My Photo

Out Now!



Coming Soon


Or Our First Book...

Blog powered by Typepad

« Nintendo Urine | Main | FreeSpeak Remembered »

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d83452030269e200d834bbef1969e2

Listed below are links to weblogs that reference Play Specifications:

Comments

I really like the idea of this. It makes me think of LEGACY OF KAIN: BLOOD OMEN. Of course trying to list every noun and verb in that game would be ridiculous, but looking at the set of verbs as a whole, I can see why the game is such a bittersweet experience to me. It offer so many verbs: suck blood, transform, cast LOTS of spells. The problem is that while so much is offered, so little is implemented and well integrated into the experience. Even with all of these cool verbs, you can see the whole game with only slight variations of: swing sword, suck blood, use magic barrier. All of the forgotten verbs make my heart ache. I like adding the idea of Occam's Razor to this (yes, I did just pull that term from Wikipedia, sue me) Make sure the verbs and nouns you use serve a purpose!

Oops, forgot, one problem I did have though, is that you so adamantly leave goals out of the play specification. From the conversation on "Play Engineer," you can see that I have many issues with full fledged paidia in videogames. Even if you believe the focus should rest on play in the design of a game, I think that it is still NECESSARY to at least make some mention of goals. Even if you want to have a fully paidic experience, I think you should mention in your specification that you are purposefully leaving goals out for that particular specification. The reason I say this also stems back to my comments on "Play Engineer," where I express my views on society as being ludic in nature and striving towards an elimination of paidia. The industry is bound to have a difficult time shifting towards less-ludic experiences because of the way society operates. Even though this shift will no doubt be a beneficial one, I don't believe it can't happen overnight. I believe that it will be more effective to bring change in gradual ways by honing games to very precise levels of ludicness, which in turn will allow designers to "construct" this shift. Ignoring the goals of a game is ignoring what I see to be the KEY to bringing paidia or "less-ludicness" to the indstry.

Also, focusing specifically on the game and ignoring the industy for a moment, the Katamari games show perfectly why addressing the absence or presence of a goal is so important. The fact that these games can be specified with so few nouns and verbs only compounds the effects that a goal will have on their use. (I've never played any of these games before, so if I don't know what I'm talking about, then let me know) I'll use the example of PONG just to be safe.

NOUNS: paddle, ball, sides of screen (I say this is a noun because the ball must reach this area for points to be scored)

VERBS: move paddle (you could include bounce, but the player has no direct control over the ball's bouncing).

We all know the objective of PONG. If we take out the ability to score we are left with three options: 1. Make the ball bounce off the sides of the screen instead of leaving the screen so that the ball will continue to bounce indefinitely, 2. let the ball continue off the screen and reset its position on the field so play can continue, or 3. let the ball continue off the screen and wait and see if it ever comes back. With these options, the significance of the nouns and verbs have changed dramatically. The motivation to 'move paddle' changes, and hence the significance of the paddle itself changes. The more complex a game is (the more nouns and verbs there are), the less likely that a lack of a goal will affect every element greatly. In turn, the less elements present, the more each will be affected by the presence or absence of a goal. So it is vital to address the goals of a game in any analysis.

Finally, I know you said that it is trivial to extend play specification to include goals if you choose, but I'm not so sure. If I choose to include goals in my specification, then ADVERBS become essential. How an action is performed is in many cases what determines an outcome. Not all games follow this, especially videogames designed for younger players, but if I'm playing Pong, moving the paddle CONSTANTLY versus moving it ECONOMICALLY can change one's performance. (Of course this is very dependent upon and derived from the player and not so much the game, and the "styles" of pong playing can be debated over to no end if people feel strongly enough about it, but nonetheless) if it affects the outcome or goal, it then also affects the nouns and verbs. Yeah, you could remove adverbs by instead breaking each verb down into a subset of verbs (example: "move mario slowly" becomes "sneak" in Mario 64) but it is essentially the same thing.

One more quick point (sorry!), the idea of the adverbs being supplied by the player made me think about player's relation to a game. Maybe by including the goal of a game in game specifications, over time we might find clues which could point towards a model for an apartus between player and game!

I really like where you're going with this "play specification," I just think we can't ignore the ludic.

This is a crisp, concise foundation upon which plenty of great discourse can be build. I agree with Donald that adverbs are very important, and I'm also interested in apply prepositions to the grammatical analogy.

Your comment about more nouns and verbs adding to the time-scale of a game's development really struck a chord with me. I'm going to be laying down the part of the MC design doc that I now realize amounts to play specifications, and if this thing is going to get out by Christmas I need to design it accordingly, which might just benifit the tightness of the game. Donald's comments about LoK are appropriate to this effect.

I should note, however, that in my case, social verbs can be relatively cheap to add, since I'll have to design and have implemented an according engine, so theres a point of diminishing increase on the time involved. Implementing spells involves spending time on particle effects and/or animations, but the social stuff with the faces can be done procedurally. One thing I'm interested in with MC is having a rough balance between the number of nouns and verbs, something that hasn't really been done before. I suspect nouns are generally more expensive since they by definition require art assets.

Whats really hot about the design is that the interface and principle data structure is an alphabetical language, so theres a synchronicity between play specification, interface and implementation. Storytron takes a similar approach, on a larger scale, a sort of inverse of the Sapir-Worph hypothesis, where the language defines the world. It seems though, from what your saying, that this is true of all interaction models.

Don: I think I should clarify the goal situation here... Goals form part of the *structure* of a game, and the play specification does not include structural elements intentionally - because it is focussed on the central play. But I don't want to suggest that the play is all of the game - just that it is the often overlooked element of the game, and a good place to begin the design process.

I believe we need a different notation for structural issues, such as goals and progress structures. Perhaps we should start thinking about appropriate notations for expressing structural issues.

And one last thing, never apologise for being verbose on this blog! People should feel free to say whatever's on their mind. Thanks for sharing your point of view!

Patrick: I'm keen on keeping to a nice tight specification for my own personal use, but that doesn't mean I'm not interested in what other people might achieve in extending or modifying the notation.

On the issue of expense, verbs are usually more expensive than nouns in my experience - verbs imply animations (which are more expensive than just basic art assets) and/or game subsystems (which are also more expensive than just basic art assets). The only exception is when verbs are rendered in the old text adventure style i.e. they have an effect but it is not delivered in the game world but rather in the player's mind (or the game substructure). Your comment on social verbs is apposite here - social verbs, where responses can be delivered in text et al, are much cheaper to implement.

I can't believe you can get Magic Circle out in eight months! This seems optimistic. I hope you have a fallback position. :)

There's always going to be a relationship of some kind between play specification, interface and implementation - and indeed, games tend to feel most ragged when these elements are out of step. But I didn't intend to make any kind of global statement in this regard. :)

Must make a note to think about notations for game structure... Thanks for the input!

I see what you're saying, I misunderstood the article when I read it the first time. After your post, I tried to come up with any other ways I could play devils advocate, but I think you've got a pretty tight formation. The system stands well on its own. If adverbs or prepositions, or any other grammatical goodies are added, then I think they are the products of game structures or the player.

I've got a somewhat related quesion for everyone. Have you all read Craig Lindley's 'Semiotics of Time Structure in Ludic Space as a Foundation for Analysis and Design'? If that's a silly question let me know. : ) When he talks about patterns and player gestalts, I have this urge to apply it to this lexicographical specification model. If you set up the play specifications by providing nouns and verbs, undoubtedly many players will squash many of these nouns and verbs together into larger entities which we could call patterns or (with caution) player gestalts (Mortal Kombat's uppercut, step forward, uppercut, step forward, continuing on and on). Could this tendency of players be taken into account within the play specification? Or is that too removed from the objective of play specification? Can we just assume that whatever game structure is developed will mediate it, and thus it should be addressed in whatever game structure model arises (such as Lindley's)? Or is there some other place for this, such as some undeveloped approach that centers on the player? Or perhaps all of the above? What does everyone think?

Yeah, I'm realizing that Spring Equinox 2007 is a more realistic release date. I was thinking it was vital to go for a release in time for Christmas, but the marketing may be more effective when the market is less crowded. Either way, it probably won't happen by November or even December. It took your reality check to make this gel for me.

Patrick: the other thing is, you are going to be an admirably small product - small products should *never* release in December. Yes, sales are up for the Winter Festival - but the benefactors of this are *always* the big titles. Smaller titles would generally do better to wait until afterwards to release.

Donald: With regards to the squashing together of verbs and nouns to produce higher level patterns... It comes down to how you want to use your play spec. For instance, I would spec Mortal Kombat as having basically one verb: fight (and two nouns: avatar and opponent). That would be the end of my play spec for this game! :)

Another person could justifiably break my verb (fight) into many constituent verbs et al to capture the minutae of the game, and have a more detailed view of the game mechanics. This isn't what I want to do with a play spec, but it's perfectly valid.

I haven't read the paper with the amusingly convoluted title! :) Here's a link for anyone else interested: http://www.gamestudies.org/0501/lindley/
Now printed, and on my reading pile.

I've had a quick skim just now... What is Craig trying to do here? It seems very reductionistic in focus, and just a little taxonomic too, but there's little clue as to what the resulting model is useful for. The abstract says: "...a foundation for more clearly integrating design choices within a coherent overall concept." This suggests he believes his model is useful in guiding game design in some nebulous way; the implication might be that we currently lack a coherent overall concept to guide game design. Is this the case? I'm open to be persuaded.

Also: "...laying the foundations for a more systematic study of possible correlations between design features and play affects." Well that would be interesting work, to be sure, but I'm doubtful this work facilitates it directly. Still, his goals are interesting - I look forward to reading his ideas.

The academic community produce some fantastically interesting work on games from time to time, but I wonder sometimes what it's all about.

I've had some ideas about structural issues over the weekend which I'll write up when I get a moment. In brief, though, it occurs to me that 'goals' (which I'm more inclined to see as 'tasks' in a structural framework) are sentences: e.g. "collect 8 stars", "kill all monsters", "Beat all other vehicles". Verb (counter) noun. 'Collect' in this case is a verb which applies at the level of structure, and not at the level of play, per se.

Not sure how useful this approach will be, however, as the interesting part about the structure is often the degree of parallelism - is it linear (task follows task) or parallel (tasks lead to tasks) or open (all tasks always available).

Will jam on this when I get a chance.

I’ve been thinking about this stuff a lot lately, and it’s starting to come together in my head. Here’s where I’m at:

Games can be defined by a hierarchy of verbs and nouns (forgetting adverbs and adjectives for a moment, though I think they are important as well).

At each level there are complete sentences (in the Noun-phrase + Verb phrase sense). The highest level is "Play Game". Verb, noun. You can break those into smaller chunks - perhaps "Beat Mission" or "Build House". These in turn can be broken up further and further.

At some point though, you get to the bottom of these hierarchies where the verbs and nouns can't be divided anymore. Jump, Select, Shoot. Player, unit, tile, power-up.

Now the top layer isn't very useful, it just sums up the whole system. But the mid and bottom levels are really interesting. It seems to me that the mid-level sentences are the subjective goals and activities that emerge from the very concrete bottom level verbs and nouns.

I think there’s a lot you can begin to see once you break a game down in this way. For example, I’ve recently been fleshing out the idea that the difference between learning and mastery (as in “Easy to learn, hard to master”) is based on the difference between verb complexity and noun complexity. Verbs being the things that the player has to learn in order to understand how to play, while the nouns define what contexts the game allows (or how big the possibility space is, as Will Wright would say). The more contexts/game states that can occur, the longer it will take for a player to master the game.

This is even measurable in a rough way when looking at breakdown of a game. To get an idea of how hard a game is to learn you look at the dimensionality of the verbs (here adverbs are useful – each of them is at least one dimension), pretty much as done with control schemes in 21st Century Game design. But you can also do the same thing with the nouns and adjectives to look at how hard a game is to master. Most games have many, many more nouns than verbs, resulting in a huge possibility space (often well over a hundred dimensions). The trouble is eliminating all the non-orthogonal dimensions and figuring out when counting multiple instances of a noun is appropriate (in Starcraft all the units on a map certainly affect your game at all times… but in Half-life do enemies that are three hours ahead of you in the game affect your play at that moment? Not really. But they does that mean that they have less of an effect on mastery… I’m not sure). Still working on that one.

Another thought is that you can use rules to connect all those little verbs to adjectives on the noun side if you want. Jump affects the Player’s position. Shoot reduces a gun’s ammunition and maybe an enemy’s health as well. Maybe you throw in some Adverbs… weapon type for example, which changes the way shoot affects the enemy…Pretty soon you have a complete description of the system of a game. I guess the trick is knowing where to stop – though to be honest a lot of games are simpler than you’d think when viewed at this level.


So yeah, interesting stuff. To give credit where credit is due, the hierarchical ideas come from Ben Cousins, and the verb/noun stuff comes from a couple places – Chris Crawford, and this blog primarily, with maybe a little bit of Koster thrown in for good measure.

Somehow this turned into a small essay. Sorry about that…I guess I got tired of lurking and not posting.

-Brad

Brad: never apologise for rambling in the comments here! I actively invite it. People should feel free to take as much time as they like to share their thoughts, and it's always interesting for me to read, and doubtless some other visitors too.

I shy away from hierarchical models (easy to get decieved by seductive reductionism...), but there's no doubt that one can apply this model usefully. The distinction between "mid level" and "low level" verbs was raised by Patrick before (although I can't find the comment now!) as micro and macro time scale issues, but I'm only just now fully internalising the idea. I have a piece on structural specification which will touch on these "mid level" verbs - I'll probably write this next week. In short, I agree with you that this mid level is intimately involved with tasks (i.e. goals) - which are sentences of a kind - and hence tied to game structure.

"The trouble is... figuring out when counting multiple instances of a noun is appropriate"

This is a facinating aspect of subjective specification, because one has so many choices! For the katamari games, for instance, it was possible to treat all the nouns in the world (everything but the katamari itself) as one collective noun. This in turn showed something about the underlying game design. At the base level, I think there is much to be said for modelling in the fewest number of nouns (letting adjectives distinguish between nouns when appropriate).

Thanks for the comment! I always appreciate it when people leave a mark that shows that something I wrote provoked some interesting thoughts elsewhere. It gives me an illusory sense of relevance. :)

"I shy away from hierarchical models (easy to get deceived by seductive reductionism...)"

I find this interesting. To me, hierarchies are perfect because they show a complete spectrum from entirely holistic (the top) to entirely reductionist (the bottom) along with everything in between. Maybe it depends on how you look at it though. In complex situations like games there is a bit of emergence at every level – at every step up from the bottom you get a little bit more than the sum of the parts. Without that I think you’d definitely end up over-simplifying. Is that what you mean, or am I deceived? 

As far as using the smallest number of nouns and differentiating with adjectives, I entirely agree. But I've noticed that how many times those nouns are instanced in a game still has a large effect. The simplest example for me would be the difference between a game of Go on a 9x9 board, vs. a 21x21 board. The smaller board is much easier to grasp. It produces shorter games with more recognizable patterns. The larger board, even though it uses exactly the same verbs and nouns, is orders of magnitude more complex.

In my breakdowns I've been noting the number of instances of each noun that exists in a game. This isn’t a perfect solution though, because in most modern games the player is exposed to only a small portion of the overall complexity at any given time (I think usually more for the designer’s sake than the player’s) and I’m pretty sure that lot of linear games are not as complex as they look – certainly not as complex as a 21x21 game of Go which takes a lifetime or more to master.

The solution seems to be to break the game into chunks, and then add those chunks together instead of looking at it as one big piece. Metaphorically count four separate 9x9 games instead of one 21x21 game, because even though the total number of instanced nouns is similar, four small games is still much, much simpler than one big game (the difference between addition and exponential multiplication).

Where to separate those chunks is the question. I’m not sure if it’s entirely subjective or just very, very complex. Either way, it makes a huge difference to the apparent size of a game’s possibility space.

And last, as far as relevance goes, I've basically been getting a cutting edge education in game theory and design by reading blogs for the last few years. For me at least, all of these discussions are extremely relevant (and I wonder how many others there are like me). So thank you for taking the time to write and post, and respond to comments.

Cheers,
-Brad

You seem quite focussed on state spaces... This is a very ludic approach, but it's also something of a limitation. It's looking at the games of Game Theory, not the games people play, perhaps. The state spaces of a GTA game are relatively flat and simple, but the play is relatively extensive. Still, different strokes... I suspect you probably enjoy playing strategy games. :)

My issue with hierarchical approaches is simply that people get very confused between the model and the reality of the situation when indulging in 'systems building' exercises, and the more complex the system, the more decieved they become. This is Alford Korzybski's "the map is not the territory" once again. There are times when thinking hierarchically is an asset - but one must be careful, because the illusion of meaning at a certain scale is an observer-mediated issue, not an objective distinction, per se. In this regard, my issues should be considered philosophical and not pragmatic.

Regarding decomposing structure into component segments - yes, I am looking at this right now with my structural notation (post pending). It really brings out the difference serial versus parallel structures make to the options for play. Is it entirely subjective? Yes, by necessity. But one can build a useful and replicable model on subjective principles. Or, one can define (subjective) instrumentation and proceed "objectively" - but here one has just shifted the subjectivity problem into the instruments. This point, however, is probably less than helpful to the issue at hand. :)

Best wishes!

Only time to read the post and first few comments, so perhaps this has been addressed - apropos of Donald Gay's comments on goals, is not a goal a collection of nouns and verbs much like the player avatar or indeed any higher in-game construct?
E.g. Football game ->
Score a goal == {N [player], N [ball], N [net], V [strike]}
(where the last includes kicks, headers and Hands of God).
Thus I'm thinking expansion may not be necessary.

Also:
"when game design is focused purely on the player's goals (as it is in many shoddy video game designs) the result is often an abysmal wreck because the play is ignored."
Hear hear! Optimal experience is supposed to be autotelic, thats almost axiomatic. Designing any experience around reward-motivated goal-states actually damages its enjoyability. After all, what do gamblers get addicted for? The certitude of financial ruin?

zenBen: yes, you can render goals in nouns and verbs, but often the goals are framed at a different degree of abstraction. I'll be putting up a preliminary post on structural specification soon which will talk about this some more.

Regarding gamblers, I feel honour bound to observe that one of my friends is a professional poker player and makes a decent living at it. In fact, he makes a lot more money than I do with my 'respectible' job. :) A story for another time, perhaps...

Looking forward to more on structural specification. I think it might be significant in the search for accurate online player profiling, although I can't detail why right now [bit busy].
Professional poker players aren't really playing though, are they? They're working, just like all professional sports people. They're just lucky enough to work on something that is extremely conducive to the Flow state [and earning more than most 'insert own adjective here' jobs].

Yes, it's a fair point; working the tables can be hard work. It's worth noting, for balance, that although the earnings potential for professional poker players is very high, it's coupled with the potential for big losses from time to time. Most jobs have no equivalent risk. :)

Hope whatever it is that is keeping you busy is engaging!

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment