Biggest problem I see: More work for you and the rest of the coders, Chris.)

Okay but that doesn't change the fact that a unit behind the smoke (wherever it may be) is damn near undetectable. So make it so smoke locations change slightly from round to round. Also, envision the effects of that in an enclosed space... say the interior of a scout. The entire room would fill, and nobody would be able to see in or out at all.

http://www.youtube.com/watch?v=d227zKrC0RA Watch that video. Aside from the unprofessionality of it (I just found it really quick) there were a few times the smoke went over a person... and he completely disappeared. That smoke is thick enough that movement behind it would be totally concealed... I'd argue, having watching that, that smoke instead of reducing accuracy or reaction fire chance, ought to (meaning to match lore rather than as a game mechanic) act as blocked lines of sight. So no reaction shots because no LOS at all. That seems like it'd be bad gameplay, but smoke that thick would be impossible to have even reasonable reaction fire through. If you want to argue that aliens would fire through the smoke, then you can make that argument but it wouldn't be reaction fire. That's fire that should be done on the alien's turn. (And no, it's not "literally tons of smoke." That was one grenade, while a full smokescreen would probably be two or three.)

Hopefully the new setup holds steady! I wasn't able to access the forums a little bit ago, but they seem faster now. Not sure if this is just one of Dreamhost's better moments or if I'm already on the new forum. Either way, hopefully it stays this quick!

You know, I lost a mission and DIDN'T get a CTD. Things about this mission that may help in tracking the bug: It was a terror mission, not UFO assault. I quit out and reloaded a couple times from an Ironman save. Has anyone else had/not had mission-lost crashes with one or both of those conditions?

Also, there's definitely something fishy with AI sight lines. First off, I was taking fire from Androns from at least 3 tiles out of my line-of-sight. I was using basic armor (LOS 16) and Androns supposedly have LOS 16. They fired on a newly-taken position (so the AI definitely didn't just shoot at where I was last turn) and there were no other aliens around for him to squadsight off of. In the specific mission I'm referring to, I actually counted it at a lost cause in order to test a couple things: One, I scouted around and made sure there were no other aliens for that Andron to squadsight from (so I'm certain that Androns operate at AT LEAST 19 LOS), and I, just for shits and giggles, turned down both the Andron LOS and the Light Drone LOS to 1 in aiprops.xml. (This was a mission with just Androns and Light Drones.) Naturally, this changed absolutely nothing, so the sight values must be defined elsewhere. Possibly hardcoded into the AI files?

You didn't mention that flashbangs were fixed! Or maybe it was a few versions ago, I haven't played much since 18.3. Anyway, flashbangs are effective at long range now. Maybe even need to be nerfed a little now that they suppress at long ranges.

It's Sebillian. Say it with me. Se-Bill-Ian. One more time: Sebillian. Why? Because we don't call them "Graymen," "Birdmen," "Robomen," or "HOLYSHITTELEPORTINGTHINGman." (Although I have a marked tendency to call them "Sebilisians" being as EverQuest brought us a city called Sebilis with lizard people living in it who were called Sebilisians.)

Well Solver did just recommend that it would "reduce" the chance of reaction fire, not prevent it. That seems very plausible from both a lore and gameplay standpoint. It would make your troop movements harder to see and therefore take shots at. Maybe it should, instead, have a doubled accuracy debuff to reaction shots? So if smoke usually gives you -20 accuracy (I don't know the real number), then it would give -40 to reaction shots? Personally I'd prefer a reduction in chance rather than that, but it's another idea. I find it hard to believe that smoke grenades would ever be too OP in favor of the Xenonauts. Remember that the smoke works both ways, so your troops will reaction fire less and be less accurate too. Mostly because you'd have to chuck 2-3 to get a good smokescreen up and you can only bring X number of grenades. At that point just bring stunners and capture the aliens.

Grenades should be able to be thrown over pretty much any cover. Or at the very least there should be a clear indicator of what would stop a grenade or not. I haven't tried it yet, and don't want to test, but do your own soldiers still count as "full cover?" I hate hitting my own guys in the back of the head with a grenade.

You know, I've always wondered why the military wouldn't have some sort of "grenade hook" as part of their combat uniform. Something firmly attached to the armor that one could easily slip a grenade's ring over and pull to prime one-handed. (Added benefit: Holds your car keys between battles.) TBH, the 10 extra TUs to throw a grenade with a two-handed weapon probably represents the slinging and unslinging of your weapon. For balance reasons that should probably be toned down, possibly to 5 TUs? It should be less than the whole cost of dropping/picking up a weapon (but it might be interesting if it was set up so completely dropping your gun for a turn would save TUs if you wanted to throw like three grenades at once).

Actually it implied that the high failure cost of Mega Man makes it appealing to hardcore gamers and unappealing to casuals. Which is probably quite accurate.

Gaud, I'm pretty sure that the file count isn't at issue here. I just flicked through my Steam directory and checked file counts. The most I could find was a little over 15,000 files from Elemental: Fallen Enchantress. I saw a report elsewhere on this forum that Civ V has about the same number of files as Xenonauts, but I don't have that installed. Xenonauts is at 37,000 files right now. So its about 2.5 times more file-heavy than E:FE, but here's the thing: I've never had a single problem with Steam hanging on game launch (excepting first-time install) like Desura gives me, and other people every time. So unless there's some break-point where Desura and Steam algorithms just break down and slow to a crawl, then there's definitely a difference between Desura's speed and Steam's. Because (Steam.FallenEnchantressStartTime x2.5 != Desura.XenonautsStartTime)

Difficulty modes are actually all the same right now, so you can play on Insane and feel like a badass The reason it didn't work probably is that there weren't any ships that can do base attacks... I couldn't figure out how to make certain ships appear sooner, like Carriers. So the game would generate a base attack but couldn't find a ship to actually do the attack.