Image by SilviaP_Design via Pixabay

I get asked questions all the time. Those coming at me through twitter I usually answer extremely briefly; if the 280-character limit (was 140) is not enough, I will either add more replies or set the question aside for answering here, if there is likely to be sufficient interest in the question. Those I get asked in person, I answer as best I can within the limits of the time I have available. And I average about 1 answer a month on Quora.

Never heard of it? It’s a site where some people ask questions and other people – who feel they have the expertise to provide a useful response – answer them. Some of my answers have been about Australia, some about Politics, some about Cricket, some about other topics – and yes, some have been RPG-related.

Until now, none of the answers have been suitable for expansion here. On Saturday, that changed.

The Question was, How do I create a good combat system for a tabletop RPG?

Below is the full text of my reply. You can confirm that (and read the other answers to the question, which are also worthwhile) at How do I create a good combat system for a tabletop RPG?

    Combat systems are the crunch of RPGs, the most intensive interaction between players and game mechanics. Before you start, you need to understand the fundamentals of time-and-motion optimization.

    If the average combat between 4 PCs and 4 equally-matched NPCs requires 6 rounds of battle, a single extra step taking 10 seconds (which seems quick and easy in isolation) for each adds 6x8x10 = 480 seconds or 8 extra minutes. If that time-span is 20 seconds, that’s 16 minutes game time lost per combat. If there are 12 rounds of combat in those typical battles, that’s 16 and 32 minutes of extra time per combat.

    If it takes five extra seconds to perform an operation – the difference between addition and subtraction – and you have to do that for 10 rounds, 8 characters, 4 times in day’s play, that’s 5x10x8x4=1600 seconds=26-plus minutes of extra game time, lost. Plus, it makes it slower to evaluate your different tactical options, and that can add anything from 5 to 25 seconds on top of that five. Two or three such operations, and you can easily lose three, four, or even five hours of game-play.

    The next thing you need to get your head around is the relationship between simulation and abstraction. Clearly, a combat system that takes account of everything that can possibly affect the outcome is going to be so slow that it’s unplayable. That means that you are going to need just enough realism that it feels like all possible combat modifiers are taken into account (i.e. feels ‘realistic’), but abstract enough that it’s playable.

    Thirdly, the system needs to contain mechanisms by which the participants can influence the basic operation of the system – ways to implement tactics, ways to differentiate one character from another, different styles of combat, and so on – and these cannot add to the combat-time overhead. Accommodating these ‘differentials’ usually requires further simplification and abstraction of the combat simulation.

    All of these principles wrap around your core combat mechanics, and don’t leave a lot of room for complexity in those mechanics. Which is why most systems come down to a chance of hitting, an amount of damage inflicted on a successful hit, and a total capacity for absorbing successful hits, and circumstantial/tactical modifiers to one, or at most two, of these. The details vary from system to system, but those are the core principles. There simply isn’t enough room for much more in the timing. Hero games splits damage into two varieties, and adds an Endurance variable that restricts medium-term capacities for combat – but uses a mechanism that requires only a single roll to calculate both damage results.

    Finally, all combat systems have to pass the sensible test. In other words, they have to feel like they respond to common sense. Put on better armor and you become harder to hit to the point of taking damage. Put on more restrictive armor (which might be the same thing) and it becomes harder for you to hit others and inflict harm with that ‘hit’.

    If you want to create a new combat system, start by understanding the ones that are already out there. Throw in an understanding of how card games work. Toss in an understanding of probability. Then follow the process outlined above. And think of the results as just your first draft, because I can guarantee that they will need revision after some real-world play-testing!

Now, there’s nothing wrong with that answer, and I stand by it. But there’s a lot of meat in there that could be better-served with some fleshing-out and annotation.

Time-And-Motion Optimization

Back in 2014 (it doesn’t seem like four years ago), I wrote a full article on the subject, but that was focused more on House Rules and the inevitable inefficiencies that come as a price tag. I had in mind things like incorporating extra classes of modifiers, or hit location mechanics, or different types of damage, or a more Rolemaster approach to weapons and the other rules tweaks that every GM seems to come up with. This is a new application of the principles, and that’s why I went into some detail in the above answer – nearly 1/3 of it relates to the subject.

Another way to phrase this aspect of the answer is to ask a distinct and related sub-question: how fast should combat take to complete?

At it’s absolutely most basic level, a single die roll that takes into account every relevant factor could resolve the battle, and the rest is GM narrative. I think this goes too far, but there may be times when it’s the right solution.

A better solution is a single die roll that then leads into a formalized and structured narrative interaction between players and GM. This essentially uses the die roll and some assessment of the “quality” of the resulting success or failure as a guide to the outcome while saying nothing about the path taken:

In this example, the characters have rolled a 16 on a d20 but don’t know what they needed to succeed because they don’t know what modifiers the GM has placed on the roll. The GM initially has the battle advantage swing this way and that before events turn decidedly against the PCs, ending in a sudden decisive twist that produces the successful outcome they earned through their die roll. They can attempt whatever they want to attempt but will have little success until the final moments of the struggle.

Along the way, the GM might dole out damage and changing tactical circumstances as the combat evolves, based on what the players attempt and what the enemy does in response. But the ultimate question, “is a 16 good enough?” is not resolved until the end of the battle.

This sort of approach permits the GM to make combat dramatic and visceral – if he’s good at that sort of thing – but can leave the players ignorant of specifics that would affect their tactical decisions, and leaves no place for in-combat strokes of genius on their part.

For that reason, a lot of players wouldn’t like it. They want to know everything about their character’s condition, they want to be able to judge for themselves how effective or ineffective a particular tactic or weapon is, and so on.

And so we end up back at the basic RPG combat system, in which round by round numeric consequences of combat are described and any narrative is something the GM has to shoehorn in, sideways.

Simulation Vs Abstraction

I’ve touched on this subject many times in many articles, but I don’t think I’ve ever addressed the subject directly – see, for example, the series on Cinematic Combat from 2015.

There’s a difference between realism and verisimilitude that many people confuse. Verisimilitude is all about the suspension of disbelief, of logical consequences deriving from improbable or even impossible assumptions, of making the game – and that includes the combat – feel as though it was real, despite the presence of all sorts of strange phenomena like Sorcery and Monsters and what-have-you. Realism means that there is no disbelief to suspend, and nothing so out of the ordinary is possible. It takes the fantasy out of the equation, replacing imagination with simulation. Verisimilitude is to be encouraged at all times; Realism may be a vector to achieving Verisimilitude, but at least as often, it gets in the way.

Simulation attempts to load up on specific details. In theory, with a well-designed combat system, these details are epicly good, because they carry inherent verisimilitude. But they do so at the expense of practicality. I have seen (home-brew) game systems in which it took 12 minutes to resolve each attempted attack in a combat. Everything you can think of was taken into account, from the capabilities of weapon-type-X against armor-type-Y, to the relative heights of the two combatants, to skill levels with a shield, to…. well you get the point. This might work well in a computer RPG because the computer can crunch this type of math faster than you can blink.

In any other sort of game, there isn’t enough abstraction; it takes too long, can be too confusing, and intrudes game mechanics so severely into the gaming process that any gains in verisimilitude are often lost. You might know that the outcome is believably accurate, but you don’t feel it, viscerally.

Every RPG, every combat system, has it’s own point of compromise between simulation and abstraction. If that was all there was to it, you could pick the point that seems most desirable to you, and all would be well.

But this feeds directly back into the question of Time-and-motion Optimization, as was implied by that home-brew system I mentioned. You can’t accurately simulate every aspect of the real world in a practical combat system; you need to abstract and compromise the realism in order to achieve both a dynamic, thrilling, visceral contest, and a sense of verisimilitude.

There are good abstractions that are easy to understand and work with, and there are poorer abstractions that are extremely complicated when you dig deeply into what seems to be a very straightforward mechanic. An example of the latter that I analyzed extensively is D&D 5e’s “Advantage” mechanic – you can read my analysis in this article.

Controls

There are always rules for things like facing, and initiative, and surprise, and movement, and cover, and flanking. These are all variables that the characters engaged in combat can manipulate in an attempt to Control the outcome. No rules system can ever cover everything, or you end up back at the Simulation end of the scale.

For example, in D&D (any variety), you might have a player state “I am attacking to draw his shield out of position. I don’t care about hitting him – yet – I want to spend a round or two defending myself and getting him into a vulnerable position.” How do you adjudicate this? Unless there is a specific maneuver listed amongst the controls that describes this, you’re in trouble.

Before you can create and implement such a Control, you would need to really understand the abstractions that are built into the D&D system. One of those abstractions is that each attempt to “hit” is actually many attempts to strike a blow against the enemy, and your chances of success at inflicting harm incorporate any skill you might have in getting a character’s shield “out of position”. Ultimately, you could craft an explicit maneuver that would work – storing up your success margins above what you need to hit over multiple rounds in order to achieve a more effective blow when you do strike, for example – but you need to understand the abstractions and foundation assumptions of the game system first.

I’ve seen a number of house rules and attempts to create combat systems where the designer lost track of the abstractions, or never understood them in the first place, and as a result, their home-brewed rules came with gaping, exploitable, system deficiencies.

You need Controls, or the characters will have no capacity to treat the combat as though it were real – breaking verisimilitude – but you can’t cover everything. D&D 5e’s “Advantage” is an attempt to umbrella everything into a single abstraction, and – to a certain extent – it works well.

The Sensible Test

The Space Gamer, and more recently, Pyramid – both by Steve Jackson Games – used to feature a page called Murphy’s Rules, which mocked (in a comedic way) rules that failed the Sensible Test. Many of these have been collected into a PDF (US$ 7.99). For example, in AD&D, a fighter specializing in Darts can do more damage in a round than one specializing in Broadsword…

Every game designer has, at one time or another, written a rule that fails the Sensible Test. If they are lucky, it’s only in exceptional circumstances, and a minor tweak can set things aright. More often, this comes about because the left hand either didn’t know or forgot what the right hand had already done.

If it comes to that, the limits of abstraction frequently pose challenges to the Sensible Test.

Take the D&D 3.x stacking limits, for example. That game system had a quite sensible mechanic designed to simplify the abstractions question, only different modifiers stack. If you had two modifiers of the same type, you simply used the biggest one and ignored the lesser. There were one or two minor exceptions.

What did various supplement publishers do? To make their products more appealing, they came up with their own “categories” of modifier. You could get one magic item from one book, a class ability from another, and add them to something from the core rules, and end up with three STACKABLE bonuses for something that was never meant to stack in the first place! If those were just +1 or +2, that would be bad enough, but some of them were +4, +5, or even more.

I once constructed a character in this way, as a test to ascertain the seriousness of the problem, and ended with a 1st-level character who had +40 to stealth. No, I no longer have the details.

What to do? Well, you could restrict the publishers that were considered canonical in your game. Or, you could rule that unless it’s explicitly differentiated in the core rule-books, all these modifiers were considered the same type under a different name, restricting the scope of the problem.

But, at the same time, there were logical challenges to the concept of stacking limits. “You mean that because of my Cloak, I can wear hobnailed boots instead of my Boots Of Elvenkind and there’s no change to my Stealth roll?” kind of challenges. The concept itself fails the Sensible Test.

You could solve that whilst retaining the spirit of the rule. Many GMs did. The simplest solution was: total all your applicable bonuses except the single largest one. Each time those other bonuses equal the largest one, in aggregate, or part thereof, it increases the benefit of that largest one by one. So +2 and +2 and +3 gives you +5, not +7. And +4, +5, +4 gives you +7, not +13. It meant that characters still got a benefit from wearing multiple items that contributed the same bonus, but those benefits were curtailed to some extent.

Understanding The Game System You Are Using

You can do a lot worse, either as a player or as a GM, than taking the time to analyze and understand the combat system of the game you are using, even if you have no desire to create a new game system. Identifying the abstractions, and understanding the implications, comprehending the mechanics, being aware of the line between Abstraction and Simulation that the rules adhere to and why – these all make you better equipped to deal with failures of the Sensible Test when they occur, and increase your capacity to arbitrate the inevitable situations that aren’t in the book.

We all encounter those situations. Quite often, they result when an unusual combination of rules comes into effect; the better the play-testing that’s been done, the more the obvious such mechanics “holes” have been identified and corrected. But RPGs deal with environments and situations that are so complex, it’s impossible to have test every possible combination; there will always be tar babies and other monsters lurking in the shadowy recesses of the rule-books.

It’s part of your responsibility to hunt these down and whack them on the head before they become a problem, if you can.


Discover more from Campaign Mastery

Subscribe to get the latest posts sent to your email.