next »

[jbarrios]

I have been thinking about buying Starfield and watching several reviews.  I have noticed some patterns of what players were expecting from Starfield.  Generally players seem disappointed because they wanted the following:
   - A large open galaxy to explore where your ship moves seamlessly between planet and space
   - Several planets to explore
   - The planets need to be filled with hand crafted, Eldin Ring level content

You see the problem right?

That would take a herculean level of effort from a game studio.  I'd even go so far to say it's impossible.

Many studios have tried to meet these expectations using procedural generation.  Let the math generate the planets so the developers have more time.  This was done in Mass Effect, No Man's Sky, and Starfield.  Players complained about the procedural generated content saying it felt too similar and soul-less.

So my question is: How would you fix this?

How would you make a game that meets all of these expectations?  Is it possible?  Could you make a game that is open and seamless while also being filled to the brim with content?

[Foxwarrior]

I would simply be Chris Roberts and get people to pay large amounts of money continuously forever in order to fund such a herculean project. Star Citizen does have seamless planets right? I haven't tried it.

[jbarrios]

I would simply be Chris Roberts and get people to pay large amounts of money continuously forever in order to fund such a herculean project

Lol.  I'm sure it's almost done.  He just needs a bit more money

[michaelplzno]

Minecraft has an ever-expanding world that is full of various kinds of mobs that are funded by more people buying the project.

Similarly, most people, when they play a game, don't explore all the content associated with the swath of stuff needed to make different pre-designed planets.

My answer is to be smart about what "special" content you create and have it on the critical path of the user, while simultaneously nudging them back onto the critical path.

It could be done with a tree structure to procedural generation where each region has a direction it pulls you in, slowly upgrading you to a higher level of predesigned stuff till you get to the critical sections.

[jbarrios]

Minecraft has an ever-expanding world that is full of various kinds of mobs that are funded by more people buying the project.

True.  But Minecraft and Starfield are very different games.  In Minecraft the world is like a resource cache to be exploited.  In Starfield the world is like a challenge to be overcome.

My answer is to be smart about what "special" content you create and have it on the critical path of the user, while simultaneously nudging them back onto the critical path.

It's a good solution.  I'm just wondering how to execute this.

What if the player arrives at a planet where the city for the main quest is on the other side of the planet?

What if you craft a section of terrain to be fun and challenging on foot, but the player flies over it in their ship?

[starsrift]

I bought Starfield, and I don't recommend that you purchase it. It's not a good game, for various reasons.

It's an interesting lesson in game design scope, though. Especially when you compare it to TES1: Arena and it's procedurally generated thousands of dungeons as opposed to the (apparently) 50 or so of Starfield. TES4: Oblivion had some 20 handcrafted dungeons* and procedural content for the rest. TES5: Skyrim had 300 dungeons, with an 8-man team doing them. I don't think Starfield could ever fulfill its ambitious scope, but their team has done more previously, than they did this time around. Especially as a huge company, I think if you approach game design as a multi-year effort, you could spend one or several years just making "pieces", or rooms to assemble in your editor, tag them appropriately, churn them through your procedural generator, and then spend enough time to create little touches that make them "seem" unique, or unique enough, you could produce a staggering amount of content. Perhaps BGS had those people working on another title, because this is something they've done before.
* Oblivion's "handcrafted" dungeons were procedurally generated and then gone over with a human touch, they weren't pure manpower. They were also created only in the final two weeks before release - this is a pace of over one a day. Again, game design being a multi-year effort, considering Starfield's creation time of 7 years or whatever Howard is bragging about - imagine 1000 "handcrafted" dungeons at this pace - by one person. There was time to create the illusion of "unlimited", they just chose not to do that.

A lot of sci-fi games consider a star system as a game "room" or area to load, and that would have been sufficient - or if the consoles couldn't handle it, cell-type preloading to create the illusion of a massive game world (which BGS figured out for TES3: Morrowind). Again, that would have been more than what Starfield does, which is to drop you in a seperate game area in orbit around each moon or planet, which you can only depart via fast travel. You have to fast-travel even between different moons of a planet; they each have their own play areas. Again, it's an interesting lesson, if not a great experience.

[LinsonLime]

As the world's leading Outer Wilds™ shill, I try to inject mention of this game into basically every conversation I have. I recommend it in every breath I exhale, but I specifically want to bring it up in this thread because it sounds like it's exactly the hypothetical game you're talking about. You explore a painstakingly handcrafted open world solar system, exploration is seamless, there's a fairly meaty amount of content, and best of all: unlike Starfield, it's actually good.

[jbarrios]

It's an interesting lesson in game design scope, though. Especially when you compare it to TES1: Arena and it's procedurally generated thousands of dungeons as opposed to the (apparently) 50 or so of Starfield. TES4: Oblivion had some 20 handcrafted dungeons* and procedural content for the rest. TES5: Skyrim had 300 dungeons, with an 8-man team doing them. I don't think Starfield could ever fulfill its ambitious scope, but their team has done more previously, than they did this time around. Especially as a huge company, I think if you approach game design as a multi-year effort, you could spend one or several years just making "pieces", or rooms to assemble in your editor, tag them appropriately, churn them through your procedural generator, and then spend enough time to create little touches that make them "seem" unique, or unique enough, you could produce a staggering amount of content. Perhaps BGS had those people working on another title, because this is something they've done before.

* Oblivion's "handcrafted" dungeons were procedurally generated and then gone over with a human touch, they weren't pure manpower. They were also created only in the final two weeks before release - this is a pace of over one a day. Again, game design being a multi-year effort, considering Starfield's creation time of 7 years or whatever Howard is bragging about - imagine 1000 "handcrafted" dungeons at this pace - by one person. There was time to create the illusion of "unlimited", they just chose not to do that.

Very interesting.  Were Skyrim's dungeons hand crafted?  They felt generated.  It seems like Oblivion's solution is the best.  procedurally generate the dungeon, then touch it up manually.  The time frame they did it in seems insane though.

As the world's leading Outer Wilds™ shill, I try to inject mention of this game into basically every conversation I have. I recommend it in every breath I exhale, but I specifically want to bring it up in this thread because it sounds like it's exactly the hypothetical game you're talking about. You explore a painstakingly handcrafted open world solar system, exploration is seamless, there's a fairly meaty amount of content, and best of all: unlike Starfield, it's actually good.

When they were released I got Outer Wilds and Outer Worlds mixed up and I missed Outer Wilds.  I'll have to check it out.