Concepts Overview


Possibly the most constricting legacy that an RPG today can inherit is that of "classes" or "professions." While some players would obviously prefer to have character customization be simple, and grouping all of a certain type of skill together and making advancement take place automatically in discrete steps certainly accomplishes this, forcing the player to choose a single class to focus all of their avatar's development in is extremely restrictive and tends to produce assembly-line characters, all alike but for the name, and players who want to be original or unorthodox aren't given the opportunity.

Too often, a set of classes conceals a certain degree of laziness on the part of the game designers, because it tremendously simplifies the character generation system (and therefore interface as well) and game balance. This is not always true, and some games even go to great lengths to provide "multi-classing," allowing you to combine diverse sets of skills, or additional disciplines or specialties. Sometimes these can increase the effective number of classes by as much as an order of magnitude, which is probably all you'll ever need for a single-player game, but for a massively multiplayer game, you need to prepare for a player base that numbers at least in the thousands, often with several characters per person, and it would be nice if every character were unique.

Standard D&D six-race, six-class RPG's give you a handful of basic character templates, forcing you to simply pick one, which makes character generation rote. Multi-classing, specialization, and other systems of that ilk tend to turn character generation into a science, as people find the "best combinations" and populate half the world with a small handful of archetypes, and even if such a system was perfectly balanced, there are still only so many basic combinations. Our goal should be to make character generation into an art.

The problem is that features like multi-classing or specialization tackle the problem from the wrong end, taking a flawed system and trying to patch it up. As any programmer can tell you, if you want to make two versions of a program and intentionally make one of them worse than the other, the correct approach is to write the better program first, and then water it down to get the lesser version.

With this in mind, let us eliminate the idea of a class altogether. Allow players to select any skills they like from any disciplines, raise and lower attributes in any combination they like, pick any special advantages or disadvantages we choose to make available straight off the master list. While we're at it, let them customize their starting equipment/cash as well. Heck, even let them trade off between starting with good equipment and starting with good skills. Happily, this also lets us reserve the word "profession" to mean exactly what it should have meant in the first place: not your education, but your job.

For the players who would rather just click two buttons and start playing, construct a list of archetypes, which are valid character designs that players can use as they are or use as a template to modify if they don't want to build a new character from scratch. You can even let players choose sub-templates if they want something more specific. Also, put in an option that allows players to specify a margin of uncertainty when designing a character (usually for when using an archetype) that will randomly lower a few attributes and skills and raise some other ones, so that they can start the game with a unique character even if they don't want to create it themselves.

This, of course, creates two obvious problems. First, it creates an obvious dilemma for balancing the game and for roleplaying in general if players can be very good at only very narrow skills in vastly different disciplines (for example, attaining mastery of swords, but not daggers, or regenerative fire-elemental underwater alchemy without the use of any other alchemical skills). Second, if players can create characters with excellent equipment but poor skills, what is to stop them from creating such a character, giving all their money/equipment to a friend, and then deleting the character? However, I think we will leave these questions for the drawing board.

Further details on the character generation system are yet to be worked out.


After inflicting the injustice of archetypical classes upon the player, the next logical sacrifice to the goal of simplicity is to put not only character generation but also character advancement into a small number of discrete steps. Not only are the avatars stratified into fighters, thieves, wizards, and priests, but they become a level ten fighter or a level six wizard.

The problem is not merely that people know exactly how "experienced" a character is at all times, or that this experience is divided into very coarse, discrete steps called "levels," but to add injury to insult, the ratio of advancement in various areas is usually fixed for each class. That is, a level X character of class Y will always have the same amount of health, attack power, etc. (possibly within a tiny range of variation). Not only does this exceed all bounds of reason or realism, it is downright boring. Even if our imagination were poor enough to allow us to suspend disbelief, we should not want to.

To make matters worse, "experience" rarely bears any connection to . . . you know . . . experience. You are almost never rewarded with advancement for varied experience, rarely rewarded for your exceptional accomplishments, and usually not even based on the difficulty of combat, the obvious minimum. For example, you generally get the same experience for defeating 3 monsters whether you fight them one at a time or all at once, even though the later is obviously far more difficult.

And of course, the absolute quintessence of poor design in the character advancement system: you don't even get experience for the fight at all; you get experience for the kill.

Requiring you to be in combat to get experience already seems awfully unreasonable. To award it all upon the deathblow, often to the person (or group) who strikes the deathblow, is inexpressibly tacky.

Well, how should experience be awarded? The answer is simple: advance the character the same way that people advance in real life! You get better at a skill by practicing it, and you improve a natural attribute (like strength) by using it. In both cases, the greatest gain usually occurs when you go right to the edge of your abilities, because if you do something that is already easy you are not pressured to improve, and if you do something utterly beyond your abilities your body or mind tends to just get hurt, not practiced.

In other words, you get better at fencing by fencing, and you get better whether your opponent dies at the end of the fight or not. You do not get better at fencing by hacking a dumb beast to pieces while it has its back to you so it can strike someone else, or at least not much. Sitting down to rest while an ally who is "in your adventuring group" fights a few battles doesn't get you any better either, nor does sneaking in in the last five seconds of a major battle and dealing the finishing blow to a poor wretch who is basically dead anyway.

You get better at casting a spell by casting it, not by killing enemies with your dagger. A healer can become masterful by working in a hospital without ever getting near a fight. Of course, they wouldn't become better at fighting, just at healing.

If you want to get better at defending yourself, you have to let someone attack you. Furthermore, your body won't get any more tolerant of physical punishment when the enemy can't get through your armor, whether that armor is physical, magical, or anything else -- even whether it is your armor or the work of a benevolent ally is irrelevant, because to the extent that it protects you, your body does not need to cope with the attacks being dealt to you, and therefore does not adapt to them.

Also, your skills need not advance in proportion to how they start. If you run around with a sword but never cast a spell, you will never get any better at magic, even if you put all your skill points into magical abilities during character generation. In fact, when you fail to use a skill for a long time, it should even atrophy! Use it or lose it.

For mechanical (and aesthetic) reasons, it would probably be a good idea to also give the player some sort of advancement points that they can distribute themselves, representing their private studies or training. This allows players to round out their abilities, makes it easier to get a toehold in new skills they want to develop, and is simply fun. However, skills that are used frequently should still be cheaper to raise than those that are never used.

And there should be a check box in one of the options panel that says "automatically spend skill points to counter atrophy of unused skills."

In addition to making character advancement much more sensible and customizable, instituting such a system would tend to eliminate such practices as "power-leveling," because experience is actually based upon how hard you have to fight. How do you "power-level" someone if they only get experience when they attack or when the monster attacks them, and to the degree that you help or protect them, you steal from them the experience they would have gotten from the fight? The best you could do would be to help the person find suitable fights, and save them if it looks like they're going to lose. This might be a little safer, but it certainly wouldn't be noticeably faster than just playing the game.

Further details on the character advancement system are yet to be worked out.


In most massively multiplayer games, players never really die -- they just lose anything they were carrying, and usually a bit of experience, and reappear back in some safe area. In the really forgiving games, everything you were carrying stays on your corpse, and no one else is allowed to take it unless you give them permission, so you can usually even get everything back.

On the one hand, this seems like a rather puzzling phenomenon from a roleplay perspective, and quite frankly the fact that none of the NPCs ever seem to notice this can get kind of creepy. On the other hand, really killing off a player's character, after they've devoted a lot of time and effort in that character, seems pretty cruel.

As far as I know, the only game that has tried to incorporate this phenomenon into the backstory is Shadowbane. I think they've done a pretty good job, but it seems like they've quietly backed away from this explanation in a few details of the game (for example, if your hired shopkeeper gets killed, you don't get him back). This also creates some difficulties for the storywriters, because they can't kill off any characters. Also, for obvious reasons, such a state of affairs can't really exist for very long, or the world would become overpopulated pretty quickly.

So, setting a pattern that you will recognize throughout the rest of this game, we're going fill a gap in the story by expanding it into a major aspect of the game. In so doing, we're going to try to arrange things such that players can keep their characters, but people can still die.

As a preliminary measure, people generally fall unconscious before they die, and normally only die if people continue to beat them while they're down. Few monsters have the desire to do this, and few criminals will risk staying around longer, since once you're down they can steal your valuables and run. Thus, it is uncommon to die; however, we obviously cannot count on this to save us.

When a person (whether a player or not) dies in Magi, they leave a corpse and enter a region called -- logically enough -- Death. Death is a sort of river, with a strong current running away from life an deeper into Death, and most things, when, they die, and sufficiently weakened by the experience that they are quickly swept far into Death. Once inside Death, it isn't easy to get out, for several reasons. First of all, you have to constantly battle this current that is pushing you deeper into Death, which (as a general rule) gets stronger the further into Death you go. Secondly, there are a series of Gates in Death, which can be easily passed going in, but which are hard to get through going out (for example, waterfalls). Each successive gate is a more formidable barrier. Nothing that has passed through the ninth Gate has ever returned.

Finally, there are a lot of other things in Death that also want to get out, and the easiest way for them to do this is to steal whatever life the weaker beings around them have left, plunging them further into Death. Consequently, the majority of beings that enter Death quickly fall past the ninth Gate. However, if you can get out of Death relatively quickly, you are free to resume your life where you left it (although there is nothing to stop people from raiding your corpse while you are in Death). After a prolonged visit to Death, if you manage to make it out, you become one of the Undead.  Death has a permanent hold on the Undead, and they need to continue to steal the life from other beings to hold Death at bay.

But even if you pass through the ninth Gate, your consciousness sometimes still escapes, becoming what most men call a Shadow. Shadows are beings of pure ether with no physical (or even metaphysical) bodies and little direct influence over the world, though they are very susceptible to the minor, everyday magics of common folk. However, they can still commune with wizards and magi.

(SIDE NOTE: Players may begin the game as Undead or Shadows if they so choose.)

Let's review. People who lose a fight usually go unconscious without dying. People who die can escape from Death, especially if they possess necromantic skills or are aided by a necromancer. If they fall past the ninth Gate, they can no longer escape, but they are given the option of playing as a Shadow.

Alternatively, when a player passes the ninth Gate, they can simply create a new character. The new character must be given a new name, and cannot start particularly near the place where the last character died, but will be given comparable skills and equipment (players will lose about 10% of their acquired abilities and wealth as a penalty for dying), and furthermore, this gives them an opportunity to rearrange their skills and attributes for a different play style -- but it must be different. Character designs too similar to the deceased character will be rejected (but, of course, the player could start from scratch with any design they please, if they prefer).

While player avatars in MMRPGs generally aren't the heroes who star in books (they're not the wizard who fell through a rift in space from another dimension, they're the wizard who just graduated from the magical academy among a class of 500), they are still exceptional individuals (they graduated in the top 10% of their class). They will therefore stand a much better chance of making it out of Death (if they work at it) than most NPCs. However, because they can really die, they need to be careful.


On a related note, we need to do something about the hordes of monsters that are routinely slaughtered by players and "respawn," often while players are sitting next to the spawn points waiting for them. Obviously, this is a tad unrealistic, but what we need in this case is really just a smarter computer.

First off, spawn points should move. So should monsters, for that matter. You shouldn't always find some creature in exactly the same position. Animals scavenge or hunt, bandits maraud, guards usually patrol, and even door guards change shifts. New creatures should be able to appear anywhere in a predefined suitable area, under certain restrictions:

  1. No creature (possibly with a few exceptions amenable to the lore) will spawn within line of sight of any player.

  2. Creatures breed, they do not spontaneously generate. There must be another creature of a similar type in the area.

  3. This should take time. Game time is different from real time (see below), but spawning should take at least weeks or months of game time, not hours.

This means that once an area is clear, it will stay clear until some creatures migrate into it. Also, no player will ever actually observe a creature spawning.

Some adventuring party storms through a cave and eliminates all the bats there, and there are no bats in nearby caves. When they come back the next day, they had better not find it full of bats again. Instead, they should discover that the caves have been taken over by a gang of belligerent imps from the nearby forest.


One of the most difficult issues in any real-time game is the problem of controlling the flow of that time. At some times, a fast pace is desirable, because we do not wish to inflict upon the player the boredom that sometimes occurs in real life. At other times, it is actually preferable to have a slower clock than in real life, because the interface between human and computer is nowhere near as good as the interface between mind and body, and if things happen too fast the player won't have time to react. This is made particularly difficult in a multi-player game, where players have to be able to interact with each other, but often should really be moving at different game speeds.

If you wish to have any sort of overarching metaplot, the average speed of life has to be fairly quick, because there is simply never enough detail to any plot to be worth playing through the entire sequence in real time. Consequently, most MMRPGs use a very quick time lapse, generally in the range of 1-5 minutes of real time for every hour of game time. Of course, this means that a one-on-one duel between powerful beings can easily take upwards of half an hour of game time. Spells can take several game minutes to cast, and I've seen a few that take days of game time before you can cast them again. Also, apparently a master swordsman is only capable of swinging his blade once every few minutes.

The solution to this dilemma is so simple and so obvious, most people fail to think of it or, having thought of it, dismiss it immediately. Let time vary from place to place!

This is by no means an original idea, but no one (to my knowledge) has tried to apply it in this context. Time travel and other forms of time manipulation have been a part of fantasy and science fiction for a long time; according to Einstein, clocks tick at different rates in different places even in real life, apparently without catastrophic results. Why it has not yet been tried here, I cannot say.

So let us try the elementary solution: in high-action locations (such as battles), time passes slowly, possibly even more slowly than real time, giving the players time to think and act. Given a good interface, this will allow players to minutely control their avatars' actions, letting them aim their attacks, issue commands to dodge or block specific counter-attacks. The larger the battle, the slower time goes. Inversely, in low-action locations (such as walking down a road between cities, with no living animal anywhere in sight), time passes quickly, allowing you to reach your destination and get on with the game.

The sun, of course, is nowhere near these low-time zones, and continues to move at the same rate. Therefore, day and night are still uniform throughout the world.

Conveniently, this also reduces the strain on the computer, because the scenes that are being slowed are those that would require the most calculations. Consequently, the technical requirements for running the game drop.

All that remains is to provide an explanation for this rather strange phenomenon of time dilation. For inspiration, let us look at how time dilation theoretically works in the real world.

According to Einstein's theory of relativity, there are two things that can cause time to slow: very high speed (significant fraction of the speed of light), or a strong gravitational field. High speed is obviously not applicable.

What causes time dilation in the presence of a gravitational field? Again, according to theory, the gravity warps the "fabric" of space-time, stretching out time. I can't provide a more technical explanation than that, as I am not a physicist, but try to visualize it: space gets all bundled up at a point, and this density hinders the passage of time. The way I think of it (which may or may not be in line with theory; we are, after all, only worried about how this will work in the game), time slows down when it has to go through a thicker medium, in the same way that it is harder to move through water than through thin air.

Fortuitously, there happens to already be a universal medium in the game, called the ether, which is where magic takes place. Prophetically, the word "ether," though it has been used in relation to magic in many other games, was the name for a hypothetical interstellar medium in the real world (which accepted scientific theory now says really does not exist, due to the failure to experimentally measure it). So let us say that time slows down as the density of the ether increases.

What controls the density of the ether? Well, as you recall from the previous section on Death (see above), when a living thing passes through the Ninth Gate of Death, sometimes its consciousness escapes as a Shadow -- a being of pure ether! Also, Shadows have little influence in the material world, and so, unable to do much, they will have a lot of time on their hands as they wander the world endlessly. It stands to reason most of them would seek out interesting events to observe, which means they would be drawn like magnets right to . . . high-action locations! Because they are made of ether, a high concentration of Shadows in an area will significantly increase the density of the ether in that area, slowing the passage of time.

A second factor is equally convenient. Ether is the medium through which magic (and also wizardry) operates; consequently, ether will flow to an area where a lot of magic (or wizardry) is operating. Since high magical activity will probably correspond pretty closely with major developments in the metaplot, this not only provides an excellent excuse to provide time-manipulating abilities in the game, but also allows the writers to slow time down for important events even if very little action is taking place.

Convenient, eh?


Let's face it: collecting spider eyes for the master of the wizard's guild isn't a quest, it's the kind of menial task that people play a game to get away from. Bring monster scalps to the guardhouse for a reward isn't a quest either, it's just bounty-hunting. If this is the kind of time-wasting chore we're going to call a "quest" in our game, we may as well forget quests altogether.

Don't get me wrong -- things like bounty-hunting have their place in the game. In fact, that makes a lot more sense as a source of income for a soldier than taking industrial-manufactured platinum coins off of the corpse of a wicked demon tree, and I see no reason why players who want their characters to bounty-hunt for a living shouldn't be provided with an effectively limitless market for their services by the NPCs. But that's such a trivial aspect of the game as to be hardly worth discussing.

A quest should be more along the lines of finding a poverty-stricken bum on the streets who blames his predicament on the death of his beloved wife and child at the hands of an evil monster three years ago and wants revenge. You should have to ask around in several different towns to determine the location of the monster (and possibly some clues as to its weakness), chase it across half the continent, finally corner it and slay it, then return and (possibly) get some sort of material reward (obviously, you can be given some sort of "adventure point" character advancement reward for finishing the quest). When someone finishes the quest, the guy no longer needs revenge and the monster no longer exists, so the quest should be removed from the game and replaced with a new quest, rather than offered to the next hundred people who pass through the town.

There should also be quests and events that influence the course of history. Wars should be fought, characters should occasionally face besieged cities or potential conscription by a less-than-benevolent government. Nomadic barbarians should occasionally stumble upon civilization and burn a city to the ground.

Speaking of cities, players should be able to construct buildings, build up or conquer cities, start their own kingdoms. No reason that the players shouldn't be given the opportunity to write a piece of the story themselves. Guilds should be real, solid organizations of players with political significance. There should be many hirable non-player characters for various tasks, from mercenaries to architects to blacksmiths to innkeepers.

Who needs a bank? Store your extra items in a house. Buy a good lock for the door, don't advertise the presence of valuables, and put the house in an area with a good police force and it is unlikely a thief (player or otherwise) will take your possessions. After all, if they get caught, they may face harsh punishments, from long-term imprisonment (unless they can find a way to escape) to death (which, in this game, has become a big deal).

Most importantly, there should be an overall story to the game; a metaplot, if you will. It should be the kind of thing legends are made of, and as many players as possible should be involved in some way. For this game, it will center around one particular Magi, whose coming was foretold . . .

Player Rivalry

Generally, people think of two possible ways player combat could be set up in an MMRPG. The first is to outlaw player-vs-player (PvP) combat entirely, never letting one player hurt another. This tends to breed the kind of jerk who will do anything they possibly can to take advantage of the system, knowing that they cannot be killed even if every player in the game would give their right eye for the chance to pinch his arm. The second, which is generally the case any time PvP combat is allowed at all, is that the majority of players get just powerful enough to prey upon those less fortunate than themselves, and anyone who travels alone is in great peril of being slaughtered by a complete stranger for a few copper coins.

Naturally, I am not very fond of either of these paradigms.

Rather, I would prefer that the risk involved in attacking another player is great enough that, though you may make such an attack at any time, common sense prevents you from doing it lightly. People who become known for robbing innocent travelers should be hunted down and brought to justice. If you are forced to kill for some reason, but others are unlikely to think the killing justified, you should have to flee the area. In short, what we should do is not to make it impossible to commit a crime, but rather to institute an in-game criminal justice system.

Of course, players (especially those who acquire political power) will have a hand in this. It makes little difference how much the NPCs hate you if other players decide to track you down and kill you. Players running their own cities will need to hire their own police force, build jails if they intend to have imprisonment as a possible punishment, etc.

The possibility of one player imprisoning another's character creates problems too, but that's more of a technical problem than a conceptual one, so it shall be relegated to the drawing board.


In many RPGs, all characters have a certain amount of mana, or "magic points." All spells and other magical skills consume a certain amount of mana, and if you currently have one point less than the spell requires, you cannot cast it. In contrast, physical fatigue, if it is measured at all, is usually not expressed numerically (though obviously it is internally stored numerically) and rarely limits your activity. Basically, any physical skills can be performed without limit in virtually any RPG. To compensate, designers generally feel it is necessary to have an appalling small number of physical skills available to you, and to make them all very similar in power and effect.

Hence, the typical distinction between boring (physical) classes and defenseless (magical) classes.

As a general rule, I like to ignore strict limits. I find that usually a law of diminishing returns is more effective. Besides, this is not remotely how fatigue works in real life, or even in most books.

I'm not sure why game designers decided to place an absolute and abrupt ceiling on the total (effective) fighting time on certain characters; I can't think of any particular advantage to this approach. Presumably, it sort of developed naturally from the original limitation on the total number of spells the character could cast, which was, of course, a rather unimaginative hack in the first place.

So in Magi, rather than using fatigue to put a ceiling on fighting time, it will create a slow withering effect. A physical example will be easiest to relate to:

When you have been doing some strenuous physical activity for some time (running, lifting weights, etc.), you easily notice a buildup of general fatigue. Your muscles become slower to respond and cannot exert as much force for the same amount of effort. Your reflexes become dulled, and the part of your body that has become fatigued may even feel slightly numb.

So in the game, as you engage in strenuous physical activity (sprinting, running over long distances, fighting, etc.) you gradually build up fatigue (note that we have a variable called "fatigue" which is counting up, with no maximum, as opposed to a variable called "stamina" that counts down to zero). As your fatigue rises, you experience a gradual overall drop in physical performance. Certain attributes and skills help to delay or diminish this effect, but if you keep up a high level of activity without a break, it will wear you down. You move slower, hit for less damage, dodge and block less often, and so forth. The higher your fatigue rises, the more a particular action will fatigue you, because your capacity to do it has diminished. Eventually, if you build up enough fatigue, you can fall unconscious even if you have not been wounded.

Fatigue works similarly but independently for other dimensions. A psionic will gradually build up "psionic fatigue" and will experience a resultant slowing of mental processes, decreases in concentration and creativity, etc.

Although fatigue is accumulated separately for each dimension, all fatigue adds to an overall feeling of tiredness that impacts your performance in other areas to a lesser degree. For example, although running generally doesn't tire out your mind, your physical fatigue is distracting and makes it harder to concentrate on mental tasks.

Depending partly upon your attributes and skills, and partly upon luck, you can continue to fight for a period of time without any specific limit, but with decreasing effectiveness as you continue. This means that you won't run out of mana at a crucial time . . . it does mean, however, that you cannot use your strongest attacks forever without taking a break.

Logging Out

No matter how immersive the game experience, reality always maintains at least one major bastion of influence: sooner or later, the player has to quit. The game world is persistent, but the player's participation is not; sooner or later they log out. Somehow, characters popping in and out of existence seems to interrupt the role-playing experience. ("Where's the king?" "I'm sorry, he doesn't exist at the moment.")

But while it is inconvenient to take the character out of the world when the player logs off, it is impossible for them to go about their normal activities. Even if some controller (either human or computer) were available to take over for every player's character any time they quit, I wouldn't trust them to do a very good job at staying in character. If we assume that hiring a skilled team of "ghost players" to monitor your on-line persona around the clock to familiarize themselves with your play style and taking over whenever you're not playing is out of the question, we need to at least remove the player's avatar from active participation in the game.

The obvious option, of course, is to say that any time the player is logged out, the character is asleep. If they log out in a locked room at the inn with a "do not disturb" sign on their door, we could claim this while technically removing the character's presence from the game. If they quit while standing in the middle of a crowded city square at midday, this verbal excuse doesn't hold up very well.

Adding in some technical support, we could actually maintain the character's unconscious presence in the world when they log out. This would make the player quitting in the middle of a city square a rather obvious target for thieves, but we'd rather that the player took five minutes to get a room at the inn before quitting anyway, so they're really inflicting it upon themselves. (We will have to make large numbers of inexpensive inns available.) If the player goes to a relatively safe location and locks the door before going to sleep, odds are that they'll be just fine.

However, sometimes this option is not available. Occasionally, we can expect that the player will have no reasonable option but to camp out-of-doors, and it would not do to penalize them too harshly.

A computer could probably manage basic defense in the event that something wakes the character before the player returns. Additionally, everything that the character sees should be recorded for review by the player upon login. This seems somewhat less harsh, although it means that someone is going to need to write a pretty good AI script to control logged-out characters in emergencies. Players will be encouraged to find an out-of-the-way location and employ any stealth skills they have before they go to sleep.

Also, to provide some additional protection, we can put a consumable item in the game called a "sanctuary stone." This item encases the player in a protective field, but the player cannot move or interact with anyone or anything outside the field until it is deactivated. A player can procure one of these stones and activate it just before logging out, and they will be protected until they return.

The field should not be impenetrable, because that would be rather extreme, but it should be difficult and time-consuming to pierce. We still don't want to give the player a cheap way to escape when he's being hounded by an angry mob, whether PC or NPC.