Symposium (4)
Symposium (end)

Symposium (5)


  • One more play spec
  • Assorted discussions about matters arising
  • Tomorrow will be the last day of the symposium

Today's Contribution

Corvus of Man Bytes Blog has contributed this detailed play spec for Thief: The Dark Project, which you can see here with its original comments.


  • Move
    • look: up | down
    • turn: left | right
    • walk: forward | back | left | right
      • sneak*
      • dash*
    • Swim: up | down
    • lean: left | right
    • shift perspective: crouch | stand
    • jump
    • mantle | climb
  • Hide
    • hide self
    • hide body
  • Choose Objectives/Difficulty**
  • Purchase
  • Pickup | Drop
  • Equip | Unequip
  • Restore Health
    • Eat
    • Drink (potion)
  • Open | close (door)
  • Open | close (chest)
  • View Map
  • Attack (use equipped object with NPC)
    • strike: chop | slash | thrust (blackjack, sword)
    • aim (bow)
      • zoom
    • shoot (bow)
      • burn (fire arrow)
      • suffocate (moss arrow)
      • knock out (gas arrows)
      • sanctify (holy water arrow)
    • throw
      • burn (mines)
      • blind (flash bomb)
      • burn undead (flash bomb)                    
      • knock out (gas bomb)
  • Interact (use equipped object with environment)
    • Pick lock
    • Shoot (bow)
      • distract (noise arrow)
      • clean blood (water arrow)
      • douse torch/fire (water arrow)
      • light torch (fire arrow)
      • muffle (moss arrow)
      • create path (rope arrow)
    • Throw
      • create noise (object)


  • Garrett
  • Bystanders
  • Enemies
    • Town Guards
    • Hammer Guards
    • Undead
      • Zombies
      • Restless Dead
      • Hammer Spirits
    • Creatures
      • Burricks
      • Spiders
    • Craymen
  • Surfaces
    • Wood
    • Stone
    • Tile
    • Carpet
    • Grass
  • Weapons
    • BlackJack
    • Sword
    • Bow
  • Arrows
    • Broadhead
    • Water
    • Fire
    • Rope
    • Moss
    • Gas
    • Noisemaker
  • Tools
    • Lockpicks
      • Square
      • Triangle
    • Explosive Mines
    • Flash Bombs
    • Gas Mines
  • Torches
  • Holy Water fonts
  • Loot
  • Potables
    • Food
      • Bread
      • Cheese
      • Meat
      • Vegetables
    • Health Potions
  • Incidental (cannot be equipped)
  • Notes


  • Health
  • Visibility
  • Noise level

* Modifies walk. I considering making velocity an adjective, but opted for this notation instead.
** Since you can do this for every mission at the load out screen, I feel it’s appropriate to include it as a gameplay verb.

Chris comments: I'm going to suggest that 'Hide' doesn't quite work here as a category, as 'Hide Self' and 'Hide Body' describe two utterly unconnected activities in the game with nothing in common. Perhaps this should be two seperate verbs: Hide and Conceal? That aside, let me just say: what a convoluted specification! The cause: convoluted game design! :) I don't want to criticise Thief, but it's small wonder the game didn't do that well in the open market, and it's equally unsurprising that it has garnered a loyal fanbase in the dark recesses of game fandom (where complexity is an asset and not a liability). I only played the demo of this game, but that was sufficient for me personally. I imagine that Deus Ex, which I also only played the demo of, has a play spec of similar complexity since it shows the exact same disparity between appeal on the open market coupled with praise from fandom resulting from excessively complex game design. Both these games are trying to be closer to a narrowly channeled RPG experience than a flat FPS experience. For certain players, the games succeeded admirably.

Thanks for taking the time to do this one, Corvus! It's a great counterpoint to the simpler games we've already looked at.


A. Matt asks if I plan to maintain an online database of play specifications... To this I reply: I'm not yet ready to become a librarian. :)

B. Peter asks if I have considered how play specs might map onto regular expression languages or EBNF... An icy shudder echoes down my spine at the mere mention of Extended Backus-Naur form. I sincerely hope that FreeSpeak was my last brush with EBNF!

C. Peter also echoes an earlier suggestion (by zenBen) that this could be adapted to an ontology. Yes, it could, but I have no interest in doing this at this time. I completely agree that this could be done, and also agree that it could be useful, it just isn't necessary for what I want to use play specs for at this time.  I am interested in people's ideas for how this might proceed, however. For the time being, it is the subjective nature of the play specification process as currently defined which interests me, as it brings the barrier for use down to a comfortable minimum.

D. Donald raises this concern:

"I have a concern regarding these different "levels" of verbs: With the middle ground verbs (those determined by the player, such as 'prepare ambush'), some games plan for and facilitate these, while others just allow them incidentally. This seems like a pretty important distinction to me, and to group all of these together into one 'middle ground' category doesn't work... If a game designs to facilitate tactics, then they are an important part of the design, if it doesn't, then the tactics aren't game design, but player reactions."

It seems to me that, like Jose, you are letting your attention be drawn to the tactical and strategic choices the player might make in order to 'win' at the game. But play specifications aren't expressly about this. In fact, it is the least interesting aspect of the process for me. Reading Jose's description of the tactics he employs in Robotron was a facinating glimpse into how he plays this game, but it didn't in any way change for me my conception of the underlying play of Robotron, which his basic specification caught completely.

The argument above rests on the assumption that whether something has been designed intentionally or emerges organically out of the game components is a cogent distinction, which obviously from some perspectives it is. But from the point of view of play specification what is of interest are the activities the player can undertake - the play of the game - whether these elements arrive from accident or design is tangential.

He adds:

Strategy games are designed knowing that "flanking" and "ambushing' will occur.

If we were doing play specs for strategy games we'd see a very different pattern from shooters - because strategy games thrive on complexity. More and more of the play content of a strategy game will rest in the middle ground, and it will not necessarily be clear what elements of the middle ground of a strategy game will have arisen from the original intent versus emerging during development. From our perspective, as a player, it doesn't matter what was intended, does it? Play specification (in regard of an existing game) is not an attempt to read the minds of game designers, but to record the play of the game from the perspective of the player.

This topic continues...

E. Jack backs up Donald:

I think that's a great point, Donald. There are games out there, particularly some of the older games, that have become highly associated with certain tactics or playstyles that are at best perhaps "suggested" by the game, but not really actively fostered or inculcated in the player. In these cases it doesn't seem appropriate to attribute the relevant verbs to something that wasn't designed in--even if many players play it that way because they are experienced or they know better, I think the case that should be assumed is the player plays the game as given.

To this my question must be: why does it seem appropriate to attribute the relevant verbs to something that wasn't designed in? If it's how players play, it's part of the play of the game. Game design creates systems - when I design games, I actively hope that I am creating something that will be more than the sum of its parts. I certainly hope to avoid 'loopholes' in the game design that cause emergent dominant strategies (which reduce the variety of play by becoming necessities), but I actively want the game to lend itself to the creation of new play inside the core components. The design of Ghost Master is built around this very assumption! That is why the game comprises of so many elements for the player to experiment with and combine creatively.

F. And Kim is in the same mental space (leaving me feeling rather isolated!):

Take the FPS example of jump. Am I jumping over a pit of spikes (precision, direction important, may take my time) or randomly jumping during a firefight to make myself a harder target (precision, direction less important; orientation important, perhaps timing important).

Here I would contend that the purpose of a base-level Jump verb is environmental negotiation, therefore the verb Jump (an immediate verb emerging from the interface) is sufficient to describe the first case - jumping over a pit of spikes. The second case, jumping to avoid being hit could (and perhaps should) be specified as a  different verb, one in the middle ground in which the purpose of the activity is to make oneself harder to hit, perhaps described as Dodging (for the sake of argument); presumably weaving from side to side is part of this same middle ground verb, showing how verbs in the middle ground emerge from those at the base level.

As an aside, the fact that FPS games often include such inane behaviours as characters hopping wildly around to avoid being hit is one of the reasons I find them hard to take seriously! :) The discrepency between the play of a game which invites this and circle straffing as part of its central play, versus (say) paintball (in the real world) is striking. Very few games offer play which relates to real world gunplay at a visceral and personal level (although those that seperate movement and shooting like RE4 are arguably somewhat closer!) I'd like to suggest this is one of the reasons why watching people compete in our current FPS games is not much of a spectator sport - unless you play these games and can interpret the deranged motions in context, it looks quite ridiculous indeed! :)


Let me take a moment to reiterate my intentions for a play specification system:

  1. Critical analysis of play; in this role, distinctions between what is and isn't intended by the game design might be more pertinent, but I still maintain that part of the goal of game design should be to create more play than the base subsystems of the game suggest. This is open to debate, but I'd be surprised to hear a professional game designer suggest that they weren't happy when players create new play inside their game systems (except when the players stumble upon a 'dominant strategy' which reduces, rather than expands, the effective play of the game).
  2. Focusing the game design process; in this role, the most important advantages I see are better interface design (by seeing the relationship between controls and verbs) and a clearer understanding of the intended play of the game i.e. an appreciation of what the player will be doing. In this regard, the middle ground is considerably less relevant except in games whose play clusters in the middle ground (e.g. strategy games).
  3. Teaching game design; this is really an extension of (2).

Now for my purposes, the base level verbs and nouns are much more useful than the middle ground, which becomes interesting only when the middle ground defines verbs which embody entirely new elements of play - as with the verb Navigate in many games - the actual tactics and strategems players employ are of less interest to me. But this, of course, reflect my interest in play specifications and is not the only approach by any stretch of the imagination.

I would suggest, however, that if one's interest is in the tactics and strategems players deploy that play specifications may not be the best tool for the job.

Jose raised this issue to begin with because (allegedly) Chris Crawford has used the simple base level verb constituents of a shooting game to criticise the play of these games. Jose's defence is to focus on the middle ground and show the complexities that result from these base level verbs. I contend: if you enjoy the basic verbs of an FPS, then the complexities in the middle ground become your entertainment and hence become relevant. But if you do not enjoy the play implied by the verbs of the FPS, then the middle ground is irrelevant and never comes into play. Either way, the stepping point is the basic verbs.

From my perspective, where my interest is chiefly in how and why people play and enjoy different games, I feel strongly that the basic verbs are the most relevant component from an overall perspective. But I also freely acknowledge that in more complex games such as strategy games, the focus of play moves away from the basic verbs and into the middle ground, and also that players who enjoy certain games will increasingly find their enjoyment rooted in the middle ground, even though the play itself is formulated at a lower level.

Well, I now turn the floor back over to everyone else. What do you think? Am I being too narrow in my perspective by minimising the importance of this particular issue, an issue which at least four contributors to the symposium all believe might be crucial? How much do you think this whole issue hinges on whether one chooses to pay attention to the agon (competition) of the game rather than the mimicry (simulation)? Do you, like Jose, find the middle ground more interesting than the base level of the play specification? And if so, why? Do you think this reflects how you personally play, or something more fundamental? And do you think the intent of the game designer(s) is pertinent to the play that results, or simply the starting point from which the eventual play emerges?

I look forward to hearing your views.


I'll tie up the symposium tomorrow to make room for the Round Table. Thanks to everyone who has contributed so far; it has been fascinating to explore these issues in an open forum like this, and I am finding the areas of agreement and disagreement to be equally intriguing.

Best wishes!


Feed You can follow this conversation by subscribing to the comment feed for this post.

Great summary of some very interesting topics. To elaborate on what I was saying about "built-in" or designed elements vs. elements that may be supported but not likely to happen without a wider awareness...
Now that's far from the common case these days, with how widespread and interconnected gaming culture is because of the internet (example: this round table :) ), but to me it seems wrong to credit the game or game designer with something they didn't come up with. Yes, it's surely a positive that the game can be played in a wildly different style than originally conceived, but is that a proper credit to the design itself?

I'm struggling with what I'm saying here, but I think this is an example of what I'm talking about. Take First Person Shooters before the widespread standardization of the WASD + mouse control scheme: Quake was playable and probably most enjoyable once players adopted WASD+mousing. And yet, the default control scheme was much more akin to Doom than the FPS games that would follow Quake; mouselook, now standard, must be manually entered as a console command (or maybe I just never saw it in the options?), and to use WASD the keys must be assigned.

My point here is that I think it would be amiss to credit the game itself, or the game "out of the box," with WASD+mousing control as a significant feature. Yes, it was quite possible and preferable once you learned it, but that knowledge must be gained outside of the game. The game allows you to play it, but offers suggestion for its adoption other than some rather esoteric-seeming options. So a hypothetically isolated gamer, or an anthropologist booting up emulated Quake three hundred years from now, would more than likely not enjoy the benefits of this control scheme, as he or she had no one to learn it from.
These days there are obviously plenty of games that build on legacy controls as well as the strength of a huge fanbase to teach or spread concepts (like MMOs, obviously), but I think there ought to be some recognition for games whose play systems and important "tricks" are self-contained, and not contingent on cultural or social knowledge.

So I hope this clarifies that I am not arguing against speccing emergent qualities of gameplay like Ghost Master or other games that take advantage of that sort of play and suggest it readily to the player; rather, that games ought not to be given credit for what they themselves did not originate or promote in play.

There is one important commonality between hiding yourself and hiding a body in Thief: use of shadows for concealment. That's why I bundled them the way I did. I don't think it hurts to separate them out though, just a matter of preference.

Perhaps, if the tendency is for people to push past the core verbs designed by a game's creator, a different label for the project might resolve the issue? 'Play Specification' seemingly puts the emphasis on the play, not on the design. That would tend to put people in the "when I play this game I..." seat, which results in the inclusion of tactical verbs.

I don't think you're being to narrow, however. One of my guiding design principles is to try and create solid verbs without dictating HOW (or in what situations) the player will have to use them. It's difficult at times to strip that expectation out of the design and I found working on the Thief spec really helped me find some clearer paths through the process.

You know, it also just helped me figure out why I don't care for the Hitman games. They designed this great core set of verbs and then structured the game so that you're forced to use them the way THEY intended. Slippery bastards.

Anyway, don't give up on figuring out how to clarify your intent. 'Design Specification' doesn't work for me, as that seems too inclusive of off-the-point design topics. 'Lexicological Specification' works a little better, or even 'Lexicological Design Specification' but then it's becoming quite the mouthful.

Regarding my comment about the online database it struck me that if these specifications have value then it would be helpful for them to accumulate over time and be open to future analysis. To do this simply you might look at a service like Dabble DB (

Also one of your commenters raised the point about games that have built in the idea of, for example, the player being able to create ambushes and games which are just open enough that players can create this kind of tactic for themselves.

I wondered whether this could be subcategorized into games which were deliberately open ended (and hence made it easier for players to invent new tactics and modes of play) and game where this happens accidentally and whether there is any useful difference between them (or even examples).

I'm just theorizing because I have little practical experience to offer but I did wonder about the real boundaries that must (since every game I have come across is finite) exist around that openness, where those boundaries get drawn, and the impact this has.

"C. Peter also echoes an earlier suggestion (by zenBen) that this could be adapted to an ontology. Yes, it could, but I have no interest in doing this at this time. I completely agree that this could be done, and also agree that it could be useful, it just isn't necessary for what I want to use play specs for at this time."

I originally conceived of using ontologies to define game structure because ontologies can be parsed and automated by pre-existing plugin software (Reasoners, I think they're called). The point was that ontologies try to bridge the gap between semantic clarity and working (not pseudo-) code. This would then serve as the first step when implementing machine learning that can look at what the player is doing at the abstraction level of verbs and nouns, rather than position/orientation vectors, other class attributes and suchlike game engine level stuff.

Didn't actually do it yet, and I'm undecided as to whether ontologies are the right tool. But I do think that starting with a solid game spec would be wiser than implementing player-centred techniques with an ad-hoc idea of what is happening in terms of play mechanics. This, I confess, is my major interest in play specs.

I've setup a wiki page where I've collected all the specifications submitted. (in answer to someone's question about whether or not they would be collected)

The wiki provides the added bonus of letting us update them based on the discussions we've had over the past couple of days.

--"I contend: if you enjoy the basic verbs of an FPS, then the complexities in the middle ground become your entertainment and hence become relevant."--

I think this varies from person to person, as well as from game to game. I don't think you have to necessarily enjoy the "basic verbs" in order to have fun. I wouldn't say that the basic button-pushing of games like Dance Dance Revolution and Parappa are instrisically entertaining... (ok, here's a side note, what would a play spec for these games look like? Would they be remarkably similar?)

-- (edited for brevity) I feel strongly that the basic verbs are the most relevant component from an overall perspective. But I also freely acknowledge that in more complex games, the focus of play moves away from the basic verbs and into the middle ground, even though the play itself is formulated at a lower level. --

I'd like to revisit the mid-level verbs we've been discussing and propose another possible term: subjective verbs

I think that we've pretty much agreed that basic verbs arise from the allowances of the interface (coupled with their interpretation by the game code). However, mid-level verbs are subjective in that they depend on human interpretation (plus context) in order to be understood. They aren't really designed per se.

Let's talk about "Dodge". In most cases, To Dodge is to move out of the way in order to avoid a collision. However, it is subjective in the sense that if you were to analyze a movie of someone playing a game, you'd have to argue if a player dodged versus simply moved somewhere and, fortuitously, avoided a collision. In the end, without the player's input, you can only hypothesize. If the player said he dodged, that's what "verb" was used. If he hadn't noticed that stray bullet and was simply moving to the corner... he "moved".

I think that the subjective verbs are what REALLY provide the richness and depth to a gameplaying experience. Most FPS games are identical at the basic verb level, and differ only in the affordances provided by the gameworld in tandem with the basic verbs.

Pac-Man arguably has only 1 basic verb: move (in one of 4 directions)! However,it provides a very rich experience because of the gameworld (the maze) and the nouns (ghosts, pellets, power pills, fruit, spawn area), afford a host of subjective verbs that players use to describe what they do when they play Pac-Man. (eat, run away, avoid, lure ghosts to the power pill, ambush them, etc.)

The basic-verb specification of Pac-Man is pretty "dry". Listen to the Pac-Man track, starting from approximately minute 3. :-)

I think the key word here is affordance. Having stuff move and hit the player + letting the player move affords "Dodge". The faster/better the player can move, and the more stuff you have moving around, the more "Dodging" the player will probably have to do. Now, if you let the player shoot anything that moves, you will minimize the "Dodging" because the player can simply shoot down what he previously had to avoid! (unless, like in Robotron, you add so many moving objects that they can't all be shot..) And so on, and so forth.

I'd argue that Pac-Man's base verbs consist of:

Move {up | down | left | right}
Eat {Pellet | Power-Pill | Fruit}

I completely agree, except for the minor fact that a pretzel isn't a fruit. :)

I agree that the middle ground "tactical verbs" should not be included in a play spec. I think what I was trying to say was that the only time anything along the lines of tactical verbs should be in a play spec is in the case of strategy games, in which case, it would be very subjective to the individual, and it would be a very fuzzy situation. Actually, I would like to see someone do a play spec on a strategy game and see what happens. I CHALLENGE SOMEONE TO MAKE A STRATEGY GAME PLAY SPEC!!! : )

Challenge accepted! :) Now, what strategy game do I know best... How about the classic Julian Gollop game 'Chaos' (subtitled 'Magic and Death on the Plane of Limbo' on the box) on the ZX Spectrum (now available for emulation):

Chaos (SLUG, 1985)
Specified by Chris


1. View (Map, Spells, Creatures)
2. Select (Spell)
3. Cast (Spell)
4. Move (Wizard, Creature)
5. Attack (Creature)


Creatures (various)
Spells (Creature, Object, Attack, Wizard, Other)
Locations (Magic Wood, Castle, Citadel)


I have numbered the verbs as there is a strict turn sequence.

The game will take 1-8 players - computer or human. Clearly inspired by a boardgame Gollop must have designed first, this nonetheless displays all the roots of a classic turn based strategy game.

Pretzels aren't fruit? Since when?

"But if you do not enjoy the play implied by the verbs of the FPS, then the middle ground is irrelevant and never comes into play."

Reading this, I can't help but wonder at a particular counterexample : myself and Team Fortress.

I *truly* despise FPSs in general (mainly because I need to perform well in order to enjoy a game and I suck at aiming and twitching) but the original Team Fortress is 3rd in my top 3 games of all time.

Despite my hatred for FPSs, the level of planning, cooperation and strategy in TF made it poignant enought for me to play it for 5 years. No matter how I look at it, it seems to fit all possible definitions for first person shooters, even if it includes non traditionnal elements like pipebomb trapping, chasing gameplay (as a scout), disguising and defensive turret building, etc. The social component also seems to be determinant in this case cause I sure as hell wouldn't have played it that much if it haden't been against other humans.

In this case, I'm positive that I don't like much of the basic verbs (shooting, moving, capturing the flag, etc.) For me, it was all about coordinating, planning, bluffing, diverting, confusing, outwitting, etc.

I wonder what can explain this:
- The profound importance of middle verbs (thus going against the initial proposition)
- A simple anomaly in my likes/dislikes
- The very narrow definitions of FPSs and the fact that they cover games like TF which offer a lot more styles of play which are not actually shooting related.

I would suggest a probably mix of the three =)

Hi Olivier:

I think this ended up being commented in the wrong post so I moved it. This post is a year old now - I had completely forgotten the symposium and had to think about what was being said here. :)

There's no doubt that FPS games with social play like Team Fortress/Counterstrike hits at a different play need. I suspect it is not the verbs that make this more compelling, but the social interplay (which, to be fair, could be expressed in verbs).

There is, on paper at least, a vast majority of players who would rather play as part of a team (of humans) than alone, for several divergent reasons. I would estimate about 75% of the population, but probably closer to 50% of videogame players, because of the bias towards introversion.

So, I guess what I'm saying is "you are not alone". :)

I'm not even sure that the specific verbs matter in this context, but I'd need to study the phenomena closer to find out.

Best wishes!

Verify your Comment

Previewing your Comment

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

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.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)