Tuesday, September 13, 2016

Local Plateau

We are in an era of flat games. Let's talk about it.

Content has a reliable return on investment. You pay more, you almost always get better content. Better models, better materials, better shaders, better levels, better quests, better music, better sound, better writing, better voice acting.

This reliability means that many devs are focusing mostly on that content. The gameplay is carefully flattened down so that it properly serves up the content.

For example, the "string of pearls" approach Mankind Divided takes. You enter Golem City. You can take it on with any variety of combat or stealth, siding with anyone you want, but in the end you finish the segment and it always works out the same way. So the devs can focus on creating Golem City, not worrying about long-term entanglements. Even the long-term gameplay is nerfed so that players won't be too strong or too weak as they move through the area.

The gameplay gives the player a tour of Golem City. That's its main purpose.

It's not just big games, either: almost everyone does it.

For example, both Cities: Skylines and Project Highrise have a 9/10 on Steam, and they're both programmed "flat". The gameplay is simple statistical dependencies that serve to unlock more content. How you play these games doesn't really matter: the progression is programmed right into the content, and doing better or worse just means slightly faster or slower progression.

I'm not going to assume you need to game dev "better". Flat games sell well and give reliable return on investment, and the fact that nobody plays them a week after launch is probably not a major concern to many studios.

Still... let's talk about how even small mechanical changes can make these games pop.

Fallout 4 is an extremely flat game, and although it sold well, a big complaint was the deeply uninspired gameplay that left it badly scarred in user reviews. It's clear that every element of the game world was designed as a self-contained pearl: you can approach most sublocations in any way you please without affecting the outcome. Similarly, the long-term play of stats and traits was nerfed to the point where it was clear it didn't affect much. The approaches you can take to any given piece of content are largely coded into the content.

However, Fallout 4's survival mode is an overhaul that disables saving, disables fast travel, and forces you to drink, eat, and sleep. This radically amps up the side effects of each sublocation and requires you to constantly go back to a settlement to resupply. The gameplay now involves cautious exploration out from your settlements, and actually rewards you for creating new settlements because it's a new hub that you can restock at. It's far from perfect, but the experience of playing the game is now radically different.

In vanilla F4, you would encounter a house full of raiders. You'd quicksave and then try a couple of things, muddle through, and then heal up and save again. But in survival, every bit of damage you take is a threat and dying will reset you to the last place you slept - which may have been an hour ago. Every success means more, and every tiny expense costs more, and it adds up. You're left agonizing over whether to push forward a bit, head back now, or head back along a detour to get a bit more of the map explored en route.

Can we codify the differences, here? Can we talk a little bit about what this means, how to use it to make better games?

Fundamentally, the best games are more than the devs created. What the players bring to the experience makes it bigger and more interesting than the content seems to indicate.

Easy example: Rocket League. The main game mechanic in Rocket League is "the soccer players are annoyingly hard to control", which doesn't sound like it'd sell well. But Rocket League was very popular and well-regarded, even though the content was limited to a few stages and cars. Well-polished, sure, but not even as much content as a single settlement in Fallout 4.

Multiplayer is kind of cheating, though: the focus is different. What's about single-player?

It's not just about player creation. Skylines and Highrise are flat, although the player can create a lot of content. Hm.

In order to make the player's content memorable, the content has to be recontextualized.

That is, the player has to make your content their own.

There are a lot of ways of doing this, let's cover a few of them. All of these involve making simple gameplay adjustments. In fact, in a lot of cases they are so small that they could be turned on and off in an options menu.

Avatar Personalization
Probably the most common method, the player will adopt your content as their own if they're allowed to personalize it. This is why so many games have adopted RPG leveling systems and costumes, or even allow players to create their own character.

That's effective as long as the customizations don't betray the player. In the old days you could spend precious skill points on things like "animal husbandry" or "tarot reading"... and they'd never come up. That feels bad.

Avatar personalization can be difficult from a writing/voice-acting standpoint. The more freely the player can customize the character, the more different approaches will have to be encoded... or the less the customization will be represented. This is why games with completely open character creation tend to have very generic encounters. IE, in Skyrim your character never "feels" anything and social encounters are extremely basic, while in Sleeping Dogs you have complex relationships with a dozen NPCs and they all feel believable.

This issue can be mitigated by having proxy avatars. This is Bioware's approach: your party members have established personalities, and those can be used to write the nuanced, directed social stuff. The main character never has a real personality outside of the player's head, but because they constantly take Wrex and Garrus with them, they get a lot of anarchist, end-justifies-the-means social stuff. While the player avatar doesn't have much to say, Wrex and Garrus can say a lot and are presumably chosen because they reflect the player's interests. In turn, members of the party (including the player avatar) get adopted and recontextualized into the player's adventure.

Home Personalization
Some games want you to personalize your home. Obviously, Animal Crossing and The Sims are examples. So is Fallout 4.

Personalizing homes can easily fall flat, just like having to choose between animal husbandry and tarot reading during avatar creation. None of the options are bad... but they feel pointless if they aren't integrated into the game.

In order to really get the player to adopt this home, you'll need to make those customizations have mechanical effects.

This is where Fallout 4 falls short. The settlements have no real role in the game, aside from "get attacked and force you to annoyingly run home all the time". Customizing them initially feels amazing, but then you realize that it's all just visual fluff. None of it matters in the game world.

Contrast The Sims. Customizing your home directly affects how well your sims can perform. Layout affects how quickly they can get things done, and more expensive toys let them advance more quickly. Gardens reduce food costs and can even make you immortal. Pools increase fitness, allow for pool parties, etc.

Fallout 4 could have done this, and they could have integrated it fairly well without requiring too much effort or distracting from the core gameplay. For example, allowing you to tag large objects in the field as salvage targets would make the quality of the settlers matter: they can salvage further, faster, safer if your settlements are set up well. Allow the player to radio for resupplies and reinforcements, in a subtler way than calling for cannon fire - again, in ways where the state of the settlers or the settlement matters. Sidequests within the settlements would have been cool, too, like playing matchmaker, helping raise kids, getting in trouble for building a walkway someone fell off of, etc.

Even small details could have gone a long way, like the locals reacting to the player as they come back, or having procedural giant bins full of melons as the melon harvest comes in, or having the locals decorate designated areas as they prefer instead of forcing the player to do it.

Any way you cut it, the basic idea is the same: the player's customizations and activities must have an in-game result.

Adventure Personalization
And then there's the nasty one. How can the player take control over the story being told? Make it their own?

We're not talking about theoretical algorithms for generating stories. We're just talking about gameplay tweaks and level design tricks.

We're going to skip the level design tricks. That's outside the scope of this essay.

For gameplay, we have to understand the overall pattern that the player will follow through the game. In a linear game, this is pretty obvious: if you play Mankind Divided, you will go through Golem City, and you will do it at this exact time in this exact context and come out of it here having had this experience. Linear games are a bit like rollercoasters, and pacing is a core part of the experience.

Nonlinear games such as open world RPGs or construction/exploration games have no rollercoaster track. How can we shape the player's adventure?

We need to create a pattern of paths the player will follow, and then arrange content to be properly paced along those potential paths.

For example, in survival mode Fallout 4, you will strike out from a settlement in an arbitrary direction, probably moving straight for a while, then curving back, or visa-versa. When designing locations, how far they are from a settlement determines how on edge the player is. You can use this to put "lure" locations (multi-section dungeons, etc) at places reasonably far from home, to make it more stressful without needing to actually raise the difficulty.

The smaller encounters nearer the settlement would be encountered either on the way out (low risk) or on the way back (high risk). Well, here's the thing: you know the direction the player is going. So if the player is moving away from the settlement, the layout should be a provocative, gristly head-on challenge, suitable for softening a player up. If viewed from the other side, the encounter should be easier to see and safely approach, giving the player the option of avoiding it while also luring them in with a slower, softer layout.

It's only possible to think like this if the gameplay creates these "safe hub" patterns of exploration. Vanilla Fallout 4 doesn't do this - the hubs are nearly worthless and the player just wanders wherever. Therefore, Fallout 4's world design is about presenting absolute barriers - putting raiders on huge elevated platforms, putting up massive walls of impenetrable buildings to funnel you into a challenge, etc. Those are also valid techniques, but the formlessness of the player's exploration means it's impossible to know how tense the player is unless the layout actively forces tension by being really long and dangerous.

"Safe hub" exploration is certainly not the only kind of gameplay that forms reliable patterns, but it's the example on my mind this week. It's a pretty small gameplay change - you can only get healing items and ammo from settlements, and they're heavy. Other than that, it can use exactly the same polished game mechanics any open-world game uses.

While it is "undirected", you can see how a player's explorations through this kind of game world would naturally form tense, interesting personal stories. Without any magic algorithms or expensive voice acting.


That's my thinking on tweaking gameplay to let players bring more of themselves into your game. Content is a good investment, but small tweaks to gameplay will serve to both highlight your content and allow the player to enjoy it much more personally.

What do you think?

No comments: