Wednesday, November 23, 2011

Extra! Extra! Read All About It!!

As an extra post to this series, we'll examine ideas surrounding proper implementation of role-based classes or niches. The mundane task of proofs behind us, we now have fun with design!

The assertions of tanking, damage, healing have been established as the fundamentals for combat to occur. These are modified via (de)buffing and utility. Lastly, logistics allows for the setting of combat to occur.

Also asserted was how players in a multiplayer game cannot be masters at everything, that spheres of influence must be of the same size. This results in a need to specialize in order to avoid perpetual stalemates in combat (flanking, balance). Lastly, groups can achieve personal specialization while maintaining collective balance as the group's center of mass is not the individuals' center of mass (unlike in a solo situation).

The Masons

And now we build our character options. Now that we've established the fundamentals and inter-workings of combat, we are free to pick up the proverbial paint brush and begin the creative work.

Before, it was suggested that if one cannot get out of doing something, they might as well get into doing it. This is to say that if separation from a given situation cannot be achieved, one might as well make the most of it, pour everything into it.

We'll look at two extremes.


On one hand, we establish a system that calls for only tanking and damage from the user. Built into the game design we do not adhere to the need for healing to be fulfilled (let's say it's achieved via outside means such as healing packs, rezzing, power-ups etc). We've reduced our game past the lowest identifying point of combat. What effect does this have?

It sucks knowing it's there, but not being able to play with it. People will look to only avoid getting hit as no one can actively heal (reducing the functionality of tanking) and will always 'overextend' because there is no fear of a target miraculously regaining health (reducing the strategic effect of damage).


On the other, we establish a system that calls for all three: tanking, damage and healing. But in this example we do not just simply 'tank' or 'damage' or 'heal'. We further divide these into 'damage prevention', 'attention getter', 'area', etc for tanking. We further divide damage into 'over time', 'chain targeting', 'attrition', 'momentum', etc. We further divide healing into...

We can quickly gain a rat-infested case of niche, niche-counter, niche-counter-counter... Fun at first, but quickly mind-boggling when continued ad nauseum?


But where do we draw the line? Obviously, more interesting gameplay is birthed from more variety of unique roles to fill, but at a certain point, the options become too numerous to wrap one's mind around.

The take-away, in case one missed it, is that interesting gameplay comes from options, not the limiting of. This is what we should be striving for, not to reduce the trinity, or take away its defining characteristics. We need to be striving to add to them.

Hybrids (a quick note)

In order to bring something of value to a group, one must be able to provide something of interest to the group. In the case of a group makeup of two tanks, three damage, one healer, and one crowd-controller, can it not be agreed that the addition of a buffing role would unanimously be value added to the group? Similarly, if we add another tank, can it not be agreed the gain would be marginal at best, redundant at least?

The value of a hybrid is the ability to fill multiple roles. For balance, hybrids would never be 100% of a purebred (ref: areas of influence). Let's give an example of a group consisting of a tank, a damage, a healer, a buffer, and a crowd-controller. The flexibility of the hybrid is what is value added to this group. Now if the group were entirely hybrids, we are only further diluting a diluted system by adding another.

How Roles (via Trinity) and Hybrids Manifest

Let's assume a generic coverage of all possible ways, for both breadth and depth, a tank could tank, with the number '3'. This means all sub-niches (direct vs over-time, etc) of tanking are contained within the totality of this number. All numerical scaling (hits for 1000 vs 500, etc) is contained within this number. The same for damage and healing.

This shows what a purebred trinity approach would look like. A 3player group covers all the bases. Content can only be scaled down, as no one else can help perform any other roles than their primary. Content can be scaled down either via the individuals (less performance is required of any individual), or via the roles (less roles are required of the group). The net result is that no one can expect to perform above and beyond their capabilities, there is too much role definition.

Recall that the area covered by a circle of influence is constant. Therefor we will modify our thresholds while considering a half-breed approach. The area of influence remains 3, however, content is scaled so that the totality of any role is contained within the number '2'. Everyone has a primary role, and a secondary role.

This half-breed model shows promise. A 3player group covers all the bases, however, content can ask more than what a given individual can perform, requiring the aid of a secondary. This results in the possibility of content being scaled in either direction, both for members of the group, and to the group at large. This reduction in predictability is the path towards dynamic content.

Let's consider a group of hybrids, again rigging our content to include all aspects of a role within the number '1'. Areas of influence remain at 3. No one has a primary role.

Like the purebred approach, the only result is combat requirements that call for members to either individually perform certain roles, or to stack up in a certain area to meet a requirement. Unlike the purebred approach, the scaling is that the group's requirements will always require more of a given role, but the possibility for the individual remains for content to scale down.

Expanded Model

A highly simplistic model, huh? Let's go a bit further and consider the applications of the combat derivatives as roles proper for content requirements.

Just for numbers' sake, the maximum of all abilities and ability output will vary to match the tallest order that a given setup allows under the area of influence rule. We would be poor game designers if we made content that required abilities or numerical outputs from said abilities that are not possibly provided by the players attempting the content! We again start with a purebred model to consider its implications. '5' is the number including the total of abilities, with an AoI of 5 (simply to match the number of players in a group).

We quickly note that the purebred in an extended model still yields the same take-aways as the simplistic. This will be true regardless of how far you extend your model.

Let's consider that half-breed approach that showed so much promise in the simplistic. '3' is the number including the total of abilities, with a constant AoI of 5.

Woah, Nelly! This is truly interesting indeed. Notice that the trend of being able to scale content in either direction is not only present, but the possibilities for performing these requirements simultaneously exists.

I wonder what an extended model might yield for pure hybrids. '1' is the number including the total of abilities, with a constant AoI of 5.

A little better than before! Stacking can now occur at different areas, which is better than the extreme predictability of the purebred approach, however, content scales only in one direction, leading to predictability. This will be a problem for a pure hybrid approach no matter how extreme the extension of the model.

Back on Track

So how do you make a dynamic game, full of both simple application and complex encounters, minimizing predictability? Well, as we've seen throughout the series, the first step is to include the trinity. The second, as just shown, is to include the derivatives. A third take-away would be to maintain the boundaries of one role within a given person, but allow a little bleed-over into a second, maybe even third or fourth, role.

But don't stop there. Flush out the aspects of each arch-role. Damage could be made to be performed with physical, psionic, elemental or cosmic rays, with matching resistances for tanking. Healing might be incorporated into buffing, with every heal performing a given buff tacked on. Thresholds for crowd-control might be used, with a given window of time whereby the controller is more susceptible to damage as a cost of controlling. Mixing and matching, sub-division, etc.

The game you make can, and should, include this sort of thinking if the aim is to create a piece of work that inspires the gamer. Establish the very basic, most fundamental aspects of any system you wish to incorporate. Identify their relationships. Pick up the paintbrush, armed with knowledge, and then begin work on the masterpiece. Don't start with a given system that one might feel is bloated or unnecessary, and cut it arbitrarily. Don't add to a system that is complete just because of a feeling to always add more.


The game we all seek asks for us to fill a needed role within a group (a future post about how solo-advocates vs groupers). How that might manifest would be a design point. But please, designers out there, don't marginalize roles nor attempt to remove them. Breath life into them, and their interdependence. When a player mentions that 'something is missing' it's because there is. This series outlines where and how to not have the combat portion of your game lacking to any player.

Previous posts in this series: combat fundamentals, combat derivatives, combat derivative derivatives, and how it all comes together.


  1. You know, you've still not proven why trinity is mandatory, as in specialized roles. All the things you mention are perfectly doable without the need for specialization and specialization, by your own words, limit the choices of players.

    Take WoW, which actually has done this right at one point or another (it just hasn't done it right lately). It's a really good example of what happens when you being your game with pretty narrow class specializations. For example, 90% of tanks in vanilla were warriors (90% is arbitrary, meant only to demonstrate the overwhelming popularity of this class in this role). Did specializations into the roles of tank/healer/dps/buffer/whatver work well at that time in the game? The truth is, it's worked better than any other iteration of the game as encounter design and group balance go.

    Now what happened when we tossed hybrids in *with* the specialists? I'll tell you what happened: there's was lots of debate (and still is) over how good a hybrid should be in any one of his specializations, since it's not a "pure".

    So the *real* problem was/is that hybrids haven't been designed into encounters. It means that encounters depend on there being 3 specialists: tanks/healers/dps. Therefore, players have less choice in how to play and how to approach encounters. The encounter dictates how the players must play.

    If you're arguing that specializations give more choice, you'll have to do better than perfect charts. They simply don't address the reality of how players play.

    And this is why I'm firmly on the side that specializations aren't all they're chalked up to be. Encounters could be far more interesting and dynamic if they could be approached from a different angle. The presence of the trinity makes it quite difficult to do this, but certainly not impossible. I've pointed this out previously: in WoW it's already been done. High King Maulgar and all other similar encounters prove the point. You absolutely don't need a tank/healer/dps and in fact your encounters become a lot more interesting when this strategy isn't revealed to you.

  2. You've still not proven why trinity is mandatory, as in specialized roles.
    Perhaps dragging the point I wanted to make through 6 posts meant missing a detail where the proof was concerned. In post 1, I chiseled out why the trinity functions are the fundamentals to combat. I then proceed in posts 2 and 3 to explain the derivatives and how they might manifest. Then, I apply this knowledge of roles to the lowest common denominator of its manifestation (the individual) in post 4, taking note on how an edge is gained compared to the competition. In post 5 (the group setting) I underline how this edge can be permanently maintained through the functionality of individual specializations. I have used the example of vehicle stability, maintained integrity of what the limits on a given individual should be able to perform. I demonstrated how perfect specializations (just like perfect hybridization) is a hindrance compared to mild hybridization around the fundamental (and support) roles of combat.
    Now, I admit there isn't a logical jump between roles and specializations other than their simultaneous occurrence, and thus the natural trend to apply the two together. I merely argue that the roles exist, and specializations are a tool to the userbase which allow for more treacherous encounters to be tackled. So, then I ask, why not marry the two concepts?

    Now what happened when we tossed hybrids in *with* the specialists? [...] how good a hybrid should be in any one of his specializations, since it's not a "pure".
    I feel I've addressed this squarely with my illustration of 3D 'areas of influence'. I considered illustrating in this post what happens when you mix pure specialist individuals with pure hybrid individuals, but the result of the combination generation was too similar to pure hybrid, so omitted it because of redundancy.
    I will post how hybrids can live up to their true potential, but already my list of topics I want to touch on grows much faster than my time allotment to blog =P

    The encounter dictates how the players must play.
    This is a truth! Perhaps I should have more explicitly mentioned this, but I'm glad you bring it up.
    At its most bare, it's the individual(s). The content can always be reverse engineered so I must start my understanding with the first thing a player encounters: class/role/hybrid/spec choice. This is done against a generic, multicolor content backdrop. From here, we obviously progress through a game as (re)actors to design biases for the content we encounter.

    Encounters could be far more interesting and dynamic if they could be approached from a different angle. The presence of the trinity makes it quite difficult to do this, but certainly not impossible.
    Again, I was not speaking to the content, and used generic applications of the simplest forms of combat, building out from a core. The trinity (and its derivatives) are the core from which to build. If the designer applies them in a formulaic way, then we get rubbish for games. If the designer takes that structure and breathes life into it with the content they generate, then we get wonderful experiences. But this has nothing to do with the structure in which gamers operate, this is a content design issue.
    Also: content should ALWAYS ask the player to think outside of a comfort zone. Allowing the player to stagnate in their approach flexibility is a designer failure.

    You absolutely don't need a tank/healer/dps and in fact your encounters become a lot more interesting when this strategy isn't revealed to you.
    Well first off, strategies should *never* be revealed to the gamer. Public PTRs etc are atrocities.
    I'm curious to learn why you think HKM doesn't require a t/h/d. I've my suspicions as to why you think it, and if I'm correct, I think there's a disconnect I can identify between my attempt at proofs and your understanding of my attempt.

  3. I guess I just remain unconvinced :)

    I don't mean to seem as if I'm not getting your points. I think it comes down to I don't agree and in any case I don't think I'm isolating the concepts in the way that you are for this series. I'm thinking of the whole, so for me clearly we're talking about roles and classes because we want to know how they translate into good gameplay.

    As to the question of HKM, it's not that the shaman, hunter, mage, or whatever you're using isn't tanking. It's that tanking isn't the primary problem. In this encounter, if you're clever enough, you avoid 90% of the damage. This encounter was designed to require players to think outside the box a little, rely less on *who* is in their group and more on *what* each player can do to help bring it down.