Space RPGs: Flight Models

I've been thinking a lot about a space RPG lately. I never played Traveller, but I've reviewed it, and it is far too complicated. I know that Traveller spawned Elite, which in turn spawned Privateer (which was a childhood favorite game of mine), but it is just too much for me to want to run as far as rules go these days.

One problem that I've been putting some brain bytes towards is how to handle space combat. There are two approaches: realistic physics or arcade style. Video games had to deal with this too. Arcade style is popular because it is easier to grok, even if it is unrealistic. An example would be violating the law of conservation of momentum, capping maximum speeds in a vacuum at something significantly less than C (speed of light), and so on.

An RPG could go with the same sort of idea: each turn, a sub-light ship moves X squares and can make Y changes in direction. This effectively taps maximum speed at X squares. You could even justify it by saying that the sub-light technology has some sort of special property that perhaps cancels out inertia or momentum (converting velocity to heat or something).

However, this totally disregards some unique aspects of spaceflight. For example, if a ship continually accelerates at 1 MPH, its velocity will eventually get to be significantly greater than 1 MPH! One of the advantages of traveling in a vacuum is that you can get up to a great speed, especially if you have enough fuel/energy to generate thrust for 1/2 your trip. I think the solution is somewhere in between: use pseudo-newtonian rules that feel unique and space-shipish but are easy to implement.

I think the key is to focus on acceleration, not velocity. After all, if the relative velocity of the combatants is zero, you might as well be stationary.

The two things that help a lot as far as limiting factors go are:

• Limited Acceleration due to Gs: The human body can take about 9 Gs before G-LOC occurs. Much equipment may not be able to handle that much. For example, a spindly vacuum-only ship may only be able to take 1/2 to 1 G. While that is not a limit on velocity, it does significantly limit acceleration. 1 G is about 10 m/s^2, or approx 22 MPH (so if you could accelerate from 0 to 65 MPH in one second you'd feel three gees). In another example, the Apollo trans-lunar velocity was something like 25K MPH, which is about 11,000 M/S. A 3G (30 m/s^2) burn would have to burn for something like 6 minutes to get you up to that speed.
  
  Note that a human body can take something like 45Gs without breaking under certain conditions. However, if you're talking about sustained, fighting capability -- 9 Gs is a good rule of thumb.
  
• Limited Fuel: Unless you posit a never-ending energy source for your sci-fi world, then fuel will not be unlimited. Say your spaceship is 2000 tons, about the size of the space shuttle. Kinetic Energy = (1/2) mass * velocity ^2. A mass of 2000 tons is close to 2,000,000 kg, and the desired velocity for a 3-day moon shot is 11,000 m/s, so the energy wrapped up in that enterprise is 121,000,000,000,000 joules. By my rough reckoning that's about the amount of energy in a million gallons of gasoline, 7.5 million lbs of coal, -- assuming 100% perfect efficiency in the engine! If you could assume that you can crack uranium-235, it'd be much less (about 7.5 lbs), but the mass of the reactor would have to get added to your 2000 ton spaceship! With the exception of drives like Ion Drives or Solar Rocket Engines which are highly efficient as far as propellent usage goes (but provide very low thrust -- and thus are not tactically interesting for most PCs...), most of the other options use a ton of fuel and are propellant inefficient (TANSTAAFL). So, the bottom line is that if you limit fuel, then it will limit the "burns" that a ship can do, which limits maneuvering and acceleration.
  

Here's my rough estimate of how to figure out "useful" scales for tactical ship-to-ship combat. Let's assume that primary weapons are energy weapons, mass drivers/cannons, and missiles/rockets.

• Energy: A light-second is 300,000 km. Particle weapons would travel much more slowly than C, but for lasers, 300,000 km is probably a reasonable maximum range. Anything longer than that and you'll start to have aiming problems. Note that this is a problem with sensors, too; a radar or lidar pulse needs to travel two ways (out and back), and the maximum unambigious range is likely going to be much less depending on waveforms and whatnot. There are also problems with sensor range due to the formula for a sphere; the returning energy pulse is reduced by a power of 4*PI*R^2 so you need VERY sensitive receivers to detect a returning pulse from extremely distant targets, even without atmospheric attenuation.
  
  I also imagine that you'd have trouble with dispersal reducing the power level of the beam as it spreads over distance. You'd need a very tight, focused beam and the power will attenuate over distance. The YAL-1, a megawatt class laser, has range of ~600 KM vs. thin skinned targets in an atmosphere. Here's a quick back of the envelope calculation:
  
  Power Density = Transmitted Power / 4 * Pi * R^2
  Power Density = 1 megawatt / 12.56 * 600 KM^2
  Power Density = 1 megawatt / 4,521,600 KM
  
  So, I think something on the order of ~1000 KM is reasonable for desired weapons effects with a megawatt-class laser weapon in a vacuum. If you assumed that a sci fi laser would be a giga-watt class system (1000 times more powerful than the YAL-1), then perhaps closer to 20,000 KM as a WAG, although I'm sure an engineer would tell me how far off I am.
  
  Power Density = Transmitted Power / 4 * Pi * R^2
  1 megawatt / 4,521,600 KM = 1000 megawatts / 12.56 * R^2
  12.56 megawatts / 4,521,600 KM = 1000 megawatts / R^2
  R^2 * 2.7777777777777777777777777777778e-6 = 1000 megawatts
  R^2 = 360000000
  R = 18973
  (Sorry I got lazy and dropped the units... this is a WAG anyways...)
  
  Likewise, if you went DOWN to a kilowatt laser, then range would shrink to something like 50 KM vs. a thin skinned target.
  
  
• Projectiles: Range is effectively infinite due to Newton's First Law. However, against a maneuvering target, there might be problems with anything at longer ranges. If you assume a mid-case 6G manuevering target, then in one second that target can change velocity by 60 meters/second^2. A bullet usually goes at a velocity on the order of 1500 meters/second; artillery shells are much slower (hundreds of m/s) but we're in the ballpark. If you want to be able to hit a 10 meter sized "kill zone" on a maneuvering target, then your projectile needs to arrive in 1/6 of a second or you need to really be able to guess where it will be. That makes the effective range about 250 meters! That makes sense, though; ~750 feet is the heart of the envelope for a fighter aircraft trying to gun another fighter, which has similar acceleration and size issues; atmospherics don't even really come into it.
  
  If you're going after a 100 meter vital zone on a 3G target, then your bullet has 3.3 seconds, so max effective range would be ~4000 meters.
  
  
• Missiles: Like mass drivers/projectiles, range is effectively unlimited. In fact, it is greater because a missile (A) accelerates after launch and (B) can correct its course with terminal guidance. The only issue is how long it takes to get to the target, and if it has sufficient maneuvering ability to catch a maneuvering target.
  

So, really, tactical space combat scales can be smaller than you might think. Hexes as small as 250 meters might make sense! Remember, velocity doesn't matter; the only thing that matters is DELTA V (the difference in velocity between combatants). If you went with a 500 meter hex then you'd get the following:

• Guns vs. Maneuvering Point Targets: 1 hex range
• Guns vs. Non-Maneuvering Area Target: 8 hex range
• Megawatt Class Laser vs. Thin Skin Target: 2000 hex range (effectively infinite)
  
• Megawatt Class Laser vs. Medium Skin Target: 1000 hex range (still effectively infinite)
• Kilowatt Class Laser vs. Thin Skin Target: 100 hex range (still effectively infinite)
• Kilowatt Class Laser vs. Medium Skin Target: 50 hex range
  Note that if you posit armored targets, or better yet, some sort of energy shields, then you could shrink the ranges significantly, especially for a kilowatt class laser. The YAL-1 takes up an entire 747. I understand that sci-tech will make major strides in a sci-fi setting, but if you're talking about a laser that can sit inside a turret like you see in Star Wars, that's a LOT of miniaturization. So saying that small turret-based weapons are kilowatt class is reasonable. A megawatt class laser might be something like a mining cutting laser or a destroyer (not fighter) class weapon.
  
• 1G acceleration over a one minute period of time: ~1 hex delta vee (and this relationship holds fairly steady, so 9Gs = 9 hexes of delta vee)
  
  

The last bullet there is actually a good argument for using a 600 meter hex for tactical spaceship combat. It would basically create a 1G acceleration = 1 hex movement direct relationship, which is pretty sweet; it wouldn't really affect the maximum weapons ranges much, except that the guns vs. maneuvering targets is getting a bit optimistic. I can also think of some cool things you could do with a D6 and 600 meter hexes. The 500 meter hex is easier to extrapolate to greater ranges, however, and you could assume there are some inefficiencies in the burn or something that keeps acceleration from being perfect.

Alternatively, you could go with some multiple (2-3G) for the hex, which would lead you to 1200 or 1800 meter hexes. You could just round to 1KM, 1.5KM, or 2KM. That basically reduces movement rates for very rapidly accelerating objects such as guided missiles, but it also reduces movement rates for the typical player-controlled ship. If you think that a 3-5G rated ship is "typical" with 9-10Gs being a peak manned combat vehicle, then using 1800 meter hexes means that the typical ship only moves 1-2 hexes, which may not be very satisfying.

Remember, the key is acceleration capability in a fight, not velocity. Assuming that the combatants are intentionally getting into a fight, then one of them has matched velocity with the other. The aggressor in that case should basically be able to enter the fight with whatever starting relative velocity they want. For example, if they enter the fight with an advantage of +3600 KM/HR, then they'll have a velocity delta of 60,000 m/s, or 100 hexes. So, they'll blow through the engagement and only get one pass, which may be what they want. They'll have to do a long ~10 plus minute burn after blowing through to come around for another run which burns a lot of fuel.

It is much more fuel efficient to basically match velocities and enter the fight with manageable Delta Vee. That way you're not burning a bunch of mass to get up to high speed then burning yet more to rapidly decelerate. So letting the aggressor pick their starting relative Delta Vee advantage makes a lot of sense. In a more reasonable and tactical scenario, the aggressor could enter the engagement with a delta vee of +260 KM/HR, which gives them 10 hexes of movement. They could then make a pass on the victim, do a 9G braking burn, and be drifting along one hex faster than the target for round two of the fight -- a manageable place to be! It is a little more complicated than that because they could apply angular delta vee but still, you could come up with rules of thumb to keep it simple and manageable.

I'm going to stop now before I embarrass myself further. It has been a long time since I took physics. But I think that with some careful thought you can create a pseudo-Newtonian feel to the system fairly easily.

Correspondences of Seven

I've been coming up with a table of correspondences for the number seven, which is core to "Septimus." This is what I have thus far.

• Quality of Associated Signs: Each planet is associated with astrological signs as a "ruler." I've extracted the quality (fixed, mutable, or cardinal) of those signs.
• Elements of Associated Signs: Each planet is associated with astrological signs. I've extracted the element (air, earth, water, fire) of those signs.
• Humour: Each planet is often aligned with one of the four humours.
• Roman Deity: Each planet is associated with a Roman deity.
  
• Liberal Art: Dante Alleghri (yes, that Dante) associated each planet with a liberal art.
• Mechanical Art: I've matched up the planets with the mechanical arts.
• Probitate: I've also matched up each planet with a probitate.
• Metal: Each planet has an alchemical association to a metal.
• Primacy of Matter, Mind, and Spirit: Each planet has one of these three over another. For example, "Matter over Mind." You can extract it from the symbol for the planet.
  

	SATURN	JUPITER	SUN	MARS	VENUS	MERCURY	MOON
Sign	Capricorn	Sagittarius	Leo	Aries	Taurus	Gemini	Cancer
Sign	Aquarius	Pisces		Scorpio	Libra	Virgo
Humour	Melancholic	Sanguine	Choleric	Choleric	Phlegmatic	Variable (Melancholic)	Melancholic
Dante's Virtue	Contemplatives	Kings	Theologians	Warriors	Love & Friendship	Good for Fame	Virtue
Vice	Sloth	Pride	Greed	Wrath	Lust	Lying	Envy
Metal	Lead	Tin	Gold	Iron	Copper	Quicksilver	Silver
Roman God	Saturn	Jupiter	Apollo	Mars	Venus	Mercury	Diana
Greek God	Cronus	Zeus	Helios	Ares	Aphrodite	Hermes	Selene
Quality: FIXED	X		X	X	X
Quality: CARDINAL	X			X	X		X
Quality: MUTABLE		X				X
Element: EARTH	X				X	X
Element: WATER		X		X			X
Element: AIR	X				X	X
Element: FIRE		X	X	X
Primacy (Over)	Matter	Mind	Spirit	Matter	Spirit	Mind	Mind
Primacy (Under)	Mind	Matter		Spirit	Matter	Spirit
Primacy (Not Present)	Spirit	Spirit		Mind	Mind	Matter
Liberal Art	Astronomy	Geometry	Arithmetic	Music	Rhetoric	Dialectic	Grammar
Mechanical Art	Agriculture	Architecture	Medicine	Armament	Fabric Making	Commerce	Hunting
Probitate	Riding	Jousting	Climbing	Fighting	Swimming	Dancing	Shooting

Most of the linkages are straightforward. I'll explain the ones that aren't directly from the classical sources.

The mechanical arts:

• Agriculture: Saturn has a clear link to Agriculture. Heck, Saturn's symbol even looks like a scythe. This one is easy.
• Medicine: The sun is associated with Apollo, the god of the sun. Historically, the sun has been associated with positive, healthful things like the healing arts.
• Armament: This one wasn't perfect, but I thought a correspondence to Mars was appropriate. Traditionally, Hephaestus was the god of smiths. However, given the martial bent of Mars and the strong association with Fire, Mars is appropriate enough. I would prefer Fire + Earth.
• Commerce: Mercury is traditionally associated with Mercatura.
• Hunting: Diana, the huntress, is the patron of hunting. Easy.
• Fabric Making: Traditionally, Athena is the goddess of weaving. However, she doesn't have a planetary association, although she is sometimes linked to Venus. I ended up linking it to Venus almost by default.
• Architecture: This, like Fabric Making, was "left over" after I matched up all the other arts with their obvious associations. I paired it with Jupiter as Jupiter is associated with kingship, mind, and matter.

The probitates:

• Riding: I linked this with Saturn due to the agricultural association. However, the Roman God of horses was Neptune which could lead to a weak linkage with the element of water, which would be a reason to put riding elsewhere.
• Jousting: Linked with Jupiter. This was a "left over" as well and I dropped it here due to the linkage with kings. I thought about replacing it with Chess (tactics) as well.
• Climbing: Linked with sun. This was also a weak association, based solely on the idea of vertical ascent. It isn't bad, though, as Helios was known for climbing across the sky in his chariot every day. It would also be appropriate to use "Poetry & Music" here as Apollo was well regarded as a patron of both.
  
• Fighting: An easy linkage to Mars.
• Swimming: I paired this with Venus primarily based on the strong presence of the element water. Venus is the only Phlegmatic planet and has the strongest association with water.
  
• Dancing: This was matched with Mercury. Mercury was known for being fleet of foot as well as being very social, which is a nice match with dancing.
• Shooting: The link between archery and the huntress seems straightforward and natural.

Gamer Typology

Earlier I mentioned tailoring characters to player psychology. I think this is worth elaborating on a bit. There are many different player typologies. There seem to be a few different examples:

• Frivolous. Some are clearly more interested in humor than any serious typology.
  
• Anecdotal. These typologies sort players into groups based on experience and perceived common traits. However, there is no correlation between them implied. For example, one group may not be the opposite of another. They are very qualitative. Some examples might be Robin Law's typology, or this typology from the 80s.
  
• Systematic. These typologies set up defined axises, then define the player groups based on how they fit to the axis. An example is the Bartle classification (Player/World vs. Action/Interaction) or the WOTC survey (Long Term/Short Term vs. Tactical/Story).
  

I like the systematic approach because there is an underlying theory that can be generalized. The anecdotal system may be useful as well, but there's really no reason that you can't add or subtract types at a whim.

There are many real-world typologies for people that describe their temperament. You may be familiar with the Myers-Briggs Temperment Indicator, Keirsey Temperment Sorter, the Big 5, Enneagram, or many of the other ones. Because of my penchant for archaic systems, I like Galen's old theory of "Humorism" which many of the other four-category systems evolved from. I also like that the four categories are easy to grok; using shades between the categories gives you up to eight.

Humorism relies on the interaction of two factors. One measured "length of response" (rapid & hot vs. slow & cold) and the other measured "duration of response" (long & dry vs. short & wet). For example, a choleric, who has a short (hot) and lengthy (dry) response, may be quick to anger and long to hold a grudge. Of course, the four humours were correlated to the classical elements, and all the correlations that entails. Modern psychology has related the former axis to either emotional stability or to task vs. relationship orientation. The latter axis has been related to extraversion vs. introversion.

I think that you can relate these to the Bartle and WOTC typologies as well. Bartle himself admitted that his original pairing of typologies to card suits was done without knowledge of the ancient correlations and that tieing his theory to humourism is viable. So, how would you relate them?

• EXTRAVERT/INTROVERT: I link this axis to the Bartle Player vs. World focus or the WOTC Short vs. Long term. In this linkage, extraverts tend to be interested in the other players and short term outcomes. Introverts tend to get absorbed in the imaginery world and require less immediate stimulus.
• TASK/RELATIONSHIP ORIENTED: I have linked this axis to the Bartle action vs. interaction and the WOTC Tactic vs. Story distinction.

That gives you these four typologies:

• CHOLERIC (FIRE): Power Gamer (Short Term Tactical) / Killer (Acting on Players)
  
• MELANCHOLIC (EARTH): Thinker (Long Term Tactical) / Achiever (Acting on World)
  
• PHLEGMATIC (WATER): Storyteller (Long Term Story) / Explorer (Interacting with World)
  
• SANGUINE (AIR): Character Actor (Short Term Story) / Socializer (Interacting with Players)
  

Are these perfect typologies? No, certainly not. However, it is clear that the idea of these enduring archetypes has been entrancing us for millenia. Plus, in a role playing game, even the idea of archetypes is helpful, I think. If nothing else, they are fun to think about.

What does this suggest about each "element's" powers?

• FIRE: This school should focus on short term effects. It should also allow action on the other allied player characters. One might think of certain leaders from 4E that can actually slide their allies around, give them attacks, and so on. You're helping your team, but you are in the driver's seat. One can also imagine some mechanics that allow Fire to get their team involved in quests -- thus giving the Fire focused character a limited amount of control over the background/narrative of the other characters! The player is likely task oriented and should be rewarded for completing somewhat shorter term tasks.
  
• EARTH: This school focuses on long term effects. It should allow action on the world. This sort of player would likely enjoy controlling fiefdoms and larger numbers of obedient soldiers or hirelings. The player is likely goal oriented and should be rewarded for achieving long term tasks.
• WATER: This school also has a long term focus, however, unlike earth, it is expert at interaction with the world and story rather than imposing action. Water is about learning, divination, and subtlety. The player is likely relationship oriented and should be rewarded for developing genuine long term relationships.
• AIR: This school, like fire, is short term in focus but instead of being focused on imposing or acting on the other PCs it focuses on interaction with them. An air character might be interested in conducting multiple quests, coming up with new and novel ways to communicate or share experiences and information. The player may actually be interested in interacting out of character with other players or be particularly theatrical. This character should literally be a breath of fresh air, and the player may want to be able to act somewhat irresponsibly without dooming the party or their character. The player should be rewarded for entertaining and engaging others.
  

Defining Characters

About a year ago, Dave Noonan, one of the 4E designers, posted a series on roles in D&D. While he got some of his history wrong and was much derided by some of the OSR types, I thought there were some interesting things to think about.

How did OD&D mechanically define characters? Sidebar: I don't care about roleplaying differentiation. Obviously you can play one fighter as a noble knight errant and another as a thieving brigand. I am referring to the system's actual crunchy differentiation.

• Class: Fighter / Magic-User / Cleric. This differentiation defined a lot of your mechanics used, abilities, and so on. I would argue that the most important thing determining what class a player should pick is their experience. Novice players would have the most success with fighters, which were relatively simple and easy to play. High level fighters had access to magic swords which made them a little more interesting. Middling players could pick clerics, who started to introduce some special abilities like turn undead and spells, although a level 1 cleric had no spells. Finally, advanced players might select a magic-user, which was definitely a challenge to keep alive but theoretically had the most power if played wisely. A party didn't necessarily need all three roles filled but it helps a lot. Dwarves are basically a subclass of fighters, and elves are the "fifth wheel" jack of all trades, switching as needed between fighters and magic-users.
  
• Primary Ability Score: Your best ability score largely drove what class you should pick. You'd get an XP bonus and sometimes other minor bonuses based on this score. Sure, you could play a fighter with 17 INT and 10 STR but it probably isn't smart. This basically allocates roles randomly if you use random ability score generation. This is good with an experienced crew because it ensures all the key roles are filled, but it is bad because the roles may not be allocated according to player skill.
  
• Other Ability Scores: Other abilities had minor effects which served to differentiate characters of hte same class. A fighter with 18 DEX might play a little differently than a fighter with 18 CHA.
• Spells or Equipment: Magic Users could pick different spells which made them different from any other MU; martial types could use different weapons or armor.

AD&D expanded on this somewhat, but not fundamentally. You basically had four core roles (Ftr/MU/Cleric/Thief) which were still based on player skill. I'd argue that fighters and then clerics are the easiest to pick up, and thieves and mages the hardest to play well. Ability score generation might allow you to arrange the scores to taste which gave the player more control over the character generated; a mixed blessing because you might end up stuck in one role ("Well, your cleric just died... But we really need a cleric on the team... So why don't you drop that 17 in WIS and make another one?"). More subclasses and more important ability score riders basically improved differentiation between the roles, as well as some cross over. A paladin and fighter both filled the same role but the paladin had a bit of cleric flavor too. Race also became a seperate thing; this doesn't seem to have a systematic effect on the system, however. Multi and dual classing allowed a bit more mixing between previously rigid niches.

4E is the most radical departure in some ways. In 4E, characters are broken out as follows:

• Role: Defender, Leader, Striker/Controller (DPS). This is the key thing that a party needs one of each to be balanced. There is also some relationship to ability scores. For example, leaders generally don't use DEX (and often prefer CHA).
• Power Source: Martial, Arcane, Divine, Primal, Etc. This determines the flavor that a character uses to accomplish their role's responsibilities. It also has a loose relationship to ability scores. For example, many arcane characters favor INT (but not all). This actually encourages a party to all pick different power sources so as to cover all the different ability scores.
• Race: Races sometimes have a slight leaning towards a role. For example, Half-Elves feel like leaders. However, there is no strong niche protection. Every party doesn't need an elf; every party does need a defender though.
  
• Class: A combination of Role + Power Source. For example, a Divine Leader is a cleric. A Martial Leader is a warlord. Sometimes there are two options within a given combination; rogues and rangers are both Martial Strikers. Thus, class provides some degree of differentiation, almost like subclasses did in AD&D.
  
• Primary Ability Scores: Each character usually has two primary ability scores that they are pumping throughout their career. One will be mandatory and determined by their class. Most classes then offer two other options to pick from which will have your character "lean" towards another role. For example, all shamans will be wise leaders; you can pick INT to lean towards striker though or CON to lean towards defender.
• Other Ability Scores: These provide minor differentiation.
• Feats: These provide differentiation between characters so that no two Rangers play the exact same.
  
• Powers: Similar to spell and equipment selection in previous edition, these differentiate between characters so that no two Rangers play the exact same.
  

Noonan writes, "the roles came from somewhere, but they didn't inevitably emerge from player psychology. Players are all too happy to hop roles whenever they get the chance." That is true, in some cases.

In older versions of D&D, I think the most succesful role hopping was in order as player skill improved: Fighter --> Cleric/Thief --> Magic-User. Assuming a group of equally skilled players, however, players might hop roles whenever they got the chance. AD&D encourages this somewhat by making it relatively easy to replace fighters with henchmen, hirelings, and even summoned creatures. There are also several more complicated subclasses for the meatshield role (Defenders/Rangers; more than any other set of sub-class!). That way a group of very experienced players could basically leave out plain fighters in order to play more complicated (and to them, interesting) characters. AD&D also allowed players to arrange ability scores to taste in many cases, which gave players more control over what they played.

In 4E, role hopping isn't really based on player skill. You do need to fill all four roles (although some argue that controller/striker are redundant), though. While there is some preference, I do think that most players are happy to hop roles when they get the chance.

Where does this get us? Well, I've got some thoughts on how to bring this together in a more coherent manner.

ROLE: Roles basically serve as niche protection. They are specialties that your character is the best at. Some examples might be Defender/DPS/Healer, or Pilot/Engineer/Gunner. A game that has niche protection needs at least two different niches. However, if each niche is vitally important (and they should be; who wants to play a role that doesn't matter?) then the more niches you have, the larger the player group must be. For example, 4E is predicated on a group of five players; thus there are four roles to fill. I personally think that a group size of 3-4 players is most reasonable and attainable, especially these days, so something like three niche roles is probably good.

I think some games have a problem with too many niches. For example, Shadowrun has Street Samurai, Adepts, Faces, Hackers/Deckers, Riggers, and Magicians. You'd need six players to cover each role. Plus, the roles are so highly specialized and use unique rule space that I think it makes it hard to have a coherent and interesting adventure. The rigger -- specialized in operating vehicles -- will love a car chase but the adept will probably suck at it.

You can have more niches with smaller groups by using some contrivances:

• Make some niches easily replaceable. AD&D does this with fighters. It is easy to find meatshields.
• Make it easier to cover multiple niches. In AD&D, multi and dual classing allows some covering of multiple niches with one character. Interestingly, the dual-classing mechanics make it very easy to cover the fighter role (hrm... poor fighter...). Many of the best dual-classes start as fighters or clerics then switch to something else.
• Have sub-classes that lean towards a second niche. AD&D has a paladin (fighter leaning cleric) and ranger (argueably a fighter leaning thief). 4E explicitly has many classes lean towards different sub roles. Paladins (defenders) can lean striker or leader, for example. If you're missing a key niche, then having two characters minor in it can help cover. For example, if an AD&D party is missing a pure cleric, then you could cover for it by having a paladin and a multiclass cleric around.

MEANS: The means that a character uses are how they fill their role. This, I argue, should be based on player typology. I think you could either use the Bartle CARD roles or the WOTC player roles study or Myers Briggs types. In theory, a given player will probably have a preference for one means over another -- and that's ok. In theory, players who are interested in social interaction with their other players should be rewarded by the system for that. Allowing some bleed over between roles isn't a bad idea as not every player can be easily pigeon holed.

PLAYER SKILL LEVEL: You should be able to set gamplay to easy for a new guy or hard for an experienced veteran. I think everyone's seen a novice player struggle with playing a magic-user at the AD&D table and get frustrated with the experience; they just want to sling spells and they keep dying and getting expected to carefully ration resources! Ideally, I think there'd by a way to move from an easy set up to a harder set up throughout play. THat was, as a player gains experience, they can dial up the difficulty without making a new character. These are some ways to control difficulty of play:

• Survivability: Easy means more surviveable. Fighters have D10s; magic users had D4s.
• Nova: Easy means that the player has little control over how quickly their offensive resources are used up. Fighters have a good THAC0 chart and good AC all the time, but find it hard to dish out a ton of hurt in a short period of time. A harder play style requires the player to carefully dish out resources at critical moments. A first level magic-user only gets one Sleep bomb. It is like a nuke, but the player must carefully decide when to drop it.
• Choice: Easy means fewer choices. Fighters choose what armor to wear and what weapon to wield and that's about it. Moreover, in OD&D, weapon choice doesn't even really matter that much. AD&D introduces some bad choices (fighters shouldn't specialize in clubs) but fighters still have few options. Hard means more choices. Spell casters must carefully assemble their spell books and choose what to memorize each day.
• Mobility: Easy characters are not very mobile. They don't derive much benefit from positioning (i.e. it doesn't matter where they are) and can't get around very easily. Hard characters derive a lot of benefits from being in the right place in the right time and have the mobility needed to make it happen.
• Mechanics: Easy characters use one unified mechanic. Roll high on a D20. Hard characters may require players to do more complicated dice gymnastics.
  

POWER SOURCE: How does the character accomplish their role? What are the "special effects" associated with their actions? Some examples might be arcane, divine, martial, etc. In a sci-fi game, maybe it is Hacking, Cybernetics, Psionics, or Vehicles. I think this is more campaign dependent than anything else. You could overlap this with one of the other factors above as desired. For example, AD&D overlaps power source with Player Skill (Arcane = Hard, Divine = Medium, Martial = Easy). I think it would also be cool to allow players to change power sources in play. For example, if you want to make magic rare and special, it could be something that can be gained only later in a character's career after a special quest rather than from level 1 on up. As another example, perhaps Primal power can only be obtained after going on a Spirit or Vision Quest.

DIFFERENTIATION: How are otherwise identical characters differentiated? Are there little unique traits associated with ability scores, feats, or other choices?

Put all this together and what do you get?

ROLE (3: tank/healer/DPS, or something else) x MEANS (4: CARD or WOTC or Myers Briggs) x PLAYER SKILL LEVEL (3: easy/medium/hard) x POWER SOURCE (4+/-3 choices; could be random, campaign dependent) x DIFFERENTIATION (4+/-3 choices; could be random).

That gives you 36 different combinations before factoring in power sources or differentiation, or any options to "hybrid" or multiclass. For example, if you have no feats/ability scores for differentiation, and only one power source, then you can have 36 combinations. You might have a Thinker Easy Tank, or an Explorer Hard DPS.