Designing a Magic System - Part Twelve (Progression)

(Although this reads as a standalone article, it is part of a larger series on Designing a Magic System. After reading this you can go back to parts one, two, three, four, five, six, seven, eight, nine, ten and eleven).

I feel like I've stumbled into a justification for classes backwards: dismissing arguments about using skills, talent trees or other game mechanics without preparing the ground as well as I should. Stephen Totilo recently argued that the defining characteristic of role-playing games is playing a role; and by that definition, included LittleBigPlanet, Guitar Hero and Spore as great role-playing games he had played recently. I am forced to disagree with Stephen's definition: while semantically correct, he is disingenuously expanding the computer genre to encompass most, if not all games, because with the exception of the Eye Toy and 'Brain Training' style self-improvement games, the player is forced to interact with the game through the proxy of an in-game avatar or character. What defines a computer RPG is progression - and at the same time, there is nothing more greatly abused in RPG design.

The grind, the treadmill, leveling up, mudflation, scaling difficulty of opponents, are all tropes of RPG progression and, as you can tell from the negative connotations of many of those phrases, few concepts have had more loathing heaped upon them, particularly in the MMORPG space. Progression, in the worst sense, can be the crutch of a lazy designer: Progress Quest typifies the ever escalating scale of identikit enemies and equipment, in which no distinction is attached to the ever increasing numbers. But the RPG progression is powerfully compulsive and increasingly adopted by other games: achievements feature increasingly in other games such as first person shooters, unlocking additional weapons, equipment and game types.

What do I mean by progression? There are at least two distinct types of progression in computer games, which I’ll label player progression, and character progression (narrative progression is arguably a third). Player progression is the increasing aptitude of the player in mastering the game: whether through learning and understanding the technical rules of the game (surface play) or the implications of those rules (deep play), better control over interaction with the game (hand-eye coordination, clicks per second) or rote memorization of in-game patterns (short term and long term memory). Player progression, while a fascinating topic in its own right involving human computer interaction, theory of game design and psychology, is not the focus of this article – instead, the poorer and often abused step-child character progression, is the intended target.

Character progression is the unlocking of additional rules of play, or altering the existing rules, by choices or actions within the game. The most common unlock is the ability: an additional in-game interaction that the player’s avatar can choose to do. But scaling upwards existing abilities is just as common in the RPG space. And sometimes, particular once the whole set of game abilities has been unlocked, or at the conclusion of the game prologue, abilities can be removed – usually through the convention of capturing the character, or having them narrowly avoiding death.

Bound up with the concept of progression, particularly in the RPG space, is choice. As the character progresses through the game, the player may be given the opportunity to choose which of several abilities to unlock or scale up. A fascinating alternative suggested here would be to create a game featuring inverted progression: where the player is forced to choose which of a starting complement of abilities to lose as they move forward in the game.

The difficulty with choice is that it makes the game design harder, as the designer is forced to provide alternate solutions or balance game-play for each choice or combination of choices that the player has made for their character’s progression. This can be mitigated by ensuring that the different choices have limited or no real consequences to game-play, simply coloring the in-game aesthetic, but this trade off can make the choices less interesting to the player. Spore has received much criticism for this decision: while the many in-game design tools allow a fascinating array of different creatures and objects to be created, the vast majority of creative differences have no impact on game-play. Only the mouth part, which dictates whether the creature is a herbivore, carnivore or omnivore, and a limited set of abilities in the creature phase provide real customization options.

Character progression can be neatly tied into player progression through the tutorial phase of the game. In this phase, the player is only given a limited subset of the total game abilities, and has to demonstrate mastery of these abilities before unlocking more. The tutorial phase extends until the full set of abilities is mastered: at which point the ‘real’ game begins.

The difficulty with the tutorial phase is that is implies a linear progression of abilities which does not sit easily alongside the choice component of game progression. The tutorial is usually gated, in the sense that the player cannot progress until mastering the ability, which may be outside their game play aptitude – leading to frustration at the tutorial element. And the pacing of the game may be affected, particularly if the tutorials have to be interleaved with the larger game play. The Zelda series of games are good examples of mixing tutorials and game play elements: it is possible to explore the overworld at almost any stage, but mastering of abilities gained at specific times is required to unlock dungeons which further test these abilities. The whole game design supports this character progression – it may not be appropriate for genres such as real time strategy games or first person shooters, where mixing intense action and tutorials may result in game play pacing problems.

If the elements of the tutorial are interesting enough to be expanded to a full game, it is possible to control character progression through introducing new abilities in later game levels or higher difficulty levels. This divides the game play up into sub-games, each of which is of increasing complexity; allowing the player to master simpler strategies before moving onto the later levels. Darwinia features this progression technique through each of its levels; and many Real Time Strategy games take this approach in their single player campaigns. The difficulty is to ensure that abilities introduced early are still relevant later in the game, and that the game is still interesting even with the more limited ability sets: otherwise levels will end up with a mismatched difficulty or under utilization of skills learned earlier in the game.

Abilities can be also unlocked once a player has mastered a particular section of the game, to make re-traversing the section less of a challenge. The early 2d Metal Gear Solid games featured this, where an initially unarmed Snake would be forced to evade guards using stealth, but after equipping himself with weapons later in the game, he could shoot his way through the same screens far more quickly. This is not just limited to geographic traversal: if a player in Resident Evil 4 has difficulty with a particular boss monster, they can purchase a one shot rocket launcher at considerable cost to bypass the monster in question, and in Spore the initial difficulty of fighting enemies in the Space phase is much easier once the player has acquired higher level weapons, made available by defending sufficient attacks at the more difficult early stage of the phase.

The acquirement of abilities can either be directly linked to actions within the game, which results in a puzzle-like structure to game play, where certain prerequisites have to be met in order to open up or make easier later parts of the game, or indirectly, by providing a resource that the player can then spend on abilities directly or indirectly. Classes, talent trees and skills are all mechanisms for controlling character progression in various ways and guiding player choices as to which abilities to acquire or improve. The class structure is the most limiting framework, where the player makes a single decision, usually at the start of the game, that fundamentally colors the game experience. Talent trees and skills allow smaller, incremental choices to be made - the difference between the two being a matter of degree rather than kind, where talent trees implies a few, spaced out decisions as to which abilities to acquire, and skills implies a more frequent investment of time into the decision making process, with abilities improving on a scalar basis with the occasional break point which introduces a new ability or opens up a new skill.

Divorcing the ability acquirement structure from the game simplifies the design requirements: you are no longer forced to ensure that the parts of the game where an ability is required fall after the parts of the game where that ability is acquired. However, the mechanic for acquiring new abilities becomes more important, and here it is very easy to end up in a position where the mechanism (experience points, money, power ups) is available without bounds in a region, even if a fraction of what is available later in the game. It then becomes possible for the player to continue to acquire new abilities in a low risk environment by trading off time instead of playing skill: in other words, to grind or farm the game. This implies that you should put a ceiling on the acquisition resource, and regularly change it as the player moves through the game and masters sets of abilities ('Your dubloons are worthless here, you need McGuffins to buy things this side of town'). Spore's achievement system implicitly does this by making the rewards for defending planets different to the rewards for terraforming them: and each reward set makes the particular task that contributed towards it much easier in future. It is easy to visualize a game of tiered abilities where each new tier requires a new resource and only a limited number of the possible abilities can be learned per tier.

But none of this answers why, as designers, we feel the need to reward the player with new abilities moving forward in the game: particularly when those abilities make parts of the game easier or irrelevant. Is the progression of player skill and narrative not sufficient to inspire the player to keep playing the game? Does the game lack depth or complexity that we should trivialize it by making the game easier and easier, as opposed to harder and harder, the more the player plays it? Or are the rewards we give purely cosmetic, the enemies scaling up as fast as the player does, so that the same sword swing at level 50, despite the gleaming blade, and cacophonous impact, change the world in only the same way that the timid stab of a level 1 character against a giant rat does?

To an extent, we are making the game more complicated, by providing more choices through progression. But when the range of choices are mastered by improved player understanding, is game progression a dressed up Skinner box? I think, and unfortunately many games shy from this, that progression must mean that the stakes are higher as well. You must risk more when you fail, which is why permadeath in roguelikes is such a powerful solution to the progression dilemma. Of the classic games, Chess and Go don't feature progression in the sense that RPGs do: the one to come closest is Poker, where your stake is built up as you play the game, and those at the table around you withdraw. In the final hands, if you manage to stay at the table, the money you have staked is the highest, the rewards the greatest and the failures the most painful. Mat Williams compares Hinterland to Poker, and although he doesn't say directly, he must have been conscious of this acquisition strategy and how risking all is the direct counter to progression's woes.

More to come in part thirteen.