Plot and Parcel: Procedural Level Design in XCOM 2
Vložit
- čas přidán 12. 07. 2024
- In this 2018 GDC talk, Firaxis Games' Brian Hess explains how the development team of XCOM 2 survived the transition from the handcrafted levels of 2012's tactical strategy game XCOM: Enemy Unknown to the procedural levels of XCOM 2 and beyond.
Register for GDC: ubm.io/341ZiaZ
Join the GDC mailing list: www.gdconf.com/subscribe
Follow GDC on Twitter: / official_gdc
GDC talks cover a range of developmental topics including game design, programming, audio, visual arts, business management, production, online games, and much more. We post a fresh GDC video every day. Subscribe to the channel to stay on top of regular updates, and check out GDC Vault for thousands of more in-depth talks from our archives.
Man, I want XCOM 3 so badly.
It's weird how I used to be lukewarm about the Fireaxis reboot at first, but the more I played XCOM 2 the more I enjoyed it, and by the time War of the Chosen came and improved the whole package even more I found myself realizing this may very well be my favorite game of the entire last decade.
It's surely the one I spent the most time on, anyway.
Agreed. XCOM:EU was superb already but with XCOM 2 they completely redefined the genre and set the bar at a new high.
I agree that XCOM2 really does improve on XCOM EU, and War of the Chosen even more so. It's very rare to find a remake that's better than the original AND a sequel that's better than the first title.
the system worked pretty much flawlessly. i remember being amazed by how good the game looked and how well the maps fit with the gameplay. this is procedural levels done right.
I find it funny he's doing all this on an Alienware laptop
He's probably a thinman.
Proof that tilesets still work even in 3D games.
I was literally looking for a video of this talk a few weeks ago as I could only find the slides, and now here it is in my CZcams Recommendations. Thanks!
He had enough of this procedural stuff!*
Disclaimer: None of the words from his talk were procedurally generated. Static test only :D
Great talk! Equally technical and conceptual. X-Com is awesome! Easily one of the best games ever.
Unrelated maybe, but the very American assumptions about what kind of urban zoning make sense make me chuckle a bit.
Big roads everywhere
general motors has entered the chat
oh great another one of you people.
"A good player maintains Full cover"
Meanwhile... My ass gets shot through a fucking wall from across the map, by a sectoid with a 2% chance of hitting. Ten times in a row
I have to agree with the cost to cinematics. I've played through the campaign several times and don't know how often I've seen the cinematics for each of the Golden Path missions. (I wish there was a way to turn them off for repeat playthroughs, along with some of the narration for "new" enemies.) It always makes me laugh whenever it tries to load the cinematic, and the entire camera is filled with an ADVENT truck or a random cliff in the exact wrong spot.
Thanks for the very informative video.
Now, where is XCOM 3?????
Hooooo! i have to kill Aliiens AGAIN! Awasome Game!
There are small towns where one residential house (usually old) is in the middle of a commercial area.
Yup. The separation of purely residential areas and commercial areas is very much a North American thing.
Cyberpunk 2077 multiplayer should have these ideas
I feel like the talk's word choice and sentence structure was too "texty" and not "talky" enough, so it feels stiff and stilted and is a bit hard to follow. The content presented is interesting and useful, though, but it requires more effort to understand when it's inside a wall of linear eloquent phraseology that would've been easier to parse with the ability to go back and re-read words. Also, the talk could've benefitted from a "lessons learned" summary at the end bringing up details like designing parts that fit together well physically (the 8-unit base parcel size) and using metadata to make sure they're placed intelligently in the level.
Content in this speech is great and clear. But the speaker needs to learn to speak in public, because this way it is incredibly boring and tedious.
The real question is, how often do level designers get to speak publicly? (Assuming you're not Romero.)
Apparently, more than they are trained for. :( There are plenty of GDC talks that are pretty decent, I don't expect them to be great speakers, but at least a bit better than this. And yes, Romero plays in a different league in this regard.
True, though I'll take boring and tedious if it's also clear and concise.
Go watch Amy Shumer. It's a GDC not TED bs for your amusement.
I’ve been watching a lot of GDC talks and while I agree his speaking skills are not the best, his overall presentation is better than most because his content is rich with information, facts, and insights. Also, he’s competently telling a story. Oral storytelling isn’t just about tone of voice or expression variety - it has to have structure like any other type of storytelling. As he takes us through the iterative nature of development, he’s giving the appropriate details (not too many and not too few) so I can really follow the narrative of their development journey. I’m actually sick of GDC talks where the presenter is a smooth talker but whose content amounts to platitudes and weak analogies. I feel like I’m actually learning about game design during this talk - it’s one of the better GDC talks I’ve heard in a while.
Oh god, XCOM has just been ruined for me. How did I never notice the lack of bathrooms?!
Congrats to whoever made the script this guy read. But they should've put someone who read that beforehand.
One of the blandest speeches I've heard
@@Beosker Because it was given to an actual level designer instead of a speaker.
Brian wrote it. No one likes giving talks if speaking is not one's actual job. But he is the expert!
I loved the maps and overall looks of the first Xcom. In Xcom 2 the maps look sterile and samey, and the dark colors are a major turnoff. Fortunately, the maps added in War of the Chosen were a major improvement in that regard. Overall, the opposite of what the man said was true for me - I prefer seeing the same maps, but well-crafted and done with love than what I saw in Xcom 2. I finished Xcom 5 of 6 times, but I still have not finished Xcom 2.
I found EU had a stronger atmosphere due to the excellent map design. I still like the fact that maps don't repeat anymore but the sterility is more due to the art design and setting overall than the procedual generation I think.
@@iuserprofile Be that as it may, the maps in Xc1 were a major draw for me. I loved the looks of the game. I did't feel the same way about the sequel.
There's a good paper titled _"DOOM Level Generation using Generative Adversarial Networks"_
Problem: Procedurally generated content is noticeable, and doesn't resemble human made maps.
Basically you have 2 ai's: one generating maps, and the other trained to spot computer vs manmade maps.
It's not until you have the 2nd, _adversarial AI,_ that your generated maps begin to feel more authentic.
czcams.com/video/K32FZ-tjQP4/video.html&ab_channel=PierLucaLanzi
@@-John-Doe- This is very interesting, thank you! But I think that they should have kept fewer but handcrafted maps than lots of samey and mediocre maps. There's just too much time to look at them from every angle not to notice the difference
.
My main issue with EU was that some maps (especially the ships) had very static (and therefore learnable) enemy emplacements, which made most of the repeated mission types very samey.
good as a fanservice but, literally, this is what indie devs are doing for 10+ years now, if you do not take other games of the past 30 years into account
XCOM 2 is great, except for the map design which is just "meh" at best.
I understand the rational for procedurally generated content, but even if there was less map diversity in XCOM1, it led to a much larger diversity of tactics.
In XCOM2 maps are too uniform: being anywhere on the map doesn't map much difference, and there's little tactical difference between maps either.
In the end, I got tired of the billions of XCOM2's maps much faster than I got tired of the "only" 80 XCOM1 maps.
We shall agree to disagree. I got bored of XCOM1 maps after just one playthrough to the point that I had a hard time finishing my second playthrough when the expansion came out. XCOM2 was a whole different story for me.
@@GabrielePalma Some maps were pretty bad in XCOM1 ("corridors" with very limited tactical options) and would be frustrating if selected multiple times (Long War mod mitigated that by adding a counter for each map, so the player had more chance to see a new one. It also added more starting points for each maps).
However, bigger maps were actually more fun to play once you knew them, especially the small UFO ones.
As a player I often thought "oh ok that's this map... So I've landed here. I know the UFO would be over there, and there's a good elevated position here. So I'll lead this way. But If I encounter aliens in the way, I know there's also another sub-optimal but decent spot over there..."
It was like having seen a satellite image of the area: it added another layer of planing between the tactical decisions during combat per se, and the strategic decision in the base.
The issue with XCOM2 maps is there way too much good positions to fight from, so the topography doesn't really matter much. You can always get to a good position in a single move.
When I stated modding XCOM2, the thing I really wanted to change was the map generator (unfortunately unlike most of the game, modders don't have access to that code).
My idea was to give each PCP/parcel a "tactical value" number.
For example a big building with lot of indestructible high cover would have a high tactical value, and a road without cover very little.
Then generate the map in the way there would be 1 high tactical value zone for every 3 mid-value zones, and 10 low-value zones.
I would also remove the opaque fog, only keeping the fog of war, so the player could see the map (I would also make senses from the lore perspective, after all XCOM saw it when landing from the Skyranger!).
Then modify the IA so patrols would slowly move toward high-tactical value zones, and then stay there.
IMAO, this would've been an elegant way to push the player to act without using a timer (which feels very artificial, and limits the tactical options by preventing the player to go slowly).
The player could control the cost/reward depending of his play style. For example choosing to go fast, hoping to secure a high-value zone before aliens (but also risking to go too fast, get caught, and end-up having to fight in a bad zone) ; or choose to go very slow to slot every single alien group, but end up having to fight them from a disadvantageous position.
@@europeansovietunion7372 I guess we very much disagree on what makes a tactical game good.
The one thing I disliked the most in XCOM1 was exactly the slow boring crawl that you were literally forced to perform at higher difficulty levels, which very much flattened your tactical options to just one strategy.
The one thing I liked the most in XCOM2 was precisely how fast paced the missions were because of the hard and soft timer mechanics that have been added which forced you to use all your abilities and utilities in a timely manner to get out of the mission alive.
The world is interesting because we are all different I guess.
@@GabrielePalma We're not talking about the same thing. I'm talking about intrinsic map objectives, this has nothing to do with how slow or fast you have to go.
What makes XCOM2 "faster" to play (by reducing boring obvious "choices" exposed to the player) is not its poor map-design, but the new concealment and activation system.
such bouring about levels what about caracters its vore intresting
WAT?
worst comment