Hacker News Reader

Comment by apacheCamel | Hacker News Reader

This sort of stuff is fascinating to me. There is so much design and forethought to level making that it just blows my mind. Whenever I am on the other end, playing the game, I never notice when the stage is built in a great way. I do, however, tend to notice whenever it has flaws. Maybe reading more about the design of it all will help me be more cognizant of the patience and dedication put into level design.

View on HN

There's a Doom playthrough with John Romero where he talks about waiting until the very end to design the first level; that is, it was the last one he made.

The idea is that by the time you've finished the rest of the game, you're really good at making levels. You have a solid understanding of what makes a passage fair, balanced, engaging, and so on.

The first Doom level is wild. It's really just two rooms and a hallway, unless you go into the third room... Or find the secret way outside, or the rocket launcher, or...

Watching pros play deathmatch in the first level is a trip.

That's the first level in Doom 2. However, the iconic first level in Doom is E1M1: Hangar.

That being true, the point of creating the first level last was also showmanship-- once you've honed your skills on all the forgettable middle levels, you're able to create a top-tier intro (and finale) that everyone remembers decades later.

The best level designs guide you to certain things while making you feel like you discovered or solved something completely on your own.

Sometimes indies use procedural generation as a shortcut to creating levels (or replayability) without consideration for these patterns, and instead of creating an engaging varied scenario they end up being obviously disjointed and break the flow of the player.

For puzzle games, discovering the answer is much more satisfying than finding it in my opinion.

How are you defining "discovering" vs "finding"? In plain language they are synonyms.

Are you saying that seeing the answer is more satisfying than solving the puzzle? (as in a joke or a magic trick that you don't even try to solve?)

i dunno if the parent post meant this, but i personally do feel that a game which seems to "let you do anything to solve the problem" plays better than one which has a "designed solution".

Think an adventure game, where in order to proceed, you have to figure out the intention of the game designer for items and combinations (and pixel hunting). Compare that to the recent zelda game, where puzzles are more free-form - where you can employ physics to solve the problem at hand (such as crossing a cravice by falling a tree trunk across it).

Most indies pay close attention to the level design, even if leaning on procedural generation to get part of the way there. I am part of a 2-person team building an indie racing/puzzle game. I am the "tools" guy and my teammate is the "level design" guy, and I provide the building blocks and the generative machinery, and he massages them into an intentional and balanced experience.

Procedural level generation is only viable in production for games if a professional level designer can step in and tweak it.

Don't know if I agree with most (after all there were 7600 games released on steam in 2017), but I'd say the ones that are successful do pay close attention.

I'm not much of a gamer but a racing/puzzle game sounds intriguing. Is that a novel concept or are there existing examples out there already?

Trackmania has both dev and user created tracks which offer a puzzle type experience (in terms of determining how to get through the checkpoints). The most complex of these are ‘RPG’ tracks which are long, sometimes technically difficult, and often themed.

As a core game concept, it isn't common. We came upon it because we want our portfolio to be games that involve a cross-section of skills and play styles (strategic, on-the-fly planning, twitchy controls)[0]

You can think of speedrunning a puzzle game as a racing/puzzle game though :)

[0] https://colorballsgame.com/

I agree, I think some of the best designs lead you along without making it obvious at all. My opinion is that some of the top tier level design has a strict path on moving forward but lets it up to the player to decide how exactly they want to move down it.

Indeed. Good procgen can be amazing, but bad procgen feels really half-assed. Derek Yu's Spelunky pretty much single-handedly started the current procgen revolution, and he wrote a book about it. :) https://bossfightbooks.com/products/spelunky-by-derek-yu

(He also wrote extensively about it in his blog, as I recall, though I don't have a link handy.)

I really liked the approach the Dead Cells developers took.

They handcrafted tiles, out of which a metroidvania style levels were generated.

It consisted of a random graph generation(with constraints), and fitting the handcrafted elements to that graph.

It took the best of both worlds - a properly designed levels with tons of replayability.

In a sense, it's pretty similar to how theme parks are designed too, with the best ones being structured in much the same as a 3D collectathon level. There's usually a key focal point in the middle of the area to act as a reference, a bunch of smaller ones off to all sides meant to draw people's attention towards smaller obstacles/attractions, and clever uses of scenery and forced perspective to give the impression of a much bigger, more detailed world than's practically possible.

In a sense, it's pretty similar to how theme parks are designed too

Many modern games are basically theme parks.

clever uses of scenery and forced perspective to give the impression of a much bigger, more detailed world than's practically possible

This is why I prefer more sandboxy games with more potential for emergence.

To be fair, being designed like a theme park isn't necessarily a bad thing. Many great sandbox type games use similar techniques in their design, and a lot has been written about how designers created open worlds to incentivise exploration/encourage people to go off the beaten track (like in this talk about The Legend of Zelda Breath of the Wild):

https://kotaku.com/breath-of-the-wilds-biggest-design-secret...

And in the olden days, a lot of this was basically necessary due to technical limitations. You couldn't get Breath of the Wild size worlds running on the N64 or PS1, so games like Ocarina of Time and Majora's Mask had to use careful design to make the worlds feel a lot more complex than they actually were. Same with 3D platformers like Mario 64 or Banjo-Kazooie.

To be fair, being designed like a theme park isn't necessarily a bad thing.

No, of course not. Sometimes I want a theme park! It's just that I'm sometimes promised a world and what I get is more of a theme park. Sometimes one wants the theme park. Sometimes, one wants a more "natural" experience. This also comes in degrees, which is part of why it's tricky.

> It's just that I'm sometimes promised a world and what I get is more of a theme park.

It reminds me of all those cheesy sci-fi/fantasy movies from the '70s that promised "AN ADVENTURE BEYOND YOUR WILDEST IMAGINATION".

I find this the most jarring thing with Rockstar Games, and it was particularly noticeable to me in Red Dead Redemption 2.

It provides a beautiful, immersive world that is ridiculous in its attention to detail. The amount of stuff to do, accurate flora and fauna, weather effects, and NPC behavior, is ridiculous.

But the moment you start a mission, things are on rails and you lose that freedom. Or the moment you commit a crime even without witnesses the law will find you like a homing pigeon. Most entities in the wild (trees, animals) are just elaborate props and there's nothing emergent about them, and random encounters are not random at all, but just selected from a pool of missions.

They're still great, but for me it did mean having to let go of the 'immersive sim' expectation and instead approach it as a ridiculously detailed version of a 'typical' GTA-style non-emergent game.

Good design is always transparent

Reply on news.ycombinator.com