Open Worlds III

Open Worlds III: When NPCs attack!

I think open worlds are compelling because they're immersive. They feel real. It really appeals to explorer-types, which is my main drive. (What's yours?)

I think creating an open world is a balance between two goals: (1) giving the player places to explore, neat things to see, and a chance to play with game mechanics on his own time, and (2) telling the player what to do next.

Any good game gives the player places to explore. Unless the game is timed in such a way that there's no save, and the player has no choice but to win or die, the player has a chance to run off and chase butterflies. I don't know of any game that doesn't give the player a chance to explore. Even in fighting games, the player can choose to not engage the opponent and spend some time jumping around or seeing what else is in the world. (Chances are there's not a whole lot; background animations maybe?)

Exploration isn't just running around and mapping. Explorers also like discovering and understanding game mechanics. They want to root out the math behind combat. They figure out how to min-max their characters. Yet this isn't part of being an open world, though. Complex game mechanics gives explorers extra goals, but it doesn't really make the world open.

An open world is non-linear. The key here is giving players options. Not only does the player not have to stay in Fooville to work on the main quest, but they also have something to do elsewhere in the world. The golden Ultimas (ie 4-7) had quests all over the world. In Mario World 64, there were places to go grab stars in each of the 'worlds'. In the Grand Theft Auto series, the player might have several independent quest chains they could pursue, or they could just run around and explore and map and try out mechanics. In all three, the player had not just the opportunity to go anywhere, but they also had something to do when they got there.

Just making random encounters and procedurally-generated dungeons to explore isn't enough. The 'something to do' should be game-relevant. Side quests and parallel quests are important here. If the player 'plays out' a region of the world in one session and then has no reason to return, the openness of the world loses its meaning. If you can go back to Trinsic but there's nothing to do there, then the choice of returning is meaningless. This ties into the next point:

With an open world, you get the economy of being able to re-use content; instead of the player only seeing your intricately built environment once, they get to run around inside it over and over. I think the biggest benefit here is that this makes balancing world richness easier. When parts of the world are made by different designers and/or at different times, some might wind up being more complex than others. Now that content is placed over the world in layers, I think it's easier to make that depth consistent.

The second pillar, the other end of the spectrum, is giving the player a goal. When you allow the player to go back to revisit content, you need to tell him he can do so. Tell him what he needs to do to advance the plot. If the player is lost in a big, open world with no idea of which of the billion NPCs in the world he needs to talk to, he'll probably become frustrated and stop playing. And return your game and tell his friends it sucks. Then not buy the sequel, that's for sure. (One way around this is to have several different threads that all point in the same direction; e.g. NPCs in each of your major cities might have their own independent quest to go talk to the next relevant guy. If the player completes one but then gets distracted, one of those other quests should send him back.)

Some players really like being given a direction. It's essential for Achievers. I think all players benefit from having a clear goal to pursue. I think this is a key part of the appeal of class-based systems: tell the player what the measuring stick is. Give the player a clear understanding of the rules in your world. Help them fit in by telling them how your world works.

Linear games might get this wrong, if they don't tell the player what the current goal is. The player might be stuck in one location (it's a linear game, on rails), but not know what it is he needs to do to move on. The player needs to know what to do, not just where to go.

These two pillars don't form a continuum. It's similar to the Libertarian argument: there's really two axes here. You want to make sure the player knows how to move the game along, but at the same time provide him flexibility in what he does next. A great game doesn't need linearity; it needs to avoid leaving the player in a vacuum.

Linear Gameplay vs Open Worlds

Continuing a thread from yesterday.

I'm fine with linear gameplay. Being shoved into a corner sucks, but if a game is good and fun, chances are you don't notice it. On some recent Lost episode, Locke told Hurley to take off (by himself) and head back to the beach. Hurley thought about it and decided it was a bad idea. Ben, perceiving that result as intentional by Locke, told him (paraphrasing) "well done, now he thinks it was his idea to stay."

If your player wants to go the same direction that you want to send him, linearity isn't that big of a deal. HL2 is linear, but was awesomeness anyway. Ditto for Portal. I thought HL2's linearity was, in some places, too linear. It dig bug me. But I got over it.

I think Sirlin would say "no annoying mechanic is worth keeping." (He, and this catchphrase, has been my ghost the past week. Maybe he wouldn't say that, but I'll save that for another day.)

It's not really that hard to add varied paths into a game. Two door and two hallways; complete three goals, but in any order that the player wants. Don't send him to find the silver key then the gold key--let him find either first. Variation has to be planned around; HL2 went through a ton of tuning, and having multiple completion paths can throw testing and tuning time through the roof. Depending on the implementation!

If the player is running in fear (the player, mind, not the character), he's less likely to notice that he only had one real option. If it looks like he had several options, he might later go back and look, on his next playthrough. People don't like thinking they have unexplored options. There is a thing as too much choice. The trick is giving players real choices (because they are going to go back and check on you), but not to make them feel like they missed something. One trick for solving this dilemma is choosing when to give choice and when to take it away.

When you put cool stuff in the nooks and crannies, some players will want to check every single flippin cranny in the entire game. For games like Doom and Quake, that works. For a narrative-heavy game like Half-Life, I think it's a detriment.

Build up a pattern language. Advertise the big stuff, so that the player is focused on an objective that you have control over. If you want to hide something, show it to the player first. That is, show them the object, but hide the path to it. Give them time to explore nooks and crannies when they're not under pressure. When the heat is on, make the path clear. Players love thinking that they're smart, so make it obvious what they should be doing. The fact that they have choice is enough. It's a bit like the Forer effect; in this case, show a player two choices, where one is obviously better, and he'll choose the better one and feel happy about it.

I think this is a good thing. It's a survival trait. Show someone a snake-filled pit and a banana on table, and he'll go for the banana and be happy that he avoided the snake. The 'danger' is almost illusory! But almost. It is real danger. People look for patterns. When there's a blank in the puzzle, their active imaginations will fill it in. Players like using their imagination. Give them a chance to.