OpenXcom Forum
OpenXcom => Open Feedback => Topic started by: michal on August 17, 2012, 08:09:27 am
-
Hello,
0.4 is released. To keep our programmers motivated, lets write how many of us want 0.5 ;)
0.5 plans:
Geoscape AI (missions, UFO spawning, bases, scoring, funding, interceptions, etc.)
Graphs
Victory/Defeat
Battlescape:
More alien units
Alien base and X-Com base missions
AI Phase 2 (grenades, blaster launcher, taking cover)
So basically, 0.5 should be first playable (from beggining to end) version. It will be important step in OpenXcom development.
-
complete geoscape AI? wow, that's really an important step :)
needless to say, that's also a very interesting field for modding... because it changes the fundamental mechanics of the storyline, and can lead to completely different spin-offs...
go, go, go!
-
Yeah, most of the essential stuff is already done, 0.5 will be a complete game (component-wise).
I bet that by the time 0.5 will be finished, most bugs will be already fixed.
-
There is no geoscape AI.
There is a probability table of missions/ship_types/races appearence for each geo zone, hardcoded by months. Plus huge geolocation chart (for each zone and each trajectory stage), plus alien ship crew variations/missions, plus alien race/rank stats/inventory list.
And I don't know how SupSuper will make it anyhow moddable or extendable.
Though it will be relatively easy to replicate it without any chance of modding.
-
I know that there is no true AI, but there is a more-or-less logical sequence of stuff happening: small ships appearing before large ships do, lots of intercepted UFOs leading to alien retaliation missions, alien bases being "constructed" and "supplied"...
a lot of stuff that pretends to be an AI acting and reacting :)
I guess no one would be angry if the openxcom geoscape "AI" would end up a bit "smarter" than the original, right?
-
Smarter? Of course. But before making it "smarter" it should walk through "the same" area, which is I believe quite difficult to make if "moddability" is a "must".
-
Smarter? Of course. But before making it "smarter" it should walk through "the same" area, which is I believe quite difficult to make if "moddability" is a "must".
Making it the same could also mean either throwing that code to a bin later or limiting yourself. I think that the original approach is completly wrong. Why can't we just have a "real AI" which generates ufos really smart - you can easily count how the player is doing, take into account difficulty and the stage of the game (based on game time, research done, ...) and deliver ufos and missions adequately.
eg. in the beginning (on lower difficulties) make the ufos fly near the players base, don't hardcode low mission count for lower difficulties (it might actually be harder, because you have nothing to savage, therefore resulting in no money), make the retaliation take place where the previous mission was unsuccessfull, do terror missions where xcom has low activity (so it's like a "surprise" attack), the same with alien bases (so they can't be detected so easily), etc. The aliens have so much technology, but are stupid enough to build own base near xcom base? That's how the original works :)
-
AI is just a bunch of rules you make, so agents appear to have intelligence, because they make decisions based on these rules, their state and the state of the world around them.
I don't see what you mean with "real AI". There is no such thing: real <> artificial. Even modern games uses tricks to make people think the AI is more advanced, but it just uses the same old "tricks".
I'm sure the original approach can be tweaked a bit to fix those issues you mention, but it doesn't mean everything has to be thrown in the bin.
If the geoscape AI in the original game didn't do anything right, why did you ever play it? Or didn't you like it at all?
-
...
That's why I wrote it in quotes. Basically I just suggest to throw out the "fixed missions" that are chosen randomly and make completly random and fitting missions instead.
-
Amunak, you must be wishing this project unsuccess? If game won't make any challenge and only react to player's actions, it may stuck on low player level and at zero progress. There always must be "hurrying up" events which can make player fail.. otherwise it won't be any interesting. I will hate this game if it will be a cakewalk to beat aliens, in this case it will simply become piece of crap, you know. It's not casual game for naive time-spending, it must present real challenge, and I want to believe - much more challenging than original.
-
Oh no, misunderstood again. You are right of course, but my idea doesn't exclude challenges and hard gameplay. That's why I'd like to see better geoscape AI that's dependant on the difficulty more than in the original. On harder difficulties you can make aliens smarter, maybe even so that they try to avoid the player, or that they do unexpected things (like draw the player from a base to get very large UFO and when he leaves attack the base). However, on lower difficulties, you have less missions in the original. That sometimes leads to very hard gameplay because you need money from savaged ufos, and since they sometimes completly avoid your base and land on the other side of Earth, I suggest making beggining on lower difficulties easier by helping the player.
But these are just few ideas, I'm simply trying to take out the almost total randomness from the original x-com and make player feel like that the aliens know what they are doing.
-
There aren't "total randomness" in original xcom. It includes Earth attack scenario: from scouts to battlesheep fleet. The only randomness it have - appearence time and order of attacks... Of cource instead of random choises it could use some evauation functions which will include player bases positions, tech level and craft/crew armament... But actually it will look like a cheat from AI. The only thing that AI can use without being a cheater - is XCOM base position, but with one notice - those bases shouldn't have "anti-detection" facilities.
So in short - randomness is not very "stupid" choice, considering there is a attack scenario.
-
Hello,
0.4 is released. To keep our programmers motivated, lets write how many of us want 0.5 ;)
0.5 plans:
Geoscape AI (missions, UFO spawning, bases, scoring, funding, interceptions, etc.)
Graphs
Victory/Defeat
Battlescape:
More alien units
Alien base and X-Com base missions
AI Phase 2 (grenades, blaster launcher, taking cover)
So basically, 0.5 should be first playable (from beggining to end) version. It will be important step in OpenXcom development.
So going back to the original topic of this post, yes i want version 0.5 this year :P
-
Wasn't there a plan to add scripting to OpenXcom? This seems like the perfect place for it. If the aliens' Geoscape AI is implemented as a script then the default ruleset could come with one that just replicates the original game's lookup tables while custom rulesets could use arbitrarily complex implementations – all the way up to "real AI" techniques like HTN planners and the like.
(Come to think of it, this would be very relevant to my thesis paper...)
-
gchevallereau experimented with python
https://openxcom.org/forum/index.php/topic,360.0.html
So, what about such topic of thesis "Scripted AI in OpenXcom" ? ;)
-
My thesis will most likely be on the subject of whether traditional AI methods (usually deemed too slow for use in video games) can be used to generate a basic dynamic plot by only being invoked very infrequently to determine basic policy, with normal scripts doing the heavy lifting. Since I suck at formal analysis I will probably take the "implement it and throw statistics at your measurements until it looks like science" approach.
Of course that means I need a game to add AI to and that game needs to be open source (or at least easily moddable). And I need to figure out how to make it behave interestingly with AI methods. Even though I never got along with Python (it's the language I would love to like) OpenXcom's Geoscape AI might be a good candidate. I'd have to discuss it with my prof, though.
But the important reason for scriptable AI is, of course, because modding is fun. :D
-
Geoscape is scripted, there's no need of any AI.
Try to look into direction of better battlescape AI.
As reminder: "xcom" actually is a "laser squad 2" game, which was "battlescape-only" game. Tactic turn-based strategy. 90% of developer's time have been spent onto battlescape logic and AI. That's the direction which should be improved, or at least made "not worse".
-
Unfortunately, I don't see a way to fit my thesis (which focuses on using AI to drive long-term actions) into the Battlescape (which is short-term by definition). Now, if we made Battlescape missions longer, say 2.000 to 4.000 turns each, so that each mission can have its own individual storyline, then I could see how I can fit that in... But I don't think it would work particularly well.
(And no, I don't think that running an HTN planner for every step of the battle would be particularly fun. There's a reason why most video games stick with regular scripts.)
-
Hi all,
I would like to tell developers, that I love their work and I would love to see EU and TFTD without bugs, playable in higher resolution and with load button in battlescape (as it already is). Continue in your great work please!
I haven't found any other way to donate, than some server support. Can I donate by Paypal directly to support/speed up development?
My opinion to AI - it should be same as original (as much as possible) - original is well balanced and very good playable. If you change AI, it wouldn't be same quality as original and it would reflect your playing style - you may think, that it is made much better than original, but someone else with different playing style may find it unplayable. Yes, original is not perfect, but it is balanced.
So for me - keep original AI, with possibility to use custom AI as option.
-
Can I donate by Paypal directly to support/speed up development?
If SupSuper won't be able to find a job after finishing uni, you could hire him to work on OpenXcom ;P
-
I'm not sure, if he will be satisfied with one time 5/10/20USD pay:D
I'm just used to support good projects that I like (DosBOX, cpu-z, CatBlock...) - because I see the effort behind this projects.
-
The fact you like the project and respect the effort is more motivating than money. (and people disliking the project is demotivating :p )
-
I love your work , i built a entire site to help support it, that got to motivate you ;)
-
Fantastic project!! Keep up the good work. Dosbox is not so lightning quick as your EXE. Also yours doesn't need any Dosbox, plus the mouse wheel, save system and other excellent features makes your work a MUST HAVE!!
I'm continually grinding my teeth looking at FiraXCom, a lot of the original features and suspense was missing from their demo and the full game isn't looking good judging from the videos. Not looking forward too much to play it.. :(
Your project doesn't need Windows 7, runs and looks superb on XP and has much bigger smoke and fire areas than the ridiculously tiny ones in FiraXCom. There are 100 annoying "features" in their re-imagination: no awesome weapon loading sounds? Whaat??! No commerce in the new game??! Eehh...
Thank You for doing this!
-
Fantastic project!! Keep up the good work. Dosbox is not so lightning quick as your EXE. Also yours doesn't need any Dosbox, plus the mouse wheel, save system and other excellent features makes your work a MUST HAVE!!
And you are still not aware of a real-good new feature (my first work in OpenXcom :) ) which is available in the newest GIT-build (2012-10-13) :
Scrolling in Battlescape with RIGHTBUTTON-HOLD like in Transport Tycoon. :)
-
Látom te is magyar vagy! :) Gratulálok a fordításhoz. Hát igen, rengeteg minden került be a GIT buildekbe és eszméletlen hangulata van már így a játéknak!
Can't wait to play the newest additions after all the GiT changes are compiled together into an OpenXCOM 0.4.5 build!! Aarrghh.. the wait kills me along with Sectoids on superhuman! :DD
-
Látom te is magyar vagy! :) Gratulálok a fordításhoz. Hát igen, rengeteg minden került be a GIT buildekbe és eszméletlen hangulata van már így a játéknak!
Wow, és én még azt hittem, hogy egyedül vagyok itt magyarként. :)
Köszi.
Can't wait to play the newest additions after all the GiT changes are compiled together into an OpenXCOM 0.4.5 build!! Aarrghh.. the wait kills me along with Sectoids on superhuman! :DD
I'm still working on a thing, i hope SupSuper waits that before releasing 0.4.5. :)
-
Waiting for it... Keep up the good work
-
I would really like to see a fully playable 0.5 by the end of 2012. Personally I'm not really doing much to spread the word about openxcom because it's not yet playable.
Geoscape AI Discussion: I think the current focus should only be to reproduce the original geoscape progression and externalize the 'AI' charts into an editable file, preferably something more appropriate to long tables then YAML.
AI is just a bunch of rules you make, so agents appear to have intelligence, because they make decisions based on these rules, their state and the state of the world around them.
I don't see what you mean with "real AI". There is no such thing: real <> artificial. Even modern games uses tricks to make people think the AI is more advanced, but it just uses the same old "tricks". ...
If it doesn't react to outside influences or other inputs and use some sort of decision making process then I don't see how it could be an AI. In MOOII if you move your fleet towards an AI empires system, that AI will 'evaluate' a few factors before 'deciding' whether or not to move his fleet to the same system to deter any potential attack. X-COMs battlescape AI likewise assesses the situation in pre-calculated ways and moves accordingly. The X-Com Geoscape 'AI' is not an actual AI because it doesn't attempt to imitate an intelligent response, and doesn't evaluate any conditions, it just launches UFOs. This could create a problem for modders who want to add more divisive AI's. Currently I don't think the geosccape does anything in reaction except have a chance to start a retaliation mission in a region when UFOs are shot down there.
-
Yep, I want 0.5 and final release in 2012 >) I'm super excited! This project is much more important to me than Xenonauts and XC:EU.
-
it'll be close, but i think we can make it!
-
Would be nice christmas gift ;)
-
So parents can put openxcom under the christmas tree for their children... o wait...no... so parents can play openxcom in their christmas holiday! :)
-
yup.
-
I asked my gf to buy me OpenXcom for Christmas. She doesn't know what it is yet... >)
-
I asked my gf to buy me OpenXcom for Christmas. She doesn't know what it is yet... >)
she will when you be spemding more time on openxcom end less time with her,...
(i remenber playing till the next morning ad being surprised that so much time had passed,... o s**t now how am i supossed to get some sleep in)
-
she will when you be spemding more time on openxcom end less time with her,...
(i remenber playing till the next morning ad being surprised that so much time had passed,... o s**t now how am i supossed to get some sleep in)
Lol :)
-
So what's the progress on this?
Kinda funny to be the one asking this time around. :P
-
tbh, i think we're there. i'm only chasing bugs here and there, tweaking things. the only 4 "features" we're missing right now are:
- Aliens picking up items in the battlescape.
- Aliens using mind control against XCom units.
- Aliens using Blaster Launchers.
- TU penalties for weight vs strength.
-
Woohoo!! 8) You're doing God's work Warboy~!
-
Woohoo!! 8) You're doing God's work Warboy~!
As a Christian i should be a little offend by hearing that blasphemy , but since its about Openxcom i will let it slide :P.
-
As a Christian i should be a little offend by hearing that blasphemy , but since its about Openxcom i will let it slide :P.
Oh.. I'm Buddhist.
-
Oh.. I'm Buddhist.
Such a refreshing change from all the Extremist Atheist i normally find on the net :D
-
tbh, i think we're there. i'm only chasing bugs here and there, tweaking things. the only 4 "features" we're missing right now are:
- Aliens picking up items in the battlescape.
- Aliens using mind control against XCom units.
- Aliens using Blaster Launchers.
- TU penalties for weight vs strength.
So, this will be included in next version (after 0.5) or should block 0.5 ? If not, i guess 0.5 can still be released this year.
-
- TU penalties for weight vs strength.
this one's done now
-
nice, scratch alien psi-attacking too.
-
So, this will be included in next version (after 0.5) or should block 0.5 ? If not, i guess 0.5 can still be released this year.
my vote is on putting it in the 0.5
with the exeption of - Aliens picking up items in the battlescape.
i'm still used to aleins stop being a threat after panicing or when they revive out of unconsciousness
-
Hey Devs, now that the Mayans are finished not destroying the world, and indigestible medical tricorders have officially been invented, how about we get a stable 0.5?
What the fuck Devs?
What .... The .... Fuck?
Naw I'm just kidding around. 2013s not going anywhere and last I checked the git-builds were more then fully playable.
-
blaster launchers prove to be elusive.
also there's far too many game-breakers to issue a release.
-
blaster launchers prove to be elusive.
Yeah, I've not really done any more on this since we were last talking about it :( Even with my code added, they seem to prefer normal shooting "methods" over my code when the unit is visible and in line of sight. Seems to skip over all my code, and not figured out why
-
Yeah, I've not really done any more on this since we were last talking about it :( Even with my code added, they seem to prefer normal shooting "methods" over my code when the unit is visible and in line of sight. Seems to skip over all my code, and not figured out why
Maybe that's your problem, you're making them too smart. :P Original aliens didn't care who or what was in the way, they would blaster launch anything possible even if it hit their team.
-
Well, last night on IRC, Warboy and I managed to fix up my code and we had the first AI launch of the blaster. Progress :)
-
i'll say.
PREPARE YOUR BOWELS FOR IMMINENT RELEASE! (https://youtu.be/GZ6hh_FQl3c)
-
i'll say.
PREPARE YOUR BOWELS FOR IMMINENT RELEASE! (https://youtu.be/B_b9iiglHik)
Urm, you've linked to a deleted video
-
Probably the release took place already ;D
-
deleted that one because it captured my whole desktop instead of just the OXC window, modified the earlier post, uploading cropped video now, but it may take a while, because australian internet is awesome.
-
Video is online now :)
-
BOOM!!! ;D :D :( :'(
-
Good job ;D Now I'm waiting for more rulesets.
-
Well, that marks the last time I'll capture an entire battlecruiser with a skyranger full of soldiers in personal armour with laser rifles, LOL.
-
Great work , the best bit was when it explodes :P
-
Well, that marks the last time I'll capture an entire battlecruiser with a skyranger full of soldiers in personal armour with laser rifles, LOL.
The difficulty curve is a lot WAY higher.
Speaking of, will the AI cheat like in the original?
-
i'll say.
PREPARE YOUR BOWELS FOR IMMINENT RELEASE! (https://youtu.be/GZ6hh_FQl3c)
Wow that explosion looks collosal. :o
-
It would have been impressive if the one taking the shot would have aimed a little further so that spotter didn't die :)