Designing a Magic System - Part Two (Choices)

(You'll probably want to read part one before starting this).

So if suspension of disbelief is the key to the designing reasons that magic could exist in your game world, what is the key to designing the game play role that magic should have? In Unangband, and I'd suggest in the design of every single player magic system, the answer is choice.

A key component of the Angband game design is the inventory limit - that is the hard limit of 23 different types of items that you can carry at any one time. From a realism point of view, limiting the player to carrying 99 potions of the same type, but using up a second inventory slot to carry an additional 1 potion of a different type makes no sense - but when considered from the paradigm of limiting choices, it fits perfectly.

The sum total of choices you can make every turn in Angband, in fact in all roguelikes, is the number of possible commands not involving equipment, plus the number of different ways you can use each item in your inventory. The importance of having a range of choices is even more pronounced when you realise the three goals in the game:
1) Stay alive
2) Kill Sauron
3) Kill Morgoth
At every turn, you must make the choice that ensures 1, and maximises your chance of 2 and 3 occurring at some point in the future. And to make the choice, you must have the choice available to you. A wide range of equipment types in your inventory increases your possible choices, and consequently increases the chance of having the correct choice available to ensure you stay alive.

As a warrior in Unangband, you have the least choices available of any class. You can move, attack, use potions and scrolls reliably, and have a slim chance of using any other item in the game. Playing a magic using class on the other hand greatly expands the number of choices you have, because each book can contain up to 26 spells, any one of which you can pick if you have learned the spell. This means, depending on your spell selection, you significantly increase the number of actions you can choose to do at any turn. The sacrifice that the magic using classes make is to trade these increased choices by reducing their effectiveness at the most straight forward choices of move and attack, and increase the overall risk at every game turn (by having less total hit points).

A class based game is uses this warrior vs. magic user as an indication of preferred playing style: a few, shallow, easy choices vs. many, deep, difficult choices. The disadvantage of class based games is that they require considerable play to determine the capabilities of the class - choosing a class is a decision you're forced to make in advance of knowing the costs and benefits of the decision. In a skill based system, not only do you have this problem, but you're also burdened with the fallacy of equivalence: the game design has to somehow equate one point of each magic using skill with a point of melee skill (or worse, figure out the relative 'cost' of each). In a classless, skill-less game, you can easily choose and discard inventory choices as you encounter new items - additionally you are not penalized by having items generated that you can never potentially use. This particularly affects magic using classes from the Dungeon & Dragons tradition, who are restricted from heavy armour or sharp weapons for seemingly arbitrary reasons.

Note that choice is not a necessary component of multi-player game magic systems. Multi-player games are usually about doing one thing and doing it well. In World of Warcraft, as an example, this consists of the holy triad of tanking, healing and dps. Tanking involves taking damage from the main monster in an encounter by manipulating the monsters attention (aggro) so it always attacks you, healing involving keeping the tank alive, and dps involves damaging the monster as fast as possible in a way that prevents it attacking you instead of the tank. A tank that can do two other things, at the trade off of not tanking quite as well, is almost entirely useless. This is the problem of hybrid classes that John Hopson examines in MMO Class Design: Up With Hybrids - An Economic Argument. His basic argument is that the tanker - healer - dps triumphvate is so effective, that you have to engineer specific encounters in the game that force hybrid classes to be useful by breaking up the ability of this grouping to work.

World of Warcraft emphasises the choices you make in the 'build' of your character prior to combat, instead of the choices you make in combat. Eve Online operates in the same way, as James Lantz of In Machinam explains in The Machinery: Eve Online. Eve space combat usually features a lead pilot 'calling out' targets in sequence, with everyone in the same combat group / fleet / wing targetting the nominated target until it is destroyed, before moving onto the next. This mechanic relies on the accumulation of overwhelming firepower to bring targets down, with very little tactical choice from individual participants.

One vs one multiplayer games are an exception to this ganging up process that demands 'excellence in a single activity'. David Sirlin advocates an asymetric rock-paper-scissors approach as a means of encouraging yomi - that is the state of knowing what your opponent will do next and countering it. In asymetric rock-paper-scissors game, the game plays the same, but the pay-off is different for each winning move: $1 for rock, $2 for paper and $10 for scissors, for instance. This means that you always have a low risk, and high risk strategy - should you go scissors for more money, or rock to counter the high-risk strategy? A good opponent should be able to take advantage of this concept of yomi to predict what you'll do. But the theory of mind behind this prediction strategy breaks down when you are not able to determine your opponent reliably, or you encounter them infrequently enough not to be able to learn their playing strategy and take advantage of it.

In this series of articles so far, the lines between 'magic system', inventory and technology, and abilities / talents / skills (call them what you want) has blurred. I'll attempt to more formally define what I mean by magic system in part three.