Only a Game

Play Specifications

Play specifications are a means of examining the core play of a game. Abstract, subjective and informal, they can be used to analyse the play of an existing game, or to specify the play of a new game concept. Perhaps an even more useful aspect is their potential to teach new game designers a basic game design method.

This piece looks specifically at a lexical play specification, although numerous other approaches are possible.

Roots

This post follows from the following previous posts:

• A Game Design Grammar, in which I explored the application of categorial grammars to game design, and thus devised what I am now terming a play specification. Some of the text here is raided from this post.
• Grammatical? Lexical? Functional? in which I debunk my use of categorial grammar by observing that I made modifications to the calculus which rendered it ineffective.
• Language Games contains the philosophical roots of the approach, drawing from the work of Ludwig Wittgenstein on the philosophy of language.

What is a play specification?

Any form of play specification focusses on expressing the elements of the play of a game in a simplified form. It is thus a kind of abstraction. It is also a form of notation, in that it sets out a method for recording (and thus communicating) the key elements of play. The goal of the approach is to provide a means both of expressing the play of an existing game in a form suitable for discussion and criticism, and also to provide a method for guiding, exploring and teaching the game design process.

Although it is conceivable to define highly formal systems for play specification, I do not personally believe this will be helpful - although I cannot dismiss the idea that it might prove useful. Play specifications can be informally constructed, and still yield considerable value.

I also do not believe in the inherent superiority of atomist or reductionist approaches to the analysis of game design. This is not the same as suggesting that these approaches have no value - different people have different focusses; different methods produce different results - but I believe attempts to objectively identify the atoms of play must eventually face insurmountable problems. Play is a subjective experience. Psychology accepts subjectivity in its remit, and any complete analysis of play would perhaps be wise to do the same.

Lexical Play Specification

The notion of a lexical play specification is to express the play of an arbitrary game in terms of lexical elements (or keywords), that is, through the creation of a miniature lexicon that describes the play of the game. The specific play specification I am presenting here has descended from my categorial grammar of game design (and the specific notation I referred to as Bast).

The short form of this specification is as follows:

The play of any game can be considered in terms of the Nouns, Verbs and Adjectives that can be used to describe it. Nouns refer to the game entities, Verbs to the actions that can be taken, and Adjectives refer to the properties of entities and actions.

The primary elements are the Nouns and Verbs. To understand what is meant by this, consider the following examples of expressing the Nouns and Verbs of certain games:

• In Snakes & Ladders (Chutes and Ladders in the US - what's the matter with the US; are they afraid of snakes?) the Nouns are the Counters, the Board, Ladders, Snakes and the Die. The Verbs are Roll (the Die), Move (Counters along the Board), Climb (a Ladder) and Slide (down a Snake).
• In Magic: The Gathering the Nouns are Permanents (Creatures, Enchantments, Artifacts and Land) and Instants/Sorceries and the Verbs are Lay (a Permanent), Tap (a Permanent), Attack and Play (an Instant/Sorcery). There are two additional Nouns - Mana and Life - which are special types of enumerated Nouns we can consider Resources.
• In an arbitrary basic racing video game, the Nouns are Cars and Courses, the Verbs are Steer (a car), Accelerate (a car) and Brake (a car).
• In the game of Sink (played by Discordians, and people of much ilk) the Nouns are The Float, the water, and the Junk. The Verb is Throw (the Float into the water or the Junk at the Float).
• In a tabletop RPG (in a general case), the Nouns are infinite and the Verbs are infinite. I knew there was a reason I used to enjoy playing them!

In terms of analysing games, the advantage of this model is that it focuses on the game objects or entities (the Nouns) and the player actions (the Verbs). Notice that it does not express elements that some people would consider essential to a game - such as a goal state. I consider this quite healthy. The most interesting thing about a game is its play, not its goals - and indeed, 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. It would be trivial to expand this specification to include goals, if so desired.

Also note that in videogames we have an extra player - the Computer - who may have its own set of Verbs, but that we need not specify this hidden layer unless we feel specific motivation to do so.

In terms of teaching someone how to design a game, the lexical play specification approach allows a game to be judiciously constructed from components - although wisdom and intelligence are of course needed to do this effectively. For instance, if we want to make a card game, we know one of our Nouns is (a) Card, and that all or most of the other Nouns will be comprised of multiple cards - these Nouns will often include a Deck, Hands and a Discard Pile. From there, we just need to consider the actions the player can take - the Verbs of the game.

When we are dealing with games more complicated than card games (and I advocate all new game designers should begin by learning to make card games) we need to add a few new categories to our specification. We need to define the properties of Nouns (and potentially of Verbs) - that is, we need Adjectives (or Adverbs). Since the grammatical distinction between an adjective and an adverb is largely irrelevant to our play specification, I suggest we only need to employ the term Adjective to describe the attributes and parameters of a particular game.

Examples of Adjectives in games:

• In Snakes and Ladders and in Sink there are no Adjectives.
• In Magic: The Gathering, the Adjectives include Tap Effect, Power, Toughness, Counters, and Casting Cost.
• In an arbitrary basic racing game, the Adjectives include Top Speed, Acceleration and Traction (Super Sprint, anyone?)
• In a tabletop RPG, the Adjectives describe the mechanics of the game; Attributes, Skills etc.

Although I have been using something akin to these lexical game design principles for many years now, it is only very recently that it occurred to me to employ them in videogame design. The inspiration for doing so was listening to Toru Iwatani and Keita Takahashi talk about game design - they were focused on verbs, because verbs describe what the player actually does.  (As Ernest is fond of saying: "Yes, but what does the player actually do?").

Applications

There are three key areas of application for play specifications:

1. Critical Analysis of Play: as a tool for the analysis of the play of a particular game, I consider play specifications to have considerable worth. We lack a critical tradition in games (game reviews are intended to guide purchasing decisions, they are not part of the critical tradition); play specifications provide one route to such an approach. Consider this analysis of the Katamari games - but note that reference to grammar in this piece refer to play specifications. (See the roots of this piece for an explanation of the disjunct terminology).
2. Focusing the Game Design Process: I suggest that beginning the game design process with a play specification allows the game design to focus on what the player actually does in the game (the verbs), as well as ensuring that the developer has a better idea of the scope of the implentation requirements for the design. The more Nouns and Verbs that are included in the specification, the more time consuming (i.e. expensive) the game will be to develop.
3. Teaching Game Design: we currently seem to lack a starting point to introduce new game designers to the process of game design. Indeed, although many people have written on game design, we still haven't developed identifiable schools of game design, perhaps because much of the discussion thus far has been too vague; too discursive. Play specifications can demonstrate to a new game designer where game design can begin. (Since many students of game design confuse the narrative of a game with its design, this could be a significantly positive step forward).

I encourage other people to experiment with this lexical play specification technique and, if they find the need, to develop new play specification notations. It is possible that the approach I provide here is general enough for there to be no need for alternative notations; it is too soon to tell. I certainly don't advocate creating alternative notations for the sake of it.

Please feel free to contribute to this approach by analysing games in terms of their Verbs, Nouns and Adjectives (remembering always that there can be any number of different ways of doing so!), by experimenting with beginning the design process with a play specification, and perhaps also by teaching students of game design this technique as a stepping point to understanding the complexities and subtleties of this unique and fascinating field.

With especial thanks to the regulars here at Only a Game who have encouraged me and helped clarify this approach, in particular Ben Cowley, Jack Monahan, Chico Queiroz and of course Patrick Dugan.

The opening image is Glass of Beer and Playing Cards, by Juan Gris, which I found here. As ever, no copyright infringement is intended, and I will take the image down if asked.

Nintendo Urine

I'm sure you've all seen by now that the new name for the Nintendo Revolution is the Nintendo Wii. I have never in my life wanted to spend much time playing with wee, personally. Still, a rose by any other name...

Something Akin to a Game Design Dictionary

I'm pleased to announce that Jose Zagel dropped by and let me know about a project which is very close to the 'game design dictionary' I discussed in a previous post. Its called the Game Ontology Project, and is a wiki-driven collaborative approach. I'm thrilled that this exists, and want to encourage anyone with the time and skills to contribute.

I want to support this project, but it will be unfeasible for me to do so directly. The sad truth is that I run a company, work as a game designer and co-run the IGDA's Game Writers' SIG - I just don't have the time to contribute directly to such a project. I have to think of ways I can contribute tangentially instead.

I'm going to raise some criticisms here; I hope it is understood that I am in full support of the project and that my goal here is to contribute to refinement, not to undermine the project:

• Hierarchical: I'm glad that this is not a taxonomic project, but I'm disappointed that it's hierarchical. I often come down hard on taxonomic and hierarchical approaches because I personally don't consider them helpful - it's easy for people to get distracted by the minutae of a scheme and lose the overall value. Perhaps, however, this reflects that my interest in such a project is in its lexicographical role, and this is incompatible with a hierarchical approach. Or, to put it another way, the result of the Game Ontology project may serve to inform a future lexicographical approach but it cannot replace it because hierarchical issues stand in direct tension to lexicographic needs. I welcome a defence of the hierarchical elements though!
• Goals: having spoken out against the hierarchical issues, watch me now get distracted by one... One can place Goals under Rules in a hierarchical sense as goals are a form of rule. The fact that you haven't is a positive step to my mind (because at a psychological level goals have different implications), but it seems contrary to the hierarchical process.
• Entity Manipulation: this is just a horrible phrase to my eyes and ears! The other three branches of the ontology have a pleasing neatness; this one screams 'work in progress'. "Entity manipulation consists of altering the attributes or abilities of an entity in the game world" it begins. I do not understand why one would value the alteration of attributes and abilities over the definition of those attributes, abilities and entities! Although much progress-driven play consists of the acquisition of new abilities, the definitions of play originate in the abilities themselves, c.f. Shadow of the Colossus which begins the player with all their abilities and attributes. Putting aside which word you choose for the nouns of the game, would not "Entities" be more than sufficient for the high level of this branch of the ontology? The attributes and abilities extend from the entities, after all.

Incidentally, the use of the term 'ontology' may be borrowed from computer science, but it behooves me to point out that ontology is a branch of philosophy concerned with the metaphysics of being. In fact, often when people use the term 'existential' it would be perhaps be more precise to say 'ontological'.

Linking to the project from my sidebar. Thanks again to Jose for bringing it to my attention!

PS: Any chance of getting updates to the wiki fed to an RSS feed? It would make it much easier for people to follow the project's progress!

Random Thoughts

1. How are we ever to escape the problem of game designers subconsciously or otherwise making the games they want to play, not the games the audience would enjoy playing? (Accepting, of course, that there will always be a place for people making the games they want for themselves).
2. If EA's turnover is 3 billion dollars per year, why can't they spare a million dollars each year to distribute to creative artists trying to make games? (Or do they, but I'm unaware of the scheme?)
3. Does Oblivion achieve so little that is new or interesting because the developer had low goals, or because they were unable to fulfill the high goals they set themselves? (I'm sure the game delivers an existing form of play competently however).
4. Is World of Warcraft impressive? In what sense? (I contend that at the design level it is rather unremarkable, but this is not the only judge of a game, of course).
5. Is game design theory more valuable than philosophy of games? Does either discipline recognisably exist yet? (I began in science, I feel fated to end up in philosophy if I live long enough. Perhaps I am already camped on the border).

DGD2: Aphorisms

While mulling over the skill sets of Temperament Theory, I came up with the following aphorisms to summarise the four player types of "DGD 1.5" (i.e. Temperament Theory applied to games):

Logistical
"Practice makes perfect"

Tactical
"Give me the tools and let me at it."

Strategic
"I can figure it out"

Diplomatic
"We can find a way."

(The Diplomatic skill set remains the least clear in terms of games, since so much focus elsewhere has been on the Idealist's interpersonal skills which don't apply in single player games. Integrative thinking and the ability to deal with abstracts seem key, however.)

Compass Reading

This week will be punctuated by a trip to London on Wednesday... I can't say anything about the project involved yet, but it's one of our concept designs coming home to roost. I'll say more when I'm in a position to do so. Anyway, blogging will be disrupted.

Still waiting for the new Fireball build. It's close - I can feel it - but just out of reach. Will post on the state of the project once we have it.

Also expected any day now, the copyedited manuscript for Game Writing : Narrative Skills for Videogames. I'm keen to get this book completed and out the door.

I'm now in a position to write my analysis of Shadow of the Colossus if I want to. Still only a quarter of the way into the game, but I don't believe the design has any surprises now; it's quite tightly constructed. This means I have to make a firm terminological decision. Previously we talked of 'keywords of design' versus 'lexical analysis'... I'm also considering 'playwords' and 'play specification'. I'm leaning towards 'Play Specification' - after all, the technique can be used to examine the play of an existing game, or to specify the play of a new game. Still thinking...

Been thinking about revisions, and whether to revise old posts, especially those that are part of a set. For instance, that ilinx piece is very out of date now.

Oh, and we have some new data. Yes, quite exciting this... It needs processing, but I should be able to provide some generalised player profiles for specific games in the next month or so. Don't know if we have enough data to make a DGD2 model and move onto DGD3, but it's certainly possible.

Anybody 'waiting' for anything, or wanting more on a topic touched in passing, just let me know. No promises, but I'm maleable.

Alright, duty calls.

Language Games

What does a word mean? How are we able to understand language? Why does so much confusion and argumentation result from apparently such gentle variation in the way individuals employ their words? To explore these issues, we can turn to the philosophical notion of language games.

Ludwig Wittgenstein (1889-1951) was one of the most influential philosophers of the twentieth century, contributing significantly to numerous areas but perhaps most importantly to the philosophy of language. Wittgenstein was the source of the idea that “the meaning of a word is how it is used” or “meaning is use”. But to fully understand what is meant by this phrase, it is helpful to explore his concept of a language game.

Wittgenstein’s magnum opus, Philosophical Investigations, commences in a manner which can be a source of great surprise to the unprepared. It begins with concise descriptions of situations involving the usage of words. The first involves the purchase of five red apples; the second describes two people using language at a construction site:

We shall imagine a language… Let this language serve the purpose of communication between a builder A and his assistant B. A is constructing a building out of building stones: blocks, columns, slabs, and beams. B is to bring the building stones, in the order A needs them. For that purpose they use a language consisting of the words “block”, “column”, “slab”, “beam”. A calls out the words and B brings the stone he has learned to bring at this call… Think of this as a complete, primitive language.

Wittgenstein calls “a complete primitive language” a language game (sprachspiel in his native German). The concept is introduced and developed in a set of notes which he circulated privately between his students and others in the 1930s commonly known as The Blue Book and The Brown Book – which in many ways are a lot like reading Wittgenstein’s blog (although of course no such concept existed at the time). The Brown Book in particular is intimately connected with the notion of a language game, and consists of a series of numbered descriptions of particular language games, accompanied by rambling commentaries in which it is hard to determine if Wittgenstein is attempting to explain his reasoning to a reader, or to elucidate his meaning for his own benefit.

Key to the way this is presented is that each language game is described in terms of how it is taught, and that we are invited to consider for each language game that it might exist in isolation – imagine a hypothetical tribe whose only means of communication is the “primitive language” of any given game, like the builder’s game above. As it says in The Brown Book:

Let us imagine a society in which this is the only system of language. The child learns this language from the grown-ups by being trained to its use… Part of this training is that we point to a building stone, direct the attention of the child towards it, and pronounce a word. In the actual use of this language, one man calls out the words as orders, the other acts according to them.”

Why introduce this peculiar concept of a language game? Part of the purpose is to stress how language is learned in the first place; the focus is on how children learn to speak for the first time. This, after all, is how language is acquired. Additionally, the training situations which are possible depend upon inborn capabilities. In the builder’s game, the capacity to carry rocks (which would not, for instance, be possible for a snake or a fish since they have no capacity to lift) is essential to the existence of the game and hence of the words involved. The emphasis in the use of language games is always towards the practical: “The term ‘language game’ is meant to emphasise that the speaking of a language is part of an activity or a form of life.”

It is easy to see, therefore, how by examining the use of language in these terms leads to a different understanding of the process of language in toto:

The whole, the language and the activities with which it is interwoven: this I will also call the “language game.”

Wittgenstein believed that the purpose of philosophy was clarity; he speaks often of muddles and befuddlement, and of philosophy aiming to remove these confusions. In formulating the concept of language games, he was able to explore how many philosophical problems originate in misunderstandings as to the meaning of words.

Following the concept of a language game, the purpose of a word is the use that it serves within its language game. He cogently examines a number of different philosophical problems in terms of the words used, and frequently demonstrates that often the problem is that a word is used in two separate language games, but appears to the individual to be the same word because the word (as a symbol; it’s letters or phonemes) is unchanged. But the word only has meaning in the context of its language game, and so if the word is used in a different context, its meaning need not be the same. (Sadly, examples to this effect are difficult to précis efficiently. The interested reader should turn to Wittgenstein’s work to explore this).

Taking this idea further, he demonstrates how we can use words in contexts that extend far beyond their original usages. For example, the terms ‘darker’ and ‘lighter’ belong in a language game of comparison of shade, associated with instructions such as “paint me a patch of colour darker than the one I am showing you.” But we could say: “Listen to the five vowels a, e, i, o and u and arrange them in order of darkness.” Many people will look puzzled and be unable to reply, but some will respond by arranging the vowels in a certain order (amazingly, this order tends to be i, e, a, o, u – which was, incidentally, my wife’s immediate response when I asked her this paradoxical question). This is a curious development! But it shows our tremendous capacity to apply words taught in a specific context to other settings which may have little or no logical connection with the original game. Given that we can have such an apparently nonsensical question answered at all, it is hopefully evident that any notion that words have objective Platonic meanings is at least problematic, and at worst beyond fallacious.

The more different language games are combined in any given culture (as language and culture are inextricably intertwined), the greater the potential for confusion. The word ‘God’ has meaning principally in a religious or spiritual language game, the word ‘True’ has meaning only in a logical language game, and the word ‘Theory’ has meaning chiefly in a scientific language game. But in everyday life, we use all of these words together in one grand system combining all the language games we have ever learned. Small wonder that careless use of words leads to bewilderment and conflict.

There is a further source of potential perplexity in the way words are employed. The fixed nature of a written word gives it the illusion of stability and invites us to consider all words as being equivalent in function. But words do not have such an equivalence; they serve very different functions. Wittgenstein suggests: “Think of the tools in a tool chest. There is a hammer, pliers, a saw, a screwdriver, a ruler, a glue jar, glue, nails, and screws…. The function of words is as varied as the functions of these objects. (And in both cases there are similarities).”

This leads to the idea of words being defined not in terms of idealised eternal concepts, but rather acting in a manner akin to family resemblance:

Consider, for example, the proceedings we call “games”. I mean board games, card games, ball games, athletic games, etc. What is common to all of these? – Don’t say: “They have to have something in common or they would not be called games” – but rather look and see if there is something common to all of them… you will see similarities and relationships – a whole series of them… We see a complicated network of similarities intersecting and overlapping one another – similarities large and small. I can give no better characterization of these similarities than “family resemblances”; for it is in just this way that the various resemblances to be found among members of a family overlap and intersect: build, facial features, eye colour, gait, temperament, etc., etc. – And I will say: ‘games’ form a family.

Language games are similarly related to one another by family resemblance, and as with other games, rules are of some significance in the formulation of many (but not all) language games. Consider how the builder’s game is predicated upon the rules which define the names of the stones: blocks, columns, slabs, and beams. The naming of things, Wittgenstein suggests, “is no move in the language game”. When one sets up the pieces on the board for a game of chess, this is not a move in the game but rather the preparation for the playing of the game. Similarly, the naming of things is not a move in a language game, but rather part of the preparation for the play of that language game.

Only by understanding what was meant by a language game are we ready to understand Wittgenstein’s most famous dictum:

For a large class of cases of the use of the word “meaning” – even if not for all cases – the word can be explained thus: The meaning of a word is its use in language.

Words are tools for communication, and these symbols have purpose only in the context in which they are employed. In discussing our choices of words and their definitions, we are serving to delineate the language games in which these signs are employed, and thus avoiding any future misunderstanding in their application. Perhaps this too is a language game of a particular sort, as it is an activity with a purpose, and our words adapt to this process as they would to any other.

The temptation to interpret words as having “true meanings” is untenable in this context. Truth is a concept from the language games of logic, and difficult or even impossible to apply to our everyday existence. A word means solely what it implies to those who use it, and what it means to them is encapsulated in how it is used. By understanding words as the pieces of a language game, we escape all manner of confusing problems which have haunted us since antiquity, and move forward into a world where our discourses can attain greater clarity. For this, we have Wittgenstein to thank.

Toru Iwatani's Escalator

A game should possess a clarity of purpose. If it is not clear what the player should be doing, the game has an insurmountable problem. Even games in which you are given a completely open hand only work if this freedom is successfully conveyed to the player. Even games in which the core play is devising what to do only work if the player is sufficiently confident that this is what they are supposed to be doing. Does your game have a clarity of purpose?

When I attended the talk by Pac-man's creator back at GDC 2004, I wasn't expecting the session to develop an almost mystical quality in me. Japanese is an unusual language in that it prefers implication to statement - a quality that often makes it rather hard to translate into the languages of more literal cultures. However, it also meant that Iwatani-san's presentation was not a carefully structured thesis, but a codex of personal experience centred on the design process behind a hugely popular and successful game. It was the most moving experience I've had at any GDC, but it took me some time to completely internalise it.

The image of the escalator was perhaps the most cryptic element of the talk. Iwatani-san asked the audience for their thoughts on what made the escalator a "perfect system"; this audience participation further obfuscated his purpose in choosing this icon. Indeed, at least one report (unless my memory fails me - which it does from time to time) erroneously supplies my suggestion (voiced on the day) as to the strength of the escalator: that even when broken it is still functional, as a broken escalator is a staircase. Only some time afterwards did the gentle zen like effect of the talk finally settle into my mind, and I successfully decoded Iwatani-san's intended message.

The beauty of the design of the escalator is that anyone can immediately see how to use it. No prior experience, no specialist skills, no instructions are necessary.  One can tell from looking at it what is expected. This was the same concept that motivated the design of Pac-man: clarity of purpose. Does your game have clarity of purpose?

Play Engineer

Would the title 'play engineer' be a better description for the role of game designer than 'game designer' (which is now so inflated as to be devalued)? Discuss, if at all interested.

My Equine Friends

Together, we bounded across the surprisingly tiny kingdom of Hyrule. She and I were the best of friends, in fact, she was my closest friend. When I was a child, I played a melody to her, and we connected in a way that transcended time. As an adult, she was my constant companion. I hated going underground into the depths and darkness, as I had to leave my friend behind. Fortunately, wherever in the world I went, I just had to play that simple melody from our mutual past and she would be there by my side. In another world, her name 'Epona', was the goddess of horses - such a powerful symbol of fertility and equine power to the Gauls that the Roman empire adopted her as part of their pantheon. She was the first of my equine friends, and still one of the greatest. I cannot wait to see her again.

In my short life as Barbie, I struggled privately with my anorexia nervosa; but I had many friends to keep me company, and together we rode through beautiful landscapes. Even if I couldn't eat, I could at least give sugar lumps and apples to my many animal friends. My stables were full of horses, although sadly this reduced my connection with any individual horse - not to mention the macabre genetic manipulations that went on to allow me to change the appearance of each horse. But all of this pales into insignificance next to the joy of riding at full canter through green fields, surrounded by butterflies and blossom, and the simple satisfaction of lining up a jump perfectly, or ducking to avoid a low tree stump. A shame that dark unseen forces dictated the need to add more 'challenge' to this game (the game developer is blameless in this), as its riding experience was quite exquisite.

And what of that incomparable destrier, Red Hare? A fire-red mare, so perfectly attuned to its rider's wishers that it would stop dead upon seeing a hare, allowing the hunter upon its back to take a perfect shot with a bow. It once belonged to the most powerful warrior of its era - the mighty Lu Bu - but despite his power he could not be trusted, and this ultimately lead to his demise. Consequently, the horse entered into the stables of the powerful warlord Cao Cao, the dominant power in the Three Kingdoms at the time. It finally came into the possession of  the legendary Guan Yu (pictured) during the time he served in Cao Cao's forces while seeking his missing oath brothers, the dutiful Liu Bei and hot-headed pig butcher Zhang Fei. When I ride Red Hare, this incomparable war horse, no one can match our speed, and the enemy forces scatter and tumble as we gallop towards our next opponent, the taste of battle burning our lungs.

Finally, we come to willful Agro. He too is my friend, for he seems to know better than I that the path upon which I travel can lead only to destruction and doom. He resists my commands, pulling at the reins and champing at the bit. Whereas other horses have obeyed me without question, Agro would like nothing more than for me to give up my insane quest and instead just ride out over the land, feeling the wind in my hair and in his mane... But this can never be. Despite his obstinance, he is also faithful and proficient. When a narrow gorge must be crossed by a stone path few horses would have the constitution to attempt to traverse, Agro sets a gentle pace and trots across confidently. Yet this is no mere riding horse, for he holds perfectly still when I take bow in hand, and a simple whistle calls him to my aid. Loyal yet rebellious in his determination to persuade me to abandon the course of self-destruction, perhaps he is the truest friend of all.