This forum is in permanent archive mode. Our new active community can be found here.

Emergent Gameplay: What is it and how can we see more?

edited June 2011 in Everything Else
So, I'd like to get a little discussion here for some of the more game design/theory people about emergent gameplay. What is everyone's definition of this idea and what are some key design elements to help foster it's use in more games? I've always found emergent gameplay to be one of the best selling points of a game and really provides a sense of character and depth to games such as Dwarf Fortress. I'm curious to see what people think.

Comments

  • I like to think of them as unintended side effects of a combination of simple ideas in a game. I suppose they don't have to be unintented, but those ones are more interesting to talk about.

    One example of something i mean is that when I play Ocarina of time, I'm usually rolling every where, or backflipping all the way across Hyrule Field whenever I don't have Epona. It's a little bit weird but I thought it made travelling around a bit more fun, you have to have a rhythm.
  • Emergent gameplay is sort of like DNA or Conway's Game of Life. You have a bunch of small things that all follow a few simple rules. You combine a whole bunch of them together, and the whole is greater than the sum of the parts. All those simple things combined in a complex system suddenly all sorts of things you never expected or intended start to appear. This could be something like rocket jumping in Quake. It can also be Putting a dragon in a trap in Dwarf Fortress to have it guard your drawbridge with its fire breath. It can be beating Zelda with no sword.

    I think I might also include small rules changes and mods. Take siege maps in NS. The rules of one game are tweaked relatively slightly, and suddenly a completely different game appears. Really siege is the same rules as NS except that the map is constructed such that certain impenetrable doors bisect the map until an amount of time has passed.
  • A system in which complex behaviour arises out of the interaction of simple rules.

    I've always thought that the more simple rules you have the more chances you have for them to grow into an emergent system. But the rules need to be able to stand on their own and also feed into each other.

    I also think that the rules need to be applied to the whole game and not small subsets. Like how you can mine/collect everything in Minecraft. Being able to then use, almost, everything to build something new and to be able to place, again almost, down on the ground to build something gives rise to tons of emergent gameplay.

    An example where I think a game designer gave this up and went for a more easily controllable system is EVE Online. There you have one set of goods produced by NPC Controlled stations/planets that can be traded between them. And you have another group of goods to be mined to produce place made good. And then there is yet another group of goods to be mined from comets/balls of ice to run player space stations. And then there is another set of goods mined by player space stations to help you produce ... and so on. They could have made the simple decision to have just one set of raw materials and have everything flow from that, rather than continuing to add more and more and more exotic materials. I see this as a missed opportunity.

    EVE does have other areas where they do keep to a more simple set of rules, or lack there of, where emergent gameplay arises.

    All the maths of science can written down in a few very small formula. The individual concepts are pretty simple too. When you start adding them together you get crazy stuff like stars and galaxies and little ants too.
  • Players reacting to systems in unexpected ways. 1up Chase is a good example.

    Back in "the day" my friends and I invented "Kill the Pimp" in GTA3; kill as many of the pimps in the bright pink suites as possible before you die. If you knowingly kill anyone else (including cops) it was game over as well.

    Monster Wars in Doom 2 was another favorite around my childhood; pick one of the big arena maps (Downtown, Suburbia, Dead Simple, etc), use WadEd to fill it with two types of monsters that can't hurt themselves (basically anything other than zombies), get them to fight and see who wins. Hell Knights v. Revenants and Imps v. Cacodemons were favorites.

    By its very nature, Emergent Gameplay can't be designed before hand. Jamming your game with tons of complex mechanisms based on simple rules is a good way to encourage it though.
  • By its very nature, Emergent Gameplay can't be designed before hand.
    Bullshit. The field of Mechanism Design is specifically designing systems to cause or encourage desired behavior.

    Player actions are a direct result of game design. If players "emerge" a new style of play, that is a mechanical extension of the game's rules.
  • maybe I'm thinking of the subset of emergent meta games as I wouldn't give Miyamoto (or insert the guy who created the system for 1ups tracking Mario here) credit for having thought up 1up Chase.

    Are you talking more about things like Portal speed runs; people using physics glitches and portal bugs to subvert the intended solution to a puzzle?
  • I wouldn't give Miyamoto (or insert the guy who created the system for 1ups tracking Mario here) credit for having thought up 1up Chase.
    The game mechanics allowed it. It's irrelevant whether or not they intended it: it was an extension of the mechanical rules of the game.

    You can try to say that "emergent" gameplay is by definition "unintended," but that's a largely useless definition. Rather, if it truly was unintended AND unexpected, then the game was not thoroughly analyzed at design-time. I'm not saying such analysis is necessary, simply that "emergent" gameplay is a literal extension of the rules of the game, no different from intended play.
  • I think a line can be drawn between the game a designer intended and the game they created; accidental genius is less laudable than intentional genius. Combos in Street Fighter II were the result of a bug (which, according to some, was detected by Capcom before release but ignored since they thought they were too difficult to pull off to matter) which eventually defined how the game was played and was retroactively embraced by the creators.
  • Oh Oh Oh! So if you mention Dwarf Fortress and emergent, what is really emergent about that game is the story being told in the background; at least this is what interests me. The story is told through narrative. Put the words together for a field of study called emergent narrative. You can do a google scholar search and actually find some papers on the subject.

    I'd also like to mention that life, in some sense, is emergent itself. Conway's Game of Life demonstrates how some configurations of the underlying rulesets can give rise to stable change propagation. There are some arguments that life itself works this way, that physics works this way (an electron being a stable configuration from a random change in some goopy substrate), that evolution works this way in part, and that socioeconomic systems work this way.

    If you want to get a very interesting perspective on this concept of how emergent, stable behaviors can be brought about by engineering substrates and encouraging the right interactions, check out Steve Grand's book Creation. If anyone played the game Creatures (1, 2, 3 or Docking Station), Steve Grand is the fellow who put those games together. I'm a little over halfway through it, but it is a fantastic read for perspective.
  • By its very nature, Emergent Gameplay can't be designed before hand.
    Bullshit. The field of Mechanism Design is specifically designing systems to cause or encourage desired behavior.

    Player actions are a direct result of game design. If players "emerge" a new style of play, that is a mechanical extension of the game's rules.
    Certainly, but designers have something in mind when they encourage behavior X and discourage behavior Y. It is called Emergent Gameplay when th interactions of X, Y, and other rules create completely unpredicted phenomenon Z. No designer expected Z to happen or designed their rules to create Z through discouraging or encouraging behavior. They were encouraging X and others, and discouraging Y and others, but never foresaw Z.


    At leas that's my working definition.
  • edited June 2011
    By its very nature, Emergent Gameplay can't be designed before hand. Jamming your game with tons of complex mechanisms based on simple rules is a good way to encourage it though.
    designers have something in mind when they encourage behavior X and discourage behavior Y. It is called Emergent Gameplay when th interactions of X, Y, and other rules create completely unpredicted phenomenon Z.
    Bullshit. The field of Mechanism Design is specifically designing systems to cause or encourage desired behavior.

    Player actions are a direct result of game design. If players "emerge" a new style of play, that is a mechanical extension of the game's rules.
    I think I'm going to agree with Rym on this one.

    If "emergent" were simply defined by being unpredictable outcomes from lower rulesets, then it would be called chaotic or statistical. Chaotic systems cannot be precisely predicted as iterations increase without having absolutely perfect knowledge and precision. Statistical systems are quite obviously random and thus unpredictable. Both lead to unpredictable outcomes (although either can have predictability within some error margin), and both are founded upon some mathematical framework (which could be a set of rules, why not?).

    Emergent properties imply that higher level behaviors emerge as interactions from and between lower level rules. Some of these properties can absolutely be predicted. Engineers can thus design for certain emergent properties to exist. The engineers are creating lower level rulesets, so the emergent play styles are not necessarily forced upon the players.

    However, if you consider cascading levels of emergence, it might be that the lowest level of rulesets are engineered so that the next level (player choices) result in a higher level (player/environment interactions) of emergence that creates behaviors and properties that are somewhat more directed than say those game designers who just make up core rules and let people have at it.

    To put it less academically:
    I would argue that tweaking the shit out of a game through playtesting is examining the emergent properties of the game and then engineering the lower level rules to yield emergent gameplay that better follows the designer's vision. It is the styles of play that aren't within the designer's vision that are most often referred to as "emergent", but I would argue all styles of play are in fact emergent. It is an unfortunate trend of using the word emergent for this isolated set of cases that causes confusion.
    Post edited by Byron on
  • edited June 2011
    well, the dictionary defines "emergent" as either being unexpected or a natural consequence so my argument being "bullshit" is clearly the opposite of true.

    It also lists "unexpected" before "natural consequence" so clearly I win ;)
    Post edited by DevilUknow on
  • It also lists "unexpected" before "natural consequence" is clearly I win ;)
    Touché!
  • RymRym
    edited June 2011
    well, the dictionary defines "emergent" as either being unexpected or a natural consequence so my argument being "bullshit" is clearly the opposite of true.
    Except the other three definitions, nay the primary definition, from your own source, don't note this. The definition you used has two components, and isn't really relevant to the discussion.

    Also, more to the point, you choosing to cite one alternative definition of a general dictionary reeks of confirmation bias. ;^) Words like "emergent," "game," "strategy," or "tactic" mean very specific things in particular fields of discussion: using a colloquial definition there is a futile endeavor, as such definitions are usually overbroad.
    Post edited by Rym on
  • I think he was joking, based largely upon the winking emoticon and flavorful green text.
  • edited June 2011
    To answer the original question about how to design in emergent gameplay, I think the best thing a designer can do when trying to create emergent gameplay is design a game that plays by itself without any other players, and still be interesting to watch.

    If a game can sort of play itself in an interesting way, and then the player gets to hop in there and take part, I think that makes for a good baseline. Dwarf Fortress did this to a lesser extent. Creatures did this very well. Black and White did this to some extent. Sims even did it. These are all games that sort of have something going on even if you aren't there, but once you are there, those goings on enrich your experience.

    I think any game in the artificial life genre sort of has to have this feature. I think games with a strong focus on agent-based modeling and simulation or learning agents will also have this feature. Perhaps all of these have emergent properties that enhance gameplay because the game engine changes actor behavior over time (as opposed to canned or scripted behaviors of the usual game which do not change over time).
    Post edited by Byron on
  • If you want to get a very interesting perspective on this concept of how emergent, stable behaviors can be brought about by engineering substrates and encouraging the right interactions, check out Steve Grand's book Creation.
    Lost my bookmark and began reading this at a random point. Hit the following which I thought was an interesting philosophical framework towards understanding very complex systems, such as intelligence, life, or similar emergent systems. Thought I'd share it. Yes, you can be uber critical of some of the analogies made herein, but again, I am sharing this for a philosophical perspective as opposed to an undeniable and logically self consistent truth.
    We could start whittling away at the brain to see which part of human intelligence is the 'most important'. We can remove the visual cortex -- we know that people can be blind and still intelligent, so that bit can clearly go. Similarly we can cut out the auditory cortex, because deaf people can still think perfectly well. What about our sense of responsibility and ability to make plans? People with damage to their frontal lobes often act irresponsibly and antisocially, but they can still reason and respond to their world, so a prefrontal lobotomy would not seem out of the question. ... we can remove individual types of memory -- the kind we use to remember faces, or the kind that enables us to remember how to ride a bike -- without the whole system collapsing between our ears. Language can go, too -- it may be quintessentially human, but people are still people, even when they have suffered damage to the parts of the brain that handle the interpretation or generation of language...

    In short, we can remove any single aspect of intelligence without destroying the whole. But if we start removing these faculties one by one, sooner or later what is left will cease to look intelligent. There will be a point at which the removal of a single component causes us to cross a subjective threshold between intelligent living thing and lump of meat, but if we start again and remove the parts in a different sequence, it will be a different component's removal that represents the last straw next time around.
    The majority of the book is a tool for generating a philosophic perspective on the emergence of intelligence and life itself as a system of processes that all coexist and interact to create something "greater than the sum of its parts."
  • edited October 2012
    Yo dog we heard you like simulations:

    Also the relevant XKCD.
    Post edited by Starfox on
Sign In or Register to comment.