03 January 2010

Sevwall's Theory of Design

This post can also be found HERE at the Privateer Press Forums.

I’ve been expounding a bit recently on the difference between thinking like a player and thinking like a designer. I thought I’d just do one big post where I explain my thoughts, hopefully to encourage some of the same thoughts in you, or at the very least a bit on introspection on how you playtest and feedback during this fieldtest. If you don’t think I am qualified to give my thoughts (I am not a game designer, nor have been trained in anything beyond conventional graphic design.) please feel to ignore this post or feel free to read and refute my points.

This largely relates to posts where people complain about abilities and models that are not overpowered from a power perspective, but instead function strangely, counter intuitively, or with unpredicatable power levels.

Definition of terms
So that we are all on the same page, here is how I define some commonly used terms.
Design Space – The ability to add new models to the game without greatly expanding the number of rules that a player must known in order to play the game at a functional level. New rules open up design space at the cost of increased complexity. It is assumed that there is an upper level of complexity that a player base will accept before it starts hurting the game.
Overpowered – A model that is more powerful than a significant majority of similar models.
Unfun – A model that reduces either players desire to play the game.
Broken – A model that is either significantly overpowered or unfun.
Please note, neither term includes the term ‘unbeatable’. Nothing is truly unbeatable, so it is a bad way to define or desribe the above terms.

Greater Tenents of Design

0. Fun is a factor
This is the zeroth rule, the underlying rule above and beyond all else. The game must be fun, or it is a failure, or a job. Abilities that reduce fun must be balanced in some way, or removed entirely. This balancing factor allows for underpowered models and abilities to be balanced by the fact that they are a great deal of fun. This is a concept that does not apply as much to high levels of the game, but must be present to make a game broadly successful.

1. Adjust models to an even level

Models must be balanced against the majority of other models in the game. That is to say, if MAT 6 is average across all melee models, then MAT 7 is good, and needs some sort of penalty when compared to the average model (increased cost, lack of ranged ability, lack of defense, lack of power etc.). If there is no apparent penalty, then the model is better than average, and can lead to power issues. This ‘baseline level’ can be used to help determine if a model possesses too many advantages for its cost or its disadvantages.

2. Balance new models to existing fair models
This is the most important thing that there is. In every game system there are some mistakes, some overpowered models that are clearly identified as such. Currently, many people think Kreoss1 is an example of such. These models cannot be used to gauge the power level of new models. The goal of design should be to limit these overpowered models, so that in time they may be errata’d to a level more in tune with the rest of the models of that type. If you own an original copy of Prime, or played during that time, you will remember Sorcha, and her dominating ability to windrush multiple times and use her feat without LOS. If we were to have designed the game around that power level, the game would quickly devolve into second turn kills where the main way to win a game was to go first. Instead, the designers tried to balance the game around the average caster, and errata’d Sorcha when the opportunity arose.

3. Limited use cannot be a large balancing factor
A model that is useless 80% of the time and overpowered 20% of the time is broken from a design perspective. Some abilities are only useful in particular instances. Vendetta, for instance, triggers only vs Blighted Models. Banefire works only against undead. Disruption works only vs Warjacks, Equilibrium only vs Hordes. The fact that these abilities are not useful at all times, or even in all games does not allow them to be much more powerful than universally useful abilities. When those abilities finally become useful, as in the case of distruption, they must be balanced, or the game quickly becomes unfun. A good example that no longer exists was Precursor Knights in their old form. They were extremely effective vs Cryx, with no real balancing factor vs other factions. Vs other factions, they are fine, vs Cryx they are a bit much. Thus, they are a bit much from a design perspective. A good example of this in the current environment is immunities to non-friendly elements. Cygnar’s widespread immunity to lightning is okay, because it allows the army to ignore some forms of friendly fire. The assault commandoes immunity to fire and corrosion is a balance issue because it is either useless or very good depending on the opponent. Thus, it is very good, and must be balanced by other weaknesses. Another example of this is Drago. Drago’s bond with either Vlad is powerful enough that he must be costed to Vlad, or else be undercosted when used with Vlad. This means that used with any caster other than Vlad, Drago is overcosted for what he provides on the table. This is the only correct way to design with limited use abilities, and its and example of just how undersirable they can be to a player base and the ability to sell and use models.

4. Unpredictability cannot be a large balancing factor
This relates largely to the above tenant, but instead relies upon unpredictable events like rolling a 1 on a D6, or a 4 or less on 2D6. An overblown example is a gun with the critical effect ‘remove target model from play’. This ability is simply too good to be balanced merely by the fact that it is a critical effect. I would go so far as to say that even a very high points cost would not balance a model with this gun, because it can potentially end the game every turn it is brought to bear. A true to the game example is the Mage Hunter Assassin. Decapitate and Weaponsmaster combine to form a melee attack that can remove almost any model from the game when rolled well. Most melee weapons in the game cannot do this. Combined with the low points cost and the difficulty in stopping the attack, this ability to kill almost anything is not adequately balanced, even by its own rarity. This applies to weaknesses as well. The Berzerker is a good example of an unpredictable downside that, even though relatively rare, reduces the models effectiveness tremendously. Flip that, and it becomes clear that rare beneficial effects also tremendously add to a models value.

Rule 3 and 4 Addendum
Rule 3 means that limited use abilities are generally undersireable unless relatively weak. They may, however, be elegantly used when paired. A model that gains weaponsmaster vs constructs and undead, and +2 to attack rolls against living models, has abilities usefull against all models in the game. This is far easier to balance, and is much less of an issue than a model that only possessed one of the above rules. This is one of the main reasons that models with Grevious Wounds have been gaining Arcane Assassin, so that they have an equal power level against all opponents.

5. Rules must function logically
Rules must be primarily utilized in logical ways. Offensive spells or attack abilities must be best utilized against enemies. Buffing effects must be best utilized on friendlies. When the best use for an offensive ability is against your own models, something ha failed somewhere in the design. An example of this is Equilibirum. It is almost never used against enemy targets, counter to its designation as an offensive spell. As such, its use is not intuitive, and it could be changed so that it either functions correctly as an offensive spell, or is instead changed to a friendly spell allowing for its current function. You can see that privateer has partially embraced this logic, from the changes to how sould tokens work, to removal of some abilities like Are You Going to Eat That that functioned best when sacrificing your own models.

6. Rules should not be unpredictable
Rules that state that a model gains an unknown rule cannot be adequately costed for power level. This includes abilities like Mirror Magic or Replication that grant unknown spells to certain casters. The spells they gain access to have not been playtested for the factions or casters that they will be used with, so the power level cannot be judged. This can lead to unknown and potentially broken situations. These particular abilities tend to function like denial abilities against good players, and as such would function in a measurable manner if they were replaced with denial abilities.

Lesser Tenents of Design

Rules must have understandable and widely applicable names
The sword must not have the rule ‘heavy mace’, as it is confusing and unintuitive. This is mostly used to name specific rules. An example in the current game is Decapitation. The mere name of the effect implies that it can only be used on bladed weapons, or by a model capable of finesse. However, the rule itself could be widely applied, since it only provides a way for low POW models to threaten high ARM ones on high rolls. If the rule were called ‘Internal Injuries’ it could easily be applied to most models or weapons, unlike the current ‘Decapitation’. Another example is ‘Eye of Menoth’. This clearly only makes sense on a Menite model, while the effect itself could be applicable to any number of casters who would support an army magically. This means that either no other model can grant the same effect, or the effect must be named twice. Naming the same rule twice is undersirable, as it increases the terms one must memorize to play the game.

Rules should be broken down to their component parts
A rule must not encompass other rules. A rule other than Parry should not grant immunity to freestrikes. An example of this is Perfect Balance, which grants immunity to Combined attacks, backstrikes and free strikes. Memorizing that Perfect balance grants these things adds rules bloat to the game, where multiple rules do bits of the same things. Instead, a model that is immune to combined attacks, backstrikes and free strikes should have the rules Perfect Balance, Aware and Parry, with Perfect Balance granting immunity to combined attacks, Aware granting immunity to backstrikes and Parry granting immunity to free strikes. This increases the number of rules on the card, but increases the ease at which each rule is memorized and recognized. This also increases design space by allowing a model to be immune to just Combined attacks without also being immunie to freestrikes and backstrikes, or by creating yet another rule dealing with combined attacks.
An example of this was Critical death roll on Snapjaw. It granted KD and an additional die of damage on a critical hit. If instead he possessed Critical KD and Critical Brutal Damage, far less people would need to ask ‘what does that model do again?’

Everything that can be easily categorized should be categorized
Design space is a difficult concept, but it boils down to keeping your options open. It is best described by example. Trenchers were a unit. Finn gives leadership to trenchers. Thus, any unit that may be released in the future must either not be comprised of trenchers, or must be affected by Finn. This can be stopped by originally releasing trenchers as ‘Trencher Mudstompers’, and having Finn affect ‘Trencher Mudstompers’ with his leadership ability. Then, future trencher units can be released free of pre-existing interactions. If Finn’s leadership is determined to be desireable on the new unit, you release a small errata in the same release that contains the new unit, changing the rule to effect the new unit.

This, in practice, also means that anything that is Blighted is referred to as Blighted, or Ogrun as Ogrun, or any number of other terms that are deemed useful in the game. This, of course, must be kept under control. ‘Legion Blighted Ogrun Warspear Unit’ is more desireable that ‘Legion Unit’, but is also more desirable than ‘Legion Blighted Heavy Ogrun Warspear Warrrior Ranger Unit’. The number of terms used for each entry should be comprehensive but limited to a reasonable number of terms. This allows for more opportunities to interact with the models in the future without annoying errata or unitended consequences.


If you've made it this far, I hope you now have some appreciation for how a small segment of the population is approaching this fieldtest. Not everyone is purely concerned with the power level of a model, and instead may be concerned with the design of a model. When debating models, try to look at it from the perspective of a designer, to see if what you are debating is a design question rather than a power question.


  1. I like your post, but I continue to believe, and I will rail on and on about this until I get heard, that a FLAT standard of models being "level" with all others at their point value is foolish. I believe that when units are balanced this way, you end up with a stagnant game in which the metagame doesn't do much shifting.

    Currently, there are some models that can be fielded in most lists, and be expected to perform well, while other models costing the same are far more "niche", or can even considered weak, unless properly supported. However, these very "situational" models can suddenly become crucial when a metagame shifts a certain way, and you need a way to bust the prevailing archetype. Now conceivably, one might say that you can achieve this diversity with an approach of trying to balance models individually with each other. I strongly disagree. Such comparisons rarely take into account the big picture, because to do so, you introduce unseen factors that extend beyond what the goodies on that model's stat card say. For example, someone (not me) might say that Skorne's somewhat expensive and inaccurate Venator Reivers are worth their points even though there are far more glamorous shooting units out there for the same cost **because they are on Skorne**

    Or to illustrate it another way, who would be willing to pay more for a command range speed boost, Trollbloods or Cryx? I'm sure of the two factions were to bid on that spell, Trollbloods would win it, with some shockingly high sum ;c)

    I think what I'm getting at is that game designers need to compare archetype to archetype, not units to units. Thankfully, I belive PP has been trying to do just that.

  2. Your theory seems sound. I wonder, however, how it matches up to the desire to have rules reflect background. I know that, for a while, the tagline 'fluff ain't rules' was batted about (and I've never liked that line, not least because 'background' gets a dismissive, cutesy synonym and 'rules' doesn't, but that's neither here nor there), but the truth is that the pieces are designed with a narrative in mind... or is the narrative defined with the pieces in mind? Does Madrak have a Thrown Axe attack because he has an axe in the story, or does he have an axe in the story because the design decision was made to give him a Thrown Axe attack?

    The point to all this (yes, there is one) is that design isn't just about rules, and that sometimes the concepts of good rules design give way to the concepts of adherence to background.

    I'm trying hard to present that as a Thing rather than a Good Thing or a Bad Thing - it's just something that happens in any game that refers to anything outside itself (even chess, presumably, where the knight moves the way it does because horses can jump over other things - even castles? - and bishops run sideways and are hard to keep your eye on). It is, however, a source of Tension, particularly in fantasy games where the background being reflected is the product of someone's imagination and consequently has high value to them.

  3. The interest in video game rules and physics has a sideline which presents a different, mathematical function basis for action with distinctive, sophisticated themes. This approach is programmatic, and capable of advanced plot with cumulative effects as well as rich variations. The key is new quantum science calculations for modeling the atoms in exact detail, which display the animated video of their motions by quantum rules. The equations guiding the 3D imaging give lively, intricate results with programmable mathematical rules. Here's an introduction.

    Recent advancements in quantum science have produced the picoyoctometric, 3D, interactive video atomic model imaging function, in terms of chronons and spacons for exact, quantized, relativistic animation. This format returns clear numerical data for a full spectrum of variables. The atom's RQT (relative quantum topological) data point imaging function is built by combination of the relativistic Einstein-Lorenz transform functions for time, mass, and energy with the workon quantized electromagnetic wave equations for frequency and wavelength.

    The atom labeled psi (Z) pulsates at the frequency {Nhu=e/h} by cycles of {e=m(c^2)} transformation of nuclear surface mass to forcons with joule values, followed by nuclear force absorption. This radiation process is limited only by spacetime boundaries of {Gravity-Time}, where gravity is the force binding space to psi, forming the GT integral atomic wavefunction. The expression is defined as the series expansion differential of nuclear output rates with quantum symmetry numbers assigned along the progression to give topology to the solutions.

    Next, the correlation function for the manifold of internal heat capacity energy particle 3D functions is extracted by rearranging the total internal momentum function to the photon gain rule and integrating it for GT limits. This produces a series of 26 topological waveparticle functions of the five classes; {+Positron, Workon, Thermon, -Electromagneton, Magnemedon}, each the 3D data image of a type of energy intermedon of the 5/2 kT J internal energy cloud, accounting for all of them.

    Those 26 energy data values intersect the sizes of the fundamental physical constants: h, h-bar, delta, nuclear magneton, beta magneton, k (series). They quantize atomic dynamics by acting as fulcrum particles. The result is the exact picoyoctometric, 3D, interactive video atomic model data point imaging function, responsive to keyboard input of virtual photon gain events by relativistic, quantized shifts of electron, force, and energy field states and positions. This system also gives a new equation for the magnetic flux variable B, which appears as a waveparticle of changeable frequency.

    Images of the h-bar magnetic energy waveparticle of ~175 picoyoctometers are available online at http://www.symmecon.com with the complete RQT atomic modeling manual titled The Crystalon Door, copyright TXu1-266-788. TCD conforms to the unopposed motion of disclosure in U.S. District (NM) Court of 04/02/2001 titled The Solution to the Equation of Schrodinger.