Combat Breakdown: Mass-Action Design
How and why modern combat designs fail
Regaining Lost Technology
It is fair to say that the vast, vast majority of TTRPG players are students of D&D versions 3.5 or later. In D&D’s third edition, there was a dramatic shift away from patterns established by previous games and an equally dramatic refocus on the party of players, their characters, and the degree of detail in which they were being carefully tracked—particularly in combat.
It is instructive to see what was lost, what was pushed aside so that today’s conventions and expectations would come to replace it. We will return to the earliest point of TTRPG total convergence and examine AD&D’s systems, style, and assumptions by understanding an example combat from designer Gary Gygax.
Only two books will be referenced: The AD&D Player’s Handbook (PHB) and the AD&D Dungeon Master’s Guide (DMG).
I. Outlining the 1-Minute Round (1MR)
(DMG pg. 61)
ENCOUNTERS, COMBAT, AND INITIATIVE
Combat is divided into 1 minute period melee rounds, or simply rounds, in order to have reasonably manageable combat. “Manageable” applies both to the actions of the combatants and to the actual refereeing of such melees.
We begin with an appeal for “manageable” combat. Resolution mechanisms must accomplish two simultaneous goals: first that we can reason about what might happen and second that the system can be adjudicated with accuracy and ease.
The choice of 1-minute rounds (1MR) is intentional, and the alternative of tracking time with more compact seconds-long rounds is explicitly rejected:
(On combat system complexity)
It would be no great task to devise an elaborate set of rules for highly complex individual combats with rounds of but a few seconds length. It is not in the best interests of an adventure game, however, to delve too deeply into cut and thrust, parry and riposte. The location of a hit or wound, the sort of damage done, sprains, breaks, and dislocations are not the stuff of heroic fantasy. The reasons for this are manifold.
Elaborate rules to track detailed individual combats (note: he is referring to breaking up the melee conceptually into a series of “duels” with this language) are easy to devise—anyone familiar enough with games can design several in a sitting—but he speaks against the designer’s temptation to do so.
Highly specific realism-serving damage models do not fit the “adventuring” theme—they “are not the stuff of heroic fantasy.” For characters and some other creatures, “hit points are not actually a measure of physical damage” taken, and saddling characters with “crippling disabilities” is really just a way of delivering certain death.
(Prefer abstraction; reject so-called realism)
Combat is a common pursuit in the vast majority of adventures, and the participants in the campaign deserve a chance to exercise intelligent choice during such confrontations. As hit points dwindle they can opt to break off the encounter and attempt to flee. With complex combat systems which stress so-called realism and feature hit location, special damage, and so on, either this option is severely limited or the rules are highly slanted towards favoring the player characters at the expense of their opponents.Damage scored to characters or certain monsters is actually not substantially physical — a mere nick or scratch until the last handful of hit points are considered — it is a matter of wearing away the endurance, the luck, the magical protections.
Hit Points are an abstraction relied upon to get away from complex damage models which inevitably lead to the ruin of player characters. This ruin is not the result of accumulation of grievous permanent injuries—the game has magic which could eliminate such wounds—but from the immediate-onset effects such injuries would imply. These effects will cripple characters when they are most vulnerable, essentially demanding they be lucky on injury tables or die.
Tracking The Passage of Time
(The 1-minute round as the basis of timekeeping)
One-minute rounds are devised to offer the maximum of choice with a minimum of complication. …
The system assumes much activity during the course of each round. … During a one-minute melee round many attacks are made, but some are mere feints, while some are blocked or parried. One, or possibly several, have the chance to actually score damage. For such chances, the dice are rolled, and if the “to hit” number is equalled or exceeded, the attack was successful, but otherwise it too was avoided, blocked, parried, or whatever. …
So while a round of combat is not a continuous series of attacks, it is neither just a single blow and counter-blow affair. The opponents spar and move, seeking the opportunity to engage when an opening in the enemy's guard presents itself.
We have a 1-minute-period melee round, where “melee” is used here in the traditional sense (a chaotic jumble of contact fighting) rather than the modern sense (up close). During these 1-minute rounds (1MR), a great deal of action and activity is happening—including feints, positioning, and other maneuvers that eventually provide a chance at a decisive attack.
These decisive attack events are what the dice rolls represent.
In 1MR, we avoid modeling things like injury and hit-location. These would cause a disproportionate explosion in complexity (“consider the many charts and tables necessary to handle this sort of detail, and then think about how area effect spells would work”) but also act mainly as an additional hindrance to the survivability of player characters (“consider all of the nasty things which face adventurers as the rules stand”).
(The role of dexterity and other speed-like traits)
Because of the relatively long period of time represented by the round, dexterity (dexterity, agility, speed, quickness) is represented by a more favorable armor class rating rather than as a factor in which opponent strikes the first blow.
Faster participants are modeled as more difficult to decisively hit rather than modeled as striking first, a concept that is dubious in the context of 1MR.
(Ease in adjudication; fairness in play)
The system of AD&D combat maximizes the sense of hand-to-hand combat and the life-and-death character of melee without undue complication. Because of this, you, the DM, are enabled to conduct such portions of a game without endless resort to charts, tables, procedure clarifications, and over-lengthy time requirements. Players, on the other hand, will not become bored with endless dice rolling and rules consulting, but at the same time will have a reasonable chance to seek escape for their characters should the affair go badly.
In summary, the design addresses the several goals already mentioned. All players have the chance to understand what is happening and what could happen. The level of abstraction allows, simultaneously, for ease of adjudication and for the possibility for characters to run if fate turns against them.
Fast but meaningful resolution is the key balancing act. Too fast and we lose the idea of what happened (and also lose the ability to imagine what could happen). Too close and we become lost in the weeds of establishing precisely which footstep happened and how.
Initiative
The 1MR with decisive endpoints modeled by to-hit rolls is clear enough. But we need to understand the ordering of actions within the 1MR.
(PHB pg. 104)
INITIATIVE
The initiative factor affects who can do what and when during the course of an encounter of any sort. …
Initiative allows one group, the party or the monster(s), to begin some course of action prior to the other group. Actions affected by initiative are many and include slamming a door, fleeing, moving to grapple or melee, a call for a truce or surrender, firing wands, discharging missiles, beginning a spell, and so on.
The initiative check is typically made with 2 six-sided dice, 1d6 for the party, and another of a different size or color for the creatures encountered. This check is made each round of play where first action is a factor.
We use initiative by side—not by individual participant. One side of the conflict takes their actions; then, the other takes actions afterwards. The following passage (from the same page) makes this expectation crystal clear.
…the first “blow” will be struck by the side gaining initiative during the round. Surviving opponents will then be allowed their attacks.
Also, we have a gentle but clear note not to even check initiative unless it could result in different outcomes. Lastly, a supporting statement on speed enhancements:
(The role of dexterity and Haste/Slow)
Because a round is a full minute long, dexterity seldom is a factor in the determination of which side acts first. However, if one group is slowed or hasted, or one or more members of the group are, the initiative will always go to the non-slowed or hasted side. In most other cases, the group with the higher die score will always act first.
II. Atomic Turn Structure (ATS)
Modern TTRPG combat designs use an atomic turn structure as their foundation. Both time and space are broken down into small elements. To understand the “action” we must take into account the summation of state changes of all these elements.
Every motion, every footstep, every potential clash of weapons is expected to be modeled with a resolution mechanism. This is accomplished by co-ordering the elements (space and time) into precise sequences.
With regard to time, every individual actor can take their actions when their allotted time in the ordered sequence comes. Even if multiple actors share the same initiative, there is nevertheless an ordering established so that 3A goes first and then 3B then 3C and so on before we get to the 4s and the 5s. In this way, a precise and definite causality is established through a sequence of “individual combats” (duels!) that are resolved on-the-spot, one after the other.
With regard to space, modern combat design divides it into a grid of cells (usually squares) that define both relative position and relative distances. Rather than being descriptive, most of these cells instead end up being heavily prescriptive. We no longer say, “Orc B is about 30' south-southwest of that gnarled tree.” Instead, we say, “Orc B is in this cell, which happens to be 3 cells west and 4 cells south of the cell the tree occupies.” Cells evolve beyond a means to understand or convey distance and location—they become locations that act as distance markers.
III. An Example Combat
Part 1
(PHB pg. 105)
A party of 5 characters — a magic-user, a cleric, a thief, a human fighter, and a dwarf fighter surprise an illusionist with 20 orcs. The opponents are 30' distant, and the magic-user immediately begins casting a sleep spell. The cleric also prepares to cast a spell, silence, 15' radius. Meanwhile, the thief darts to the rear of the party to attempt to hide in the shadows and attack from behind when opportunity presents itself; the human fighter nocks an arrow and shoots it at the illusionist; and the dwarf hurls an axe.
Analysis: 1MR
Even though we begin with surprise, we can see the pattern for how things will unfold with 1MR. “I’m casting Sleep,” “I’m hiding in shadows behind the others,” “I’m attacking the illusionist” etc. There is no consideration of where exactly anyone is “standing” or positioned—details like these are expected to evolve throughout the course of the round.
Analysis: ATS
Since the opposing party is 30' distant, there is an ambiguity. We are already in a position of needing to precisely define every individual participants’ relative distance.
Where exactly does the thief stand to hide “behind” the party? And when? If he stands behind someone and they move, it was all for naught! Even this simple idea (“I want to stay out of sight behind the party”) demands a fully tactical formulation and execution.
Part 2
The surprise segment is over, and initiative is determined. The illusionist/orcs win initiative, and while the former begins a spell of his own, the latter rush to attack, hurling spears as they come.
Analysis: 1MR
This one-sentence description is the entire turn, though the designer leaves some detail out. Since there are 20 orcs, we might expect they have 20 throwing-spears (or 10 or even only 5, depending on the role these orcs are playing).
Since their throws are simultaneous (in the abstract and mechanical sense), it is sensible for the orcs to split their attacks evenly between the four visible party members (or some variation thereof)—a simple hail of throwing-spears. In that case, we roll four sets of 5d20*, looking for some target number on each to correspond with the character AC. Afterwards, the orcs close distance.
*(There is certainly room here to roll even fewer dice by usage of simple scaling rules, the discussion of which is not essential to these points.)
Even this additional level of inferred detail is simplistic in the extreme compared to the modern style.
Analysis: ATS
It is important we consider the stark difference between the brevity of description offered and the relative slog of enacting 21 combatant turns.
In modern designs, each orc would get their own turn—their own declaration and attack. Orc B throws a spear—at whom? And where does he move, exactly? And does the spear hit the target? Okay, now onto Orc C, and so on down the line.
Consider the orc turn pictured above. The orc throws his spear, successfully striking and damaging player character 1 (PC1) then uses his movement to move next to PC2—his turn is then complete. It would be most logical, given how this system works, to have orc #2 focus fire the target already struck and then move in a similar fashion. Each orc would focus fire PC1 until they manage to bring him down to 0 Hit Points.
Then we move onto the next PC in line. But if orc #13’s turn comes up, he may actually be unable to throw his spear (without penalty) at PC2 because several orcs have already ended their turn in the cells around PC2.
And what happens if PC2 is eventually felled by spear-throws somehow? All the orcs that spent their movement allowance surrounding him will have, in the final accounting, failed to close to melee!
Not only is this scenario painfully drawn out in comparison to a 1MR interpretation, the action itself is changed. Rather than a simultaneous hail of spear-throwing, we have a sequential snipe-snipe-sniping until PC1 drops, and then the next PC and so on.
Breaking the group’s actions down into individual turns creates a silly story compared to “the orcs toss a hail of spear-throws at the party and close to melee.” Even in small groups of 2 or 3, it takes away the teamwork feel that is possible with 1MR.
We must ask ourselves: how long does enacting 20 separate orc turns take? And how long does it take to roll four sets of 5d20 and declare crowding in?
Part 3
A spear hits the magic-user, so the sleep spell is spoiled. The orcs are attacked by the fighters, the cleric casts his spell upon the illusionist, and the magical silence both spoils his prismatic spray spell and enhances the chances for the thief's attack, for he is successfully slinking and sliding around in the shadows. Thus, after surprise and 1 melee round, the party has inflicted 2 hits upon the illusionist, spoiled his spell attack, and felled one orc and wounded another. They have taken 3 spear hits and had one of their spell attacks ruined.
Analysis: 1MR
After the orcs’ hail of spears and crowding in, initiative moves to the party. One of the spears struck the magic-user, ruining his spell (which would not have happened had the party won initiative!). The cleric is not so unlucky, and manages his magical silence.
The designer proposes a particular possible tally of decisive strikes after a surprise segment (favoring the party) and one full 1MR of back-and-forth. Note that in mass-action thinking, it only needs to be a tally. We do not need a precise accounting of which orcs were struck and which spaces they occupy, and so on.
Analysis: ATS
Where exactly is the thief hiding in shadows? Can he justify that his cell (and potentially all the cells he moved through) really provides enough obfuscation to stay hidden?
A tally of casualties and damage on the enemy is not sufficient information to continue in ATS. We need to know in which cell the dead orc is located—he has likely become either a gap or a travel obstacle in tactical consideration. Additionally, we must know in which cell the wounded orc is located so that he specifically can be targeted with follow-up attacks.
Part 4
Initiative is now checked for the second round. The illusionist/orcs again win initiative and attack first, 5 orcs going after each fighter to grapple, 6 rushing the magic-user, and 3 heading for the cleric. The fighters are pulled down, as is the magic-user, but the cleric avoids their grasp. The illusionist begins casting another spell, one which requires no verbal component; he does not hear the thief behind him.
Analysis: 1MR
The orcs use their superior numbers to swarm-grapple the party. Even though grappling is “infamous” in D&D circles for its complexity, this isn’t difficult to figure out. Each fighter has five orcs facing them. Grappling in AD&D amounts to a d100 roll, modified by a few things like the type of armor the opponent is wearing. Roll 5d100 for the one fighter, then 5d100 for the other—each fighter will have a different target number. The orc tactic is to hope at least one d100 rolls a high result—and they happen to manage it this time, except against the cleric.
Analysis: ATS
Again, the questions about the thief’s actions demand answers.
We run into another silly story here where orcs, one at a time, approach the fighters and individually attempt to grapple them. Maybe the orcs will get lucky and only need e.g. 2 and 4 orcs instead of 5 orcs each to pin the fighters. The others can then be used more efficiently, from a tactical standpoint.
Just like in Part 2, we are going to need to justify exactly where each orc ends up standing—so as to maximize future orc grappling attempts if necessary.
Part 5
It is now the party's turn in the round. The cleric smites 1 of the onrushing orcs and kills it, and the thief stabs the illusionist from behind with his sword, killing him; the fighters and magic-user are held fast by orcs, so they can do nothing. Round 2 is over.
Analysis: 1MR
In mass-action thinking, it is easier to understand how a lone actor could go unnoticed. He is after a distracted target and amidst a brutal melee.
Analysis: ATS
The thief needs to specify precisely how he navigated, what path or stratagem was taken, what cover was being used, and so on, in order to get to this point.
One of the issues with breaking complexity down into atomic elements is that much is lost. It is easy enough, in a 1MR context, for a thief merely to say, “I use the confusion of the melee to continue advancing unseen on the illusionist.”
In the ATS context, where is this confusion? Which cell does it occupy? In which turn is it present? These questions are slightly unfair, since we are free to use the same creative reasoning in ATS contexts, but it illustrates the fundamental point that mass-action thinking is more correspondent with our intuitive grasp of “action.” In ATS, we are consistently reasoning at the level of cells and turns, but in a 1MR context we are grappling with less precise notions more open to interpretation.
Part 6
The initiative roll in the third round goes to the party. The cleric kills another orc, while the thief rushes at the orcs holding the nearest fighter. It is now the orcs’ turn, and as their leader is dead and they still face 2 powerful opponents, they will check morale. It is probable that they will kill the pinned characters with dagger thrusts if their morale does not break, or that they will release the pinned characters and run away if their morale is bad.
The party took a risk and got in over their heads—now it all comes down to a morale roll. The fighters and mage lie helplessly pinned, outnumbered. Only the thief’s swift assassination of the orcs’ leader is giving them an opportunity for life.
The Path to War
In modern combat designs, each individual actor is painstakingly modeled in atomic fashion until each atom has had its spotlight turn. But following mass-action thinking, we have the same amount of action happening (with nearly the same resolution) but without the demand for all elements to be so precisely laid out.
What if the fight were not between 5 and 21 but between 50 (on the player character side) and 200 (orcs)? It is trivially easy to have a marching party of 50 with high-level AD&D characters and their combination of followers and henchmen. And in the wilderness encounter tables, one is expected to run into between 30 and 300 orcs in an encounter.
We already know how this 50v200 can be played out with 1MR. Does ATS have any hope of approaching it? Yes, but only if we relax and expand and smudge. Larger atoms, more clustering structures, more support for true simultaneity—in other words, moving closer to the mass-action 1MR design.
Future TTRPGs must be wargames, and it stands to reason that a combat system supporting all scales of conflict could be the glue that binds their elements together. ATS designs will never be capable of this without relaxing the atoms, the turns, and the structural rules that hold them together. As soon as the scale of conflict overshadows the size of our atoms, the action is lost in a sea of detail and concern over quantities of low relevance.
Whole generations of designers have pursued increasingly atomic systems to the point where it is a cultural norm among tabletop gamers to joke about someone taking 20 minutes to finish their turn. This level of mistake requires very serious reflection on our part.
To avoid furthering this cataclysm, we must reverse course entirely—beginning with the presumption that ATS will not work to model our game systems. Gygax, even at the beginning of TTRPG convergence, warned about the danger of ATS design—and it fell on deaf ears. The time has come for us to listen.