Fundamentally, how were you supposed to play the original Dungeons & Dragons? What form did the participation of players take? The text published in 1974 gives us no shortage of rules. With imposing pages of charts and exposition, the books illustrate how to generate characters and underworlds and how to dice against tables to derive random encounters, decide combat results, and disburse hard-won plunder—but they provide strikingly little instruction on how to integrate those activities into the moment-to-moment play of a game.
In a traditional board wargame, say, players take turns. A rulebook for the game would explain how turns break down into phases of unit movement and combat. Some games specify that a turn represents a specific interval of game time, which in strategic wargames might be months—each Diplomacy turn represents half a year—but in tactical wargames can equate to minutes or, as in Korns, even just a few seconds. But the D&D rules do not instruct players to take turns; the game retains the word turn, but defines it solely as a measure of time for deriving movement rates and similar calculations salient to a tactical situation. In the absence of turn sequences and without the familiar constraints of boards or pieces, it was not at all obvious from poring through the books what the participants in a D&D game actually were supposed to do. As Ted Johnstone put it at the dawn of the hobby, “It requires comprehension of lots of scattered elements throughout the rules to begin to get even a vague idea of how the D&D game is structured” (APL 511).
At the most basic level, players participate in the game of D&D by talking to the referee. In lieu of any overview of its operation, D&D tries to teach by example, through a sample transcript of a dungeon adventure that records a spoken exchange between a referee and a “caller”—albeit without anywhere defining what a “caller” is. The two hold a conversation in which they might be said to take turns making statements: the caller proposes the actions of a party comprising several characters, and then the referee describes the results of attempting those actions. This structure has obvious precedents in both of the two cultures—free Kriegsspiel and collaborative authorship—but the presentation of a transcript in lieu of a turn sequence borrows most directly from Korns’s Modern War in Miniature, which indeed opens with just such a transcript rather than stashing it toward the back of the rules, as D&D does.
When daily newspapers breathlessly transcribed the Fischer–Spassky chess matches in 1972, each reported move took up the same amount of chess notation. But unlike the turns taken by traditional game adversaries, there is little consistency or parity in the length or composition of the spoken “moves” in D&D: as in any conversation, the utterances of each side are as long and complex as necessary to serve the speaker’s purpose. During a period of exploration, the referee in the transcript rattles off movement over time in staccato 10-foot increments and enumerates potential directions the party might explore, while the caller navigates with curt and direct instructions such as “Go south.” Both speakers phrase their statements as contributions to a common story, as if they are taking turns adding sentences to a fictional work in progress—in the course of the transcript, neither challenges the other’s authority to make any utterance. The most obvious difference in the nature of their statements is the pronouns: the referee throughout addresses the party in the second-person, while the caller generally describes the actions of the party in the first-person plural. For example, after the referee informs the caller that “there is a door to your left across the passage on a northwest wall,” the caller submits the action: “Listen at the door—three of us.”1
The conversation transcript in D&D is punctuated by die rolls, which seem to regulate both the information the referee shares with the party and the flow of events. When the party listens at a door, the dice induce the referee to deliver terse reports such as “You hear nothing” or “You hear shuffling.” At certain intervals, die rolls also determine if “wandering monsters” in the underworld confront the party. When the party bursts in on a group of gnolls, the transcript glosses over the die rolls of combat with the simple placeholder “melee conducted” in order to focus on the victorious party’s subsequent efforts to collect the vanquished foes’ treasure—and it is here that the true power of a dialogue becomes apparent. By way of ascertaining what the room looks like, the caller sets the party to work with statements such as “We’re examining the walls, ceiling, floor, and contents of the room itself.” The referee then provides a detailed description of the gnolls’ lair, and through a process of steady inquiry into the details of the space, all phrased as specific actions such as “Each trunk will be opened by one of us” and then “Check the trunk for secret drawers or a false bottom,” the intrepid caller uncovers hidden plunder.
This example of D&D play illustrates something that its designers apparently found so obvious that it went without saying—that the game takes place in a dialogue. This was no doubt novel to many who chanced on the rules. But the roots of these practices in the legacy of wargaming meant that some would recognize their source: Jack Harness, one of the early adopters in science-fiction fandom in Los Angeles, noted that “the Dungeonmaster is sort of a cross between a Kriegspiel referee . . . and God and His Chosen People.” Watching Lee Gold run D&D, he was perhaps reminded of Kriegsspiel because “Lee rolled a die repeatedly and observed it before answering questions sometimes.”2 Twin Cities gamers in Arneson’s circle had exhumed the Kriegsspiel principles of the nineteenth century from Totten’s Strategos; in 1880, Totten knew well that that the fundamental structure of play would not be obvious without an explicit description of the interaction between player and referee:
The Referee, therefore, should generally require a positive statement of intention, as the basis of his decision; the attempt must be willed into operation by the player. It is not until then that the Referee may properly exercise his functions. He may then duly consider all the pros and cons. Losses, Tactical, Strategical, Topographical, and Accidental Difficulties etc. must be calculated and examined, and, the crucial moment having in due time arrived, as indicated by the circumstances of the particular case, he should make his decision, and, if desirable, state his reason, which, however, etiquette must protect from dispute.3
In this Strategos passage, we find something of an informal sequence for each effective “turn” in a Kriegsspiel dialogue: a direct account of how play is structured, something that D&D sorely lacked. A Strategos player must first furnish to the referee “a positive statement of intention” through which an attempted action is “willed into operation by the player.” With this statement of intention, the referee can proceed to the second phase of the turn by consulting the system, including any necessary calculations and die rolls. Finally, in the third phase, the referee decides and reports the result of the intended action, which the referee may justify with a rationale that the rules “protect from dispute”—the referee’s decision may not be contested. Based on the resulting circumstances, the player must then at the start of the next turn formulate a new “statement of intention” for the referee, and so the cycle continues for the duration of play.
Games in the Kriegsspiel tradition allow players a qualitatively different sort of agency than they enjoy in a traditional board wargame or a game like chess.4 A chess player who lifts pieces and moves them across a board exercises direct and highly conspicuous control over game events—even if the consequences of a given move may not be easily predictable. But when Reiswitz transposed moves into written orders and then Verdy du Vernois further adapted them into conversational snippets, a layer of mediation formed between players and their actions. Now a referee must parse and interpret the verbalized intentions of the player, reporting results that may be surprising in a different way than a simple chess blunder. For the Kriegsspiel referee, this power of interpretation brings the opportunity to construe vague orders in ways the commander never intended, which may lead to disastrous consequences on the imagined battlefield and a lesson hopefully learned.
But this combination of dialogue and interpretation also brought the opportunity for tremendous versatility: when compared to the rigidity of a traditional move–countermove board game like chess, it can hardly be overstated. Anything the referee can describe verbally can become an element of the game, and anything a player can articulate as a statement of intention can potentially translate into an action. This innovation proved useful enough that it influenced many later works with little connection to the original Kriegsspiel tradition. We hear an echo of it even in the way Patterson described his postal Midgard game in January 1971, in describing the responsibility of the “umpire or gamesmaster, to whom the players send their moves and from whom by return they learn the results of their actions.”
Hobby wargamers of the 1970s not directly acquainted with Prussian Kriegsspiel literature might still have recognized gameplay with this structure thanks to Korns. In an early wargaming fanzine review of D&D, Arnold Hendrick could in 1974 read the turn structure between the lines, describing the play of the game as follows: “The referee is informed of each action, and after consulting the maps he has made, the basic tables and information in the booklets, and his own imagination, gives the player a response” (CO 6 (6)). Hendrick predicted that “those who remember Korns’s Modern War in Miniature will see the parallel” to D&D. Korns’s referees served to isolate players in the situation of the game, as his referees were “the only ones who need to be familiar with the rules,” and so too can a D&D referee let players state intentions without any insight into how the system would be executed.
The Cambridge University student Sandy Eisen described in the beginning of 1975 his own introduction to D&D: “I found the first few games intensely enjoyable and exciting; I really lived the part and my ‘willing suspension of disbelief’ found myself there—in the dungeon. My actions (and of course my thoughts about these actions) were dictated by real-life considerations and no thought of wargame mechanics entered my head to distract me from the ‘events’ going on” (EU 6–8). But, for Eisen, eventually learning the game system was a source of acute disappointment. “Inevitably when you are aware of the rules, you play out each situation with an eye to obtaining best odds/chances for survival, etc., considering the rules rather than the situation you are in.” In fact, Eisen felt that it impaired his experience of the game so much that he vowed, “To avoid this I have decided that when I design and run my own dungeon I will not permit the players (people who do not know about D&D yet) to discover the rules. Of course this will put them at a great disadvantage, and I feel I may have to put over quite a bit of information in the form of legend/folklore/tales so that they will have some idea of what they are up against and what to try, but all without disclosing the game mechanics.”
In some respects, Eisen’s vow falls in line with Korns’s thinking about “isolating” players from the rules; that Eisen recognized this isolation put players at a disadvantage is perhaps a clue to how players could find certain referee styles adversarial. But where Eisen departed from wargame precedents was in the purpose of this deprivation, the property that he hoped to recover: his feeling as a novice player that he “lived the part” and, in some sense, that he found himself through suspension of disbelief “in the dungeon.” It was the power of the dialogue with the referee that made it possible for Eisen’s actions to be solely “dictated by real-life considerations” with “no thought of wargame mechanics.” Nor was Eisen the only one in 1975 who valued “living the part,” a sort of theatrical understanding of D&D, as Jack Harness could already attest: “The play’s the thing, not the winning of the battle. It’s impromptu improvisational theater, where all the audience are players, including the Dungeonmaster” (EM 21). To understand how a game can deliver this feeling, we need to explore the sort of agency that players have when they play D&D.
It took a little time for games in the D&D tradition to incorporate rule language explaining how the dialogue worked. In TSR games, teaching by example remained for some time the preferred method of introducing the fundamental structure of the game: the company’s follow-up, Empire of the Petal Throne (1975), also furnishes a sample dialogue, one that gives more detail on combat and event resolution. Metamorphosis Alpha (1976) provides its own “Example of a Referee Moderating an Adventure,” which further clarifies some of the fundamentals of the process. It gives a long-overdue definition of the “caller” as “the player representing the group.” Although callers pronounce the statements of intention, we learn from Metamorphosis Alpha that these calls are informed by group discussion; in a few cases, “the caller momentarily consults with the other players on what the group should do” until “the group has reached a consensus.” When the system so requires, the referee consults the dice: “after he finished the rolls, he announces the results—or the results discernible to the players.”5 That final caveat recalls how Korns envisioned the role of the referee in 1966, where players “know only what the judge tells them that their [characters] can see or hear” instead of the imperceptible causes of those ostensions.
Outside of TSR, published games building on D&D incorporated the dialogue more explicitly into the rules. Monsters! Monsters! (1976) gives something like a formal turn sequence, and in the first phase of it “the characters tell the game master what direction they go, what actions they take, etc. As they progress, the GM tells them what they see, hear, or otherwise sense.” In this phase “the characters may question the GM if they want more detail.”6 The participants may exchange several such statements before actions that constitute a full turn of time have been taken; at the end of that interval, the referee executes a number of system functions that round out the rest of the phases, such as checking for wandering enemies and applying any healing effects. Bunnies & Burrows (1976) simply and succinctly instructs the referee to start a game session by kicking off the dialogue: “Tell them where they are to begin with. From then on, the players tell you what they want to do and you tell them the results of their action.”7
Sometimes, for various reasons, players do not get to do what they want. Although Totten famously promised in Strategos that “anything can be attempted,” he did not mean everything should be. He qualified that dictum in the same paragraph with the principle that “the advisability of an attempt is another thing, and one that it is the object of the War Game to make evident to all concerned by the results.”8 Any statement of intention might turn out to have been ill advised, and Totten stressed that the referee must make players experience the consequences of misguided or misstated intentions. His wargame first and foremost served as a training tool for soldiers, and as such its referee had a corrective responsibility, one that might sometimes prove difficult to distinguish from a game opponent’s antagonism. The referee of Strategos is perhaps best thought of as a teacher: although a teacher can adopt an adversarial posture toward students—and vice versa—both are nominally participating in a collaborative process. The referee, by establishing the general situation of a wargame, constructs a sort of test and then judges players’ performance on a moment-to-moment basis through their statements of intention.
D&D inherited something of this didactic responsibility from the wargames it imitated. Designing a dungeon has a certain kinship to designing a classroom test, in the sense that Kevin Slimak surely meant when he said that a “dungeon designer sets the problems for his adventurers and they try to solve them.” An adversarial teacher can always devise an unfair test that students will surely fail, and, similarly, some referees take pride in devising tests they know more than half of players will flunk. The D&D guidance on dungeon design cautions, “There is no question that a player’s character could easily be killed by falling into a pit thirty feet deep or into a shallow pit filled with poisonous spikes, and this is quite undesirable in most instances,” advice that the designer of the “Tomb of Horrors” might well have heeded. The rules instead recommend specifying “as many mystifying and dangerous areas as is consistent with a reasonable chance for survival,”9 just as a teacher amenable to the edification of students will design tests that challenge pupils yet still provide opportunities for the worthy to excel.
Because D&D transpires in a conversation, assessing a player’s worthiness hinges on statements of intention. Figuring out what to say and how to say it is fundamentally the measure of a player’s skill: a canny referee handles a statement of intention based on what a player states rather than on what a player might have tacitly intended. Nothing better exemplifies this adversarial fidelity to player statements in D&D than the treatment of wishes. A magical wish presents players with a rare opportunity to share in the referee’s world-shaping authority, but wishes also have a proverbial tendency to backfire. The wish rules in D&D place the entire burden for articulating the intended outcome on the player, clearly directing the referee to exploit ambiguities in a wish’s wording so that “a wish . . . could be fulfilled without benefit to the one wishing.”10
In practice, a wish becomes something of a contest between players and referees, where referees punish careless players for lax phrasing. In the original game, wishes could be granted only by magic rings and swords, but the publication of the Greyhawk supplement in 1975 brought with it the “Limited Wish” and “Wish” Magic-user spells as well as items such as the “Deck of Many Things,” which could dispense wishes to lucky players. Play reports from 1975 demonstrate a particular consternation regarding wishes and the way referees managed them, one that is emblematic of the broader nature of the dialogue itself.
John Brennick related an early example of the adversarial resolution of wishes in A&E 8 at the end of 1975. Brennick described how two of the three wishes bound to a particular ring were interpreted by a referee who clearly disapproved of the party’s acquisitiveness: “We went outside the town walls and wished for a Vorpal Blade. This wouldn’t come, so I wished for a pair of mated young adult Pegasi. Soon a great fog came out of a nearby forest. When it was about ten yards away, a man came out, riding a Pegasus with another following, swinging a Vorpal Blade. He attacked Hrothgar and cut off his head before Hrothgar even had a chance to swing back with his Holy Sword!” Fortunately, the decapitator was swiftly dispatched by the remainder of the party, and Hrothgar was made whole by an obliging Cleric. Similar accounts of referees interpreting a wish for material gain as an invitation to manifest the sort of powerful foe likely to possess such extravagant goods filled the fanzines of the day. Wishes were made to be spoiled, the conventional wisdom went. “You’re supposed to crock wishes,” Blacow summarized bluntly in A&E 11. “Not only is it traditional,” a fact corroborated by numerous fairy-tale protagonists insufficiently careful what they wished for, but “it’s the universe’s way of easing strain on itself. You give them exactly what they ask for, not what they want but can’t figure out how to say.”
Ultimately, every “positive statement of intention” that is “willed into operation by the player” in the course of the D&D dialogue is a sort of disenchanted wish, which the referee must interpret and weigh before deciding its impact on the game world. Sometimes the referee must simply disallow the stated intention: in the course of navigating a dungeon, a caller could always propose that the party “go south” when that is impossible, and the referee would then be obligated to report something like “you can’t go that way.” But the versatility of the dialogue introduces the possibility of far more verbose and ambiguous statements of intention, which in stressful situations players might struggle to articulate clearly—it might not be obvious how to say what they want. That makes it inevitable that the referee will sometimes misinterpret or correct or reject the intentions of the player.
In 1977, a new game in the tradition of D&D called Space Patrol set the statement of intention on a pedestal as the fundamental operation in the game: “The basic rule of play is that of statement” by the players. But reflecting a few years of experience in the hobby, Space Patrol has to provide some caveats around what qualifies as a viable statement of intention. “It is not enough for the player to draw his pistol and then say, ‘I should fire at it.’” The player must instead make a more affirmative statement; the referee should reject vague statements that express no clear direction. Correspondingly, Space Patrol must also admonish the referee to “never assume anything about the actions of the players. Nothing happens unless the players declare it.”11 This careful language surely reacted against problems already witnessed around the table, such as referees inferring unstated player intentions.
But statements of intention once uttered become irrevocable: “Once a player declares an action, that decision is beyond recall and the player must suffer the consequences,” Space Patrol ominously warns.12 Sir Pellinore’s Game (1979) similarly insists that a declared action cannot be undone: “Once a character says he is going to try to do something he must go through with it. If he says, ‘I’m going to shoot an arrow,’ and he misses and hits a friend he can’t take it back.”13 Even actions with the best intentions can have unforeseen consequences, and as players will these attempts into operation, they must be careful what they wish for. Totten stressed that the educational value in playing his game is to reveal such blunders—sometimes the teacher should fail the student. Game designers thus understood the role of the referee, following the wargame referee of old, to encompass tutoring players in judicious fantasy dungeon exploration. Reiswitz hoped to teach officers to write clear orders; a D&D referee grades players on the circumspect dictation of wishes, whether they have the backing of magic or not.
In this light, it is plain how players might perceive a referee as a sort of opponent and indeed how referees might surmise that their own mandate included strictly—sometimes even uncharitably—interpreting statements of intention put forward by players. Much of the competition that early players perceived between referees and players has its roots in this instructional legacy of wargaming, the test grading inherent in the dialogue. But referees enjoyed the latitude to fulfill this responsibility with either lenity or cruelty, and although the more draconian referees surely alienated gentler gamesters, virtually all gamesmasters reported that they strove to achieve balance as they saw it.
How close did these practices come to what the game’s authors intended? Gary Gygax would write in SF&F Journal 87 in 1976 that he saw D&D as a game of “interaction between the players and Dungeonmaster, and it is as challenging and varied as they are. The two factions alternately act as sounding boards.” That neutral term interaction locates the game as neither a collaboration nor a competition; it is simply a discussion in which the two sides take turns acting as “sounding boards” for one another. Although the referee effectively makes the first move because “the referee creates the basic area in which the players act,” Gygax stressed that an alert referee “will temper his own particular wishes with the tastes of the playing group,” a gesture of accommodation rather than aggression. But Gygax did see something like a competition in how the referee must continually refine the game world in order to test players: “As they learn of this creation, and seek to outwit and out-imagine him, the Dungeonmaster must make further efforts to challenge the participants.”
Although we should hesitate to represent Gygax’s view of the referee’s relationship to the player as an entirely adversarial one, the game was for him a battle of wits. The differences in approaches to D&D perceived by early adopters call into question how fair a referee—the “arbiter of all fortune,” as Gygax put it—would allow that battle to be. The play of the game exposed a tension at its core, perhaps a flaw, relating to how much influence the players’ statements of intention truly have over the course of events.
The dialogue at the core of Dungeons & Dragons would hardly serve as fair grounds for a contest were players unable to craft their own statements of intention. But due to a variety of factors, including the nebulous role of the “caller,” some players might not always enjoy that privilege. The exact degree of participation players really enjoy in the game is largely a matter of referee discretion.
The play transcript in the D&D rules takes place entirely between the referee and a single caller, so we might fairly say that responsibility for crafting a statement of intention in that example resides solely with one player rather than with any other members of the party. Players would offer any necessary input to the caller during the course of each effective turn in the dialogue, and the caller would translate these into a proposed action. Playing with a caller meant that novices could ease into the game through something close to spectatorship: Barry Gold suggested to a prospective player that in practice “you will find your character being played by the party leader until you get enough understanding to call for yourself. Like many other games, D&D is easier to learn if you first watch a few games” (APL 513).
So the presence of a caller meant that players might have little practical agency in a D&D game: they might never formulate statements of intention for themselves. For experienced players, this caused predictable frustration. George Phillies noted at the beginning of 1976 that when it came to appointing a caller to run a party, “‘run’ of course means different things to different people. My own taste is that each person says what he is actually doing” (WH 1). He recalled playing in a party where the caller ordered that Phillies’s Magic-user and a fellow spellcaster move into melee with a wounded orc: “I objected, and moved back, but the other MU accepted that the leader of the party could do this. How one does things depends somewhat on the amount of initiative that people have.”
When Mark Swanson participated in the famous “Tomb of Horrors” tournament at Origins I in 1975, he seized the position of the caller of his fifteen-person party because only “four of the fifteen had any previous experience” with D&D (AE 4). Although Swanson promptly “announced the imposition of military discipline,” in practice he found that even novice players sometimes discovered hidden reserves of initiative. At one point during the adventure, when he had decided for the party “it was time to charge through all together” into another room, he noted that “our Patriarch and the 7th level Dwarf decide to stay put” in defiance of his instructions as a caller. After other players had spectated long enough, Swanson reported that “the rest of the party had now gotten the idea and ordered their characters around to a limited degree.”
In very large parties, even ones made up of experienced players, callers could bring welcome order to groups that might otherwise become bogged down with chatter and mutual recrimination. Virginia Bauer gave an account in the spring of 1975 of how she “found if you have more than five or six players—not characters—players, there is too much nattering, bickering, and confusion! A Lawful party of ten or eleven becomes Chaotic!” (APL 520). However, she still insisted on polling individual party members for their actions; the “leader of the expedition should state at the beginning of the expedition that we should talk in turn, when asked by the leader what our character(s) will do at a given moment.” Some referees had only a limited tolerance for intraparty discussions and disputes; many recommended, as Robert Hollander did in A&E 3, that “too much standing around in one place and chattering should double the chances for a wandering monster.”
Referees could shock dawdling players into action with a surprise incursion of monsters, but the referee’s dominance over game events also meant that the referee could simply advance an encounter without any player’s input. In A&E 12, Charles McGrew sadly observed of his hometown, that “in Raleigh, most DMs allow a basically infinite time of decision (the players always decide in under two seconds or so, but that’s not the point) during which time one assumes the monsters freeze in mid-charge or mid-breath and wait for the group to bring forth their best weapons and fighters to ward off the threat.” McGrew found such leniency unrealistic, so he recommended that referees “give the players a time limit for their decision.” He then gave an example dialogue transcript in which a bumbling caller proposes a succession of unworkable actions, only to finally vacillate, “Well . . . uh,” at which point the referee interjects, “Well here are the orcs.” We can infer a similar instruction to referees of Metamorphosis Alpha, where in its sample transcript of a stressful situation “the referee pauses, awaiting the responses of the players and noting their quickness in acting in the face of this sudden danger.”14
Early reports suggest that many referees who strictly managed the clock treated the lack of a timely statement of intention as a sort of forfeiture: they wanted to force players to make decisions as quickly as characters would. Sheldon Linker from UCLA reported, “Something I have tried recently that seems to work well is to give players six seconds per melee-round decisions. This approximates actual timing, necessitating the player to think as fast as he would have to in the actual situation. If, by the end of the melee-round, the player has not yet decided on a move, then that melee-round is defaulted” (AE 12). A key property of this real-time requirement for statements of intention is that it fixes an interval of time that any statement should cover: it must describe what the character proposes to do over the next six seconds, and any action that would take longer to accomplish may be rejected by the referee or at best be split into segments and completed on an installment plan. Just as a wargame approximates the experience of command in war, so this approach to the dialogue strives to approximate the need for snap decisions that an adventurer would encounter in the underworld.
Linker extended this principle to players who voiced challenges to the authority of the referee instead of statements of intention. He ran a sample dialogue in that same article: “Ref: ‘Four more archers come through the doors. What are your actions?’. Player: ‘Wait a minute, I still think that back around that last corner there should have only been . . .’ Ref: ‘You have taken no action this melee round. The following people have been hit . . .’” (AE 12).
Nicolai Shapero in the next issue of A&E spoke from similar experience: “The DMs I tend to get involved with require quick action. It tends to be, ‘Alright this is what you see, you have ten seconds to consider—what do you do?” Sean Cleary reported in A&E 14 that the rule that “the players have ten seconds to think of something” also prevailed in Boston, where he, Mark Swanson, Glenn Blacow, and Kevin Slimak, among others, had adopted it. In these games, the dialogue became something more like an interrogation conducted by the referee, with its urgent refrains of “What are your actions?” and “What do you do?,” which must be met in timely fashion by a thoughtful statement of intention, or else the players would be reduced to nonparticipants, mere spectators to the unfortunate events that follow.
Sometimes no player had any say in what characters would do, such as when the system dictated that characters had to take certain actions or when players proposed statements of intention that contradicted the rules. As early as 1975, Lee Gold had already grasped that “there are some times that a dungeonmaster should legitimately overrule a person’s call for his character” (APL 520). Gold gave two prominent examples: one where the character wields a magic sword with a high ego, in which case it is the sword, as directed by the referee, rather than the player that will decide certain courses of action for the character. The second is when a “character has been charmed/held by a spellcaster and the player refuses to obey the spellcaster’s commands,” a case where Gold insisted the referee should intervene—she considered this an instance of “general pigheadedness” on the player’s part. Along these lines, she related an anecdote from play about a Fighting-man subject to a Confusion spell who, according to a system die roll, should have attacked his own party. His player instead proposed that his left hand was fighting his own right hand, hoping to persuade the referee that this fulfilled the letter of the law—wishful thinking, it turned out. Here, Gold commented, “the dungeonmaster simply stated that Frank’s character was attacking and chose who, since Frank wasn’t up to confronting the fact he had been spellbound.”
Outside of open insubordination, players might also provide statements of intention that are not actionable, along the lines of the example “I should fire at it” from Space Patrol. Flexible as the dialogue is, a player could phrase statements in any number of ways that no referee could translate into a result. The game What Price Glory?! (1978) illustrates the problem explicitly: “Although a player is free to attempt anything, this doesn’t give him a license to be vague in describing how he will attempt it. If a player merely says, ‘I’m going to try to become king,’ and doesn’t tell how he will try to accomplish it, his statement is meaningless.”15
In part, the difficulty with a statement of intention such as “I’m going to try to become king” is that it is misaligned with the time interval that the dialogue assumes for the adjudication of actions. The refrain “What do you do?” is seeking a proposal for how the player or party will occupy some vaguely scoped but short period of time. In combat, that might mean just six seconds, as Linker recommended, but D&D established a widely followed precedent for supporting different time scales that apply to different modes of the game. In the overworld travel mode, each turn lasts a day, whereas in the underworld exploration mode a turn is just ten minutes, but during combat time compresses into mere rounds. Korns, for example, tuned statements in his dialogue to represent just two seconds of character actions. A statement of intention had to represent something achievable in the implicit timeframe of the dialogue. But, more significantly, “I’m going to try to become king” is a statement that the system of a game like D&D simply has no means to adjudicate.
In a section called “How to Referee an Expedition,” Sir Pellinore’s Game offers a sample diagnosis of failures that can arise when translating statements of intention into actions. Although it stresses that a referee must “let the players have freedom of decision,” it furthermore instructs, “Make the player tell you exactly what his character does. ‘I’m going to escape,’ isn’t good enough. How is he going to escape? Is he going to dig a tunnel? Ambush a guard? How does he hit the guard?”16 Pellinore directs the referee to challenge an impracticable statement of intention and to compel the player to refine it into actionable steps or events that the system can adjudicate. This process might, as the example shows, require multiple exchanges, but once a player identifies how he or she intends to do something—say, to hit the guard—that action should be resolvable by the system.
The question of whether a statement of intention requires or even admits of resolution by the system—which here means recourse to a quantified model and usually to a die roll—must depend on both the sorts of actions that the system covers and the referee’s interpretive powers. Sometimes it is obvious that the baseline rules can resolve an action. The sample play transcript in D&D incorporates die-roll checks that result from various intentions expressed by the caller, such as listening at doors or attempting to force doors open. A die roll to determine the success of both of those activities is stipulated in the baseline D&D rules; for example, “doors must be forced open by strength, a roll of a 1 or 2 indicating the door opens.”17 But affirmative rules of this form are rare in the original books, and they cover only a limited set of actions, mostly ones specific to dungeon exploration. Thus, most of the statements of intention that the referee processes in the sample transcript are resolved not by recourse to a die roll but by simple referee fiat: when a party member scours a pile of refuse for any concealed treasure, the referee makes no system check because the original D&D rules offer no quantification for determining the success or failure of search attempts, so the referee just relays the result.
Supplements to the D&D rules gradually expanded the set of resolvable actions in the baseline system. The first-draft Thief rules, which Gygax circulated around the summer of 1974, introduced percentile skill checks for opening locks, removing traps, moving silently, and hiding in shadows. Shortly after the official publication of the Thief rules in Greyhawk, the Ranger class introduced a similar percentile skill check for tracking monsters. Empire of the Petal Throne around the same time pioneered a professional background skill system, which could enable a character to attempt useful tasks common to a vocation. In Empire of the Petal Throne, all spells have a chance of failure based on the caster’s level, and the skill system reuses that percentile check system to determine the chances of success for actions that include creating alchemical potions, recognizing salubrious or poisonous herbs, and persuading with the power of speech. There is no mention of similar chances of success or failure for other professions, such as building ships, so the results of related statements of intention once again devolve to the discretion of the referee. The community quickly adapted the Petal Throne profession system for D&D; Hendrik Pfeiffer gave his own version in A&E 8, which included a new percentage chance for characters with the proper disguise skills to impersonate someone else, for example. A set of “Birth Tables for D&D” in The Dragon 3 determined, in addition to background descriptions of parentage and social status, what skills or crafts a starting character might know.
As the actions resolvable by the system proliferated, this naturally encouraged design experiments to consolidate action resolution into a single universal rule. In 1976, Richard J. Schwall observed, “It should be possible to replace the plethora of charts in D&D for combat, saving throws, opening doors, thief skills, etc., with a single unified system for calculating the chance of success for any action” (AE 13). From his experience developing his own D&D variant, the “Realm Fantastic,” Schwall then confirmed, “it is possible, for I have done it.” However, Schwall hastily raised some qualms about his own solution. First, he worried that his universal action-resolution mechanism could introduce delays: “D&D mechanics work fast . . . because they are mindlessly simple.” This point recalls the familiar distinction between realism and playability in wargame design, where an exhaustive and thoughtful system for simulating events might prove tedious and impracticable in implementation.
More significantly, Schwall intuited that the realism of his universal resolution system was misaligned with the practical needs of play: “D&D is basically a game of such crude approximations in its very nature that it doesn’t warrant mechanics accurate enough to be used in a wargame,” he suggested. This curious aside makes an important general observation about the purpose of simulating reality in a referee-driven game and about the practical impact of simulation on play. When the referee, in the model of Sandy Eisen, is the only party to the execution of the system—and is thus free to alter the rules at will—players have precious little insight into the factors that determine the resolution of a statement of intention: a referee might have studiously consulted well-considered simulation models or simply blurted out a shrug of an answer. If most actions have at least a chance of failure, and the success or failure of the action is the only feedback on the exercise of the system that the player receives, then the player’s experience of the game is unlikely to demonstrably improve with any strenuous labor on the referee’s part to calculate action resolutions precisely. The dialogue conceals all this from view: before any physical product was sold as a referee’s screen, for hiding maps and paperwork from prying eyes, the reductive power of selective reporting served as the first and most powerful shield for the referee. Without the precision incumbent on the public use of a board and miniatures, as in wargames, and with only a few unrehearsed words to cement the state of a world, Schwall’s misgivings about superfluous accuracy seem well founded.
It is no accident that the concept of a referee guiding the players through a conversation as the “moves” of a game and the concept of a referee exercising discretionary power over the system arose simultaneously in the history of wargaming. The two properties are difficult to decouple. When anything can be attempted and anything can be proposed as a statement of intention, the referee necessarily takes responsibility for improvising new rules to account for unanticipated intentions. Although the D&D transcript does not show the referee making up rules on the spot, the Petal Throne transcript does, in a parenthetical aside, with the referee “mentally giving the warrior a 20 percent chance of being hit by the tiny poisoned projectiles hidden in the hasp, rolling a die and finding that the spines missed the man.”18 People immediately grasped that this was a tacit rule of D&D: in 1976, Howard Mahler would list among the responsibilities of the referee “deciding the chances of success for actions not strictly covered by the rules” (QQG 1). The rules could not anticipate every possible statement of intention players might propose, and when faced with a request that the system lacked the means to adjudicate, a referee had to make a stark choice: either decide the results by fiat or invent some rule on the spot, estimating an appropriate probability of success via some “crude approximation” and then rolling the dice for it.19
In the legacy of wargaming, almost as soon as Verdy du Vernois popularized wargame designs where a referee could “make up” the rules and apply them as he went along, this raised the question of whether events in the game were being decided arbitrarily or, worse, with partiality. Any hobby wargamer could know from that Strategy & Tactics article back in 1972 that “Verdy [du Vernois] advised that most of the rules and the dice be thrown out” of wargaming and that instead “an umpire experienced in actual warfare” would simply decide what should happen under a given game circumstance (ST 33). But the article went on to relate how, when free Kriegsspiel based on Verdy du Vernois was introduced to the United States, it “was itself criticized. Several officers argued that free Kriegsspiel replaced arbitrary written rules with even more arbitrary unwritten rules.” Rather than resulting in a lasting schism between “free” and “rigid” Kriegsspiel, instead, by the twentieth century “there seemed to be a tendency for the two systems to coalesce into one, becoming semi-rigid (or semi-free) Kriegsspiel.” Under this murky compromise, devotees of Verdy du Vernois “were found on occasion to be consulting charts and rules,” and William Livermore, a referee of the opposing philosophy, “was reported to disregard his own tables and charts as often as he consulted them.” Although this gloss on the state of wargaming at the end of the nineteenth century is a bit of an oversimplification, its presence in the pages of the flagship magazine of the wargaming hobby provided ample warning of the ambiguity that could surround event resolution in a referee-driven game. Two years before D&D was published, it put the wargaming community on notice about the perils of these philosophical extremes and the possibility of a compromise. Not everyone may have gotten the memo then, but, as we will see, this history lesson would be periodically reshared with the community into the 1980s.
Crucially, players would never know how the referee approached resolution—or, indeed, even if a given situation falls outside the coverage of existing rules—unless they have some visibility into the execution of the system. If the referee does decide that dicing against a chart or table should determine the consequences of a player’s stated intentions, a further question is to what degree the player gets to participate in that resolution process. The most prominent staging area for this question in the early literature was disputes over whether the player should roll his or her own dice for saving throws, to-hit rolls, damage rolls, and similar checks—or if not, whether a referee should even permit players to witness those rolls and thus to understand how actions and events come to a resolution. The latter philosophy must recall Korns’s principle that players should “only know what the judge tells them that their troops can see or hear” and that, indeed, it is the function of dialogue with the referee “to isolate the players within the confines of the knowledge of their troops” or, as the case may be, their characters.
It is not obvious how involved in system resolution D&D intended its players to be. The play transcript does not clarify who rolls dice: passive constructions such as “a check is made” sometimes mask the dicer’s identity. This ambiguity is especially interesting for one case in particular: the roll for listening at doors. A character listens successfully, according to the rules, only on “a roll of 1 for humans.”20 When a character listens at a door, should a player get to observe the die roll? Presumably the referee should roll secretly, so that the report “you hear nothing” might mean either that the roll has succeeded and the room is empty or that the roll has failed. The D&D rules offer little direct clarity on the subject, but there are other places where the system encourages the referee to roll secretly for some result: the “Fly” spell, for example, will last for a “number of turns equal to the level of the Magic-user plus the number of pips on a six-sided die which is secretly determined by the referee.”21 These rules entice a Magic-user to take a calculated risk, one that may prove fatal depending on an unseen die roll. We have some indication that Gygax reserved checks for game events such as damage rolls for the referee: in a July 1974 letter explicating the adjudication of combat, he recommended that “the referee secretly rolls a die (or dice if the hit warrants) and removes the number shown from the total of possible damage for who or what was hit” (GPGPN 10). Why keep it a secret? We might presume that it is for the sake of isolating players into the situation of their characters: Gygax did not want them to participate in the quantified determination of how successful their hits are and thus how close to defeat an adversary might be. Or perhaps keeping the die roll secret let Gygax exercise his “Divine Intervention,” saving worthy characters from doom, without alerting his players.
The Petal Throne play transcript states more clearly who does the rolling: it shows the referee casting a die when players listen at a door or attempt to break one down. The referee also does some last-minute dicing to determine things such as how many hit points a monster might have. But the transcript goes on to show a player rolling to hit monsters and then rolling for damage and even rolling to see if a creature is surprised. Indeed, the player takes control of events in the sample combat in a way that goes beyond an ordinary statement of intention, suggesting all in one breath that his character “is slashing at the one nearest him” and then, consulting a rolled die before reporting, “he hits with a 19,” and then after rolling again, “he does six points of damage.”22 Although the referee could overrule any of those steps, this example shows a certain awareness of system execution on the player’s part, such as knowledge that a roll of 19 will hit a Biridlú, which runs contrary to Korns’s principle. So already at the dawn of this new genre of game, there were divergent practices for allowing players access to the resolution of events.
Early adopters of D&D treated player participation in die rolls as a matter of referee discretion. “Some dungeon master types prefer to do all the die rolling, providing a narrative for the players,” George Phillies observed early in 1976 (WH 2). When Sherna Burley first played with Lee Gold around that time, she informed Gold afterward that her own dungeon was “going to adopt your practice of letting the players roll their own attack, etc. dice. It was fun, for me as a player, to do it, and I’ll bet it will be more fun for me as a DM not to have to do it” (AE 10). Burley had previously played with groups that did not permit players to roll dice for themselves: she explained that in the New York area there were referees who insisted that “if the characters are attacking with weapons that can’t harm the monster, they shouldn’t know that even a 20 won’t hit” (AE 13). These referees clearly believed that players should not have the same participation in the execution of the system that the Petal Throne transcript demonstrates. John Boardman, another New Yorker, elaborated that some local referees forbid “the players to lay hands on dice once they have set up their characters” in order to keep players in suspense about why attacks might fail (AE 14). But Boardman disapproved, and he stated a weighty consideration as a counterargument: players might lose “the sense of immediate involvement that . . . players have as they roll the dice.”
Even if the players possessed only limited insight into the execution of the system, placing the dice into their hands could transform their attitude toward play. It created a sense of personal responsibility for the outcome so powerful that it sometimes overwhelmed them. Mark Swanson attested that “while I agree that characters should roll their own attacks, some players find it impossible to do this quickly. At least one was reminding me of the Guys & Dolls scene, complete with impatient chorus” (AE 14). That iconic scene, in which Sky Masterson sings the entire song “Luck Be a Lady” while clutching a pair of dice he hesitates to throw for a life-altering bet, aptly characterizes the trepidation a player can project onto a die roll in D&D. Players can feel as if they they are gambling their character’s fortune when they cast the dice.23
But, of course, Swanson did not literally mean that “characters should roll for their own attacks”; he meant that players instead of the referee should roll for their own characters’ attacks. This distinction between players and characters lies at the heart of how statements of intention are formed and the privilege that rolling dice confers. How much players know about the resolution of the system can have a profound effect on which intentions they voice, and thus what kinds of actions characters can attempt in a game. Delivering a statement of intention in the first person, as we see in the sample transcript of D&D, blurs this distinction—and when the referee addresses the party or the caller in the second person, is it to give information to the player or to the character? Attuned to this confusion, Mahler in 1976 carefully worded one of the referee’s responsibilities as “rolling the dice and telling the players what their characters can sense of what is going on,” but by the following year he needed to illustrate a starker distinction between ways the dialogue might be conducted either to include players in the execution of the system or to bar them from it (QQG 1).
In A&E 22, Mahler articulated two “extreme” positions by showing a pair of hypothetical dialogue transcripts that describe the same basic encounter. A party runs across a group of weretigers, one of whom wears a collar granting a defense bonus. In the first dialogue example, the referee explicitly identifies the creatures as weretigers, which can be damaged only with magic weapons, and so the party’s caller directs that only characters with magic weapons bother to attack. Because the referee is obliging enough to give hit point totals for the creatures, the caller’s statement of intention focuses attacks on the weakest one: “we’ll go for the one with 21 hits.” The referee allows the players to roll their own dice: since the armor class of weretigers is a matter of record in the rules, the players will know whether their blows should land, so the referee must further explain that one swing missed “because the weretiger has a collar which increases his A.C. by one.” But after showing that approach to the dialogue, Mahler switched to his second hypothetical transcript, in which a more parsimonious incarnation of the referee tells the party at the start only that “you see what appears to be tigers,” and the caller replies with the statement of intention, “All our front line fighters attack.” After rolling for attacks in secret, the referee this time reports that several attempts missed and that in one case “Joe’s sword seems to have been blocked by a collar around his tiger’s neck,” with no further mention of the collar’s properties. The players here know no better than their characters and must discover for themselves the nature of their adversaries.
Mahler argued that this difference in approach can help to explain the relative lenity and cruelty that players perceive in referees: the second group “has been faced with a much tougher situation, in spite of the fact that the situations ‘are’ the same.” Like Sandy Eisen, Mahler recognized that depriving players of access to the system makes it necessarily more difficult for them to succeed. When viewed side by side, his invented transcripts demonstrate a fundamental philosophical distinction. In one, the conversation refers openly to the system, to the quantified model of the game, from the player’s perspective; in the other, the statements deal only with the game world as the character would perceive it. “The basic question here is the difference between what the player is told, and what the DM knows,” Mahler explained. “I do not believe in withholding any information that the characters would have; however, neither do I believe in giving the players any more than this.” Korns couldn’t have said it better himself—nor could Eisen. Shielded by the dialogue, the referee can isolate the players from the system, which furthermore necessarily grants the referee the latitude to “mentally” calculate the rules for resolving any action without the players ever being the wiser.
Mahler acknowledged that these examples are “extreme,” but questions about the degree to which players are first parties to the resolution of the system, rather than outsiders, recur throughout the early literature. D&D seemed compatible with either extreme, but neither was entirely satisfactory to either of the two cultures, as we shall see in the coming chapters. Without sufficient visibility into the tactical operation of the game, wargamers would feel helpless; with constant interruptions of system mechanics, the “story people” would feel their epic adventure had been reduced to number crunching. But countervailing incentives push against pigeonholing the two cultures that way: wargamers also respected Korns’s principle of “isolating” the players for the sake of simulating reality; and for authorship of the game’s story to be truly collaborative, all of the participants needed the authority that Sir Pellinore’s Game calls “freedom of decision.” Both cultures felt this tension, and neither came to D&D with an easy resolution for it. Precisely where to situate player participation on the continuum between these extremes became one of the first and most important design decisions for the games that followed D&D.
Eisen did not recommend withholding the system from players in order to ratchet up the difficulty, however; he believed that a unique and desirable experience was unlocked in players by shielding them from the rules. It would take a few years of experience before critics started trying to pin a name on that feeling Eisen had, that he “lived the part,” but some early systems did encourage referees to limit the information given to players to solely the sensory data that their characters could gather. Monsters! Monsters! makes this distinction clear by instructing referees, “As the players’ characters enter a given locale, you, as GM, will describe to them what they see, hear, smell, or otherwise sense about the area.”24 Of course, as the sample transcript in D&D illustrates, players can always pick up cues from that description and explore further; Monsters! Monsters! allows that “the players may ask questions for any fine details if they wish,” though presumably only those fine details visible to characters.
But then if we look at High Fantasy (1978), we see a very different approach to play, one with a combat situation not far removed from Mahler’s first hypothetical example. In it, there is no caller: a Fighter and Wizard speak directly to the referee. When the Wizard casts a binding spell at an animated jade statue, he announces, “I have a 58 percent chance,” only to be told by the referee, “This creature has a 20 magic resistance, therefore the chance of success is 38. Roll.”25 The Wizard’s player duly casts the dice, rolling an unsuccessful 82. As the Fighter engages with the statue, the referee obligingly reveals its hit points—its “defensive total” in the High Fantasy system—though the overmatched Fighter has little opportunity to apply this rich information to the tactical situation. We might say, when we read this dialogue, that it is not really an exchange of statements of intention and results; affirmations such as “I have a 58 percent chance” have a different status, and they point to a different mode of participation.
It is again hard to say how well the authors of D&D understood the potential trade-offs here. The rulebooks do imply that referees might inform players of changes to the “basic rules” and that players are to “note them in pencil (for who knows when some flux of the cosmos will make things shift once again),”26 and surviving copies from the era do bear such telltale marks. Early in 1976, Gygax explained how “I am generally uncertain of what ‘laws’ govern things when I play in ‘Blackmoor,’ Dave Arneson’s campaign” (EU 12–13). He accepted this uncertainty without complaint but made it clear that “this is not to say that the players should be denied rules,” in the sense of keeping them ignorant of “the general laws which govern their world.” However, that does not extend to sharing with players information that characters would not know: “As a referee I never tell players what they have found, I simply describe an object, and it is up to them to determine what it actually is and what it does.” Preserving that level of uncertainty about the world is one reason Gygax aspired to “to keep the rules for D&D as amorphous as possible,” permitting vast differences between campaign systems and ultimately greater referee discretion in the moment.
Dungeons & Dragons and the games that closely imitate it take place in a conversation as referees and players discuss game events. But that dialogue can cover radically disparate subjects: it can be a conversation restricted solely to what the characters themselves would know of the game situation, or it can incorporate the execution of the system that resolves game events, or it can fall somewhere in the middle. Where it falls is crucial to determining what it is that people do when they play a character, how lenient or cruel a referee might appear, and to what degree statements of intention should reflect an awareness of the system. Eisen’s vow would steer players toward statements of intention that make no reference to system, and indeed, in retrospect, we should understand it as one of the earliest theoretical stances expressed toward D&D.
It is of paramount importance to recognize that the roots of this design question stretch back throughout the legacy of wargaming, into the extremes of “free” Kriegsspiel, where only the referee understands how game events are resolved, versus “rigid” Kriegsspiel, where the referee executes established rules with little latitude, or the “semirigid” compromise position between them—and that, moreover, the historical disputes around this choice were reviewed in hobby wargaming literature in the years leading up to the publication of D&D. Parsing a dialogue transcript, it is easy to see the players and referee as equal parties to the game, each taking turns adding statements to a narrative work in progress, but this impression rapidly admits of all manner of qualms and caveats. How much control over the game world—or even their own characters—does a statement of intention entitle players to? To what degree are the contents of both parties’ “moves” influenced or even determined by the rules? These questions are intimately bound up in what it means to play a role and what it means for something to be a role-playing game.