Factions or Keepers

[svenskmand]

I have posted my current ideas for the backstory, if we go with the alignment system, on the wiki, it is at a very early stage. If you have comments please discuss in this thread (There is also a link to the wiki page in there).

[Keldaryth]

In regards to RTS games with factions... has anyone played Majesty 1 and 2? I love these games and they're kind of like Dungeon Keeper above ground and with heroes...



The Majesty series is built around indirect control - you hire people, and they do their own thing, with incentives from you (i.e. cash).

However, the more relevant part is that they used a very basic 'faction' choice system for game one

Each player had access to the same basic hero guilds:
+ Warriors
+ Wizards
+ Rangers
+ Thieves

You then had a choice of one of three demihuman races - choosing one meant the other two became unavailble
+ Dwarves (defensive ballistae tower, slow, magic resistant and built stuff)
+ Elves (very good archers and kicked your economy into overdrive, while bringing some vice into play to waste your heroes time and drain some money out of the system)
+ Gnomes (easy to kill, insanely fast builders and very difficult to get rid of, to prevent players from using them to build a base then switching races)

You had a choice to two tier two temple factions:

Life/Order: (strong, defensive and suited for turtling)
+ Healers (only healer unit in game - would sometimes heal enemies. Planted herbs that rangers could collect for healing potions)
+ Monks
+ Paladins at the Warriors Guild

Death/Chaos: (heavy damage dealing, light defense, swarmed with summons and charmed creatures)
+ Cultists (charmed wild animals and beasts, planted poison plants for thieves)
+ Priestesses (of death - drained life, summoned/charmed undead, restored health in undead by draining life)
+ Warriors of Discord at the warriors guild (crazy people with light armour and massive damage)

Then a choice of tier three temple factions

Sun:
+ Solarii (agressive melee combatants, sought out enemy buildings)

Moon:
+ Adepts (defensive melee, patrolled your buildings, very very fleet of foot)

Alternatively you could bypass all the temples by picking the last faction:

+ Barbarians (massive damage, rage ability)

In Majesty 1 the rules were absolute - picking one 'faction' meant the other(s) wouldn't help you. In Majesty 2 the lines were blurred so that you could have multiple factions, but some didn't get along, so dwarves would stun elves and elves would entangle dwarves in vines, paladins would attack the undead minions of a priestess, and so on.

From an outsiders perspective I think you could combine the two approaches - each unit should belong to a faction and have an alignment tag - creatures of diametrically opposed alignments would react negatively to each other, by sulking, leaving, fighting or underhanded pranks (sorry, did I trip you up and stun you? bad me! Oh, was that your head? I thought it was a spittoon for my corrosive acid spit... etc).

From a keeper perspective it allows you to cross factional boundaries if you wish, and it also allows for each major 'faction' so to speak, to have a light and a dark side - so you could have mages and warlocks as the 'light' and 'dark' side of the human faction for instance, or a paladin and a death knight (although a death knight would also fit into the undead faction I suppose - there's nothing to stop a creature from being part of more than one faction).

Things get interesting if you consider creature attraction. Assuming the DK style of portals is continued:
* basic rooms will attract a range of creatures from different factions
* specialised rooms will attract a range of creatures from a specific faction and/or alignment

Regardless of your room choice, each player would also have an alignment score: the weighted average alignment of all of his/her creatures. At the start of the game, this would either be dictated by the level builder, or alignment of the initial builders (unless the ruling is that kobolds begin as magical workers who have no alignment and are therefore neutral).

The alignment factor becomes important because it would determine how likely a creature would be to come through a creature portal.

Assume alignment ranges from 1 to -1, 0 being neutral, 1 being good and -1 being evil.

At the start of the game a player is neutral (0) and builds (for example) a food place, a lair and a library.

Say for the sake of argument the Library attracts the following:

Wizards (neutral human - probability 0.25)
Warlock (dark human - probability 0.25)
Necromancer (dark human/mythological - probability 0.25)
Fairy (light mythological - probability 0.25)

With no other factors, each is as likely as any other creature to appear in the creature portal

However, if the fairy is the first one through, it would be logical to assume that some of the other creatures might not want to come and join her, or her 'keeper', so we'd give each creature an alignment coefficient - a representation of how much they care about the keeper's apparent alignment. This coefficient would then be multiplied by the keeper's alignment score and the result used to modify the probability of that creature arriving at the dungeon.

First let's say the wizard, as a self serving egoist only interested in knowledge for its own ends, doesn't care what alignment the keeper is, and he gets an alignment coefficient of 0.
The Warlock, being largely interested in dark magic, is nonetheless still interested in finding out new things - he might be able to subvert light magic after all - gets an alignment coefficient of -0.15
The Necromancer, being a true creature of darkness detests all things light and has an alignment coefficient of -0.8
The fairy, being light and bubbly has an alignment coefficient of 0.7

In this example, the keeper starts with an alignment factor of 0, and attracts a fairy - alignment 1
Assuming the keeper has four kobolds (neutral), the average alignment factor is (1+0+0+0+0)/5 = 0.2

This means the creatures now have the following probabilities of appearing:

Wizards probability 0.25 + (0*0.2) = 0.25
Warlock probability 0.25 + (-0.15*0.2) = 0.22
Necromancer probability 0.25 +(-0.8*0.2) = 0.09
Fairy probability 0.25+(0.7*0.2) = 0.39

Rescaling those out of 100% gives the final probabilities of
Wizard: 0.26
Warlock: 0.23
Necromancer: 0.09
Fairy: 0.41
(1% missing due to rounding errors)

Thus you can see that a wizard is slightly more likely to appear than a warlock once a fairy has come through, necromancers will typically steer well clear and a second fairy is likely to join the first. Assuming more fairies arrive, you'd pretty soon find a situation where the Necromancer would not show up at all.

If you wanted you could add other modifiers to represent things like the attractiveness of any given faction to the creature (constructs might not care, but undead would be drawn more to the necromancer than the fairy).

This also poses an interesting question regarding how rooms and such would affect creature attraction.

I would suggest giving each creature an immutable trigger threshold - unless this condition is met, the creature will never consider coming into your dungeon (i.e. the library in the above example). However, after that all sorts of bonuses/penalties could apply - room size can be used as a modifier (probably one all to itself), and this also opens up the idea that a room can have an alignment impact - a torture chamber for example could have an effect of -0.02 alignment modifier per tile, meaning that in this example a player building a ten tile torture chamber would cancel the alignment effect of his or her first fairy.

It might be interesting to have this effect other areas of the game as well - alignment and faction weightings might dictate what spells/rooms you can research, or even what 'skin' is applied to the rooms you build.

Personally I'd like to see a range of keeper placed objects outside of traps and doors - things like statues and other dungeon furnishings. As well as being decorative, they would have an effect of modifying the alignment score or faction score of a player to allow for a strategy shift as well as potentially soothing/pissing off existing creatures or bringing in new ones to conflict with.

Obviously there's then scope for keeper personalities to affect the starting faction and alignment scores, as well as dungeon dressing for people who build levels.

Okay, that turned out to be more detailed than I first expected.

[svenskmand]

I need to check those games out

[TheAncientGoat]

Majesty is awesome, I've pointed it out a few times in the design discussion. Also, I really like what you're saying here Keldaryth, glad to have you on the team

[Skorpio]

I've played Majesty 2 and it was OK, but I wasn't thrilled. TBH I've somehow lost the interest in RTS in the last years, I tend towards turn based games like Heroes of Might and Magic or RTT like Close Combat 5 ... hmm actually I prefer multi-player shooters atm. So what am I doing here? *lol*

[Keldaryth]

I personally like the aspects of Majesty 1 more than 2. Two has been designed much more as a standard RTS, which has lost the flavour that made Majesty 1 so great.

You should also check out Startopia, which I understand was made with several ex-Bullfrog members. It does a lot of what people have been talking about. I've done some backstory tweaking on the wiki to give the factions a bit more depth but they'll all need some serious work to give them some alignment variation.

Is there any reason why the name 'Corpar' is being used? I haven't found any references as to how the name originated and while I've made something up, I'd rather use an alternate name that is instantly meaningful - I dubbed them Keepers of the Earth for fun and for a DK reference.

If people like the idea of alignment/faction attraction factors I can do a mock up of a simple system for it.

However if this is something that you'd like to look at the number of 'factions' would jump immediately. Essentially you'd need the following:

Good/Neutral/Evil Humans (not too hard to do, given that most of them typically fall into neutral anyway)
Good/Neutral/Evil Constructs - I thought this was initially impossible, but given the backstory, constructs were given reasoning, which means they could go either way. Still, in most cases, I would represent this as constructs being typically neutral.
Good/Neutral/Evil Copars/Mythological Creatures - probably an easy implement. You could add some demihumans to flesh this out (or add them to the 'human' camp) undead fill out the neutral/evil category and the current copars can go all across the section (okay, I rewrote them a bit).

If people don't mind me making a complete mess of the proposed creatures I'm happy to see what I can come up with.

[svenskmand]

TheAncientGoat and I have actually had a good discussion about a alignment system which I believe that we will use. I will see if I can find a link to it.

EDIT: Heh I could not find it because it was already in this thread , the post containing the system is here. I thing we should stick to that system, as it is very flexible

EDIT^2: I have attached a simple OpenOffice Calc-sheet where you can play around with the system for the four factions (I have not given attributes to the other "factions"/forces yet, and the attributes are not final at all, just how I would assign them to our 6 axises of the alignment system.

EDIT^3: It is important that we use a higer-dimensional simplex and not a hypercube, because with a simplex the "factions"/forces can be archtypes/extreme points in the simplex. In a hypercube it would not necessarily make sense to have the "factions"/forces as corner points in the cube - this would make it very hard to come up with sensible alignment axes.

[Keldaryth]

Okay, the issue I have with the proposed alignment system is that unlike the Lawful-Chaotic Good-Evil matrix, the proposed alignments here are subjective.

To take the example of indecent and decent - how do we define what is decent or indecent? Is an unannounced invasion of another societies land to plunder their resources a decent act? How is carving your name in the flesh of someone you sleep with (to use a recent example from the wiki) a decent act allowing humanity a decency score of 1?

We can argue averages or semantics or points of view and intentions until the cows come home - and then we can argue about the ethics of keeping cows for meat as well, but I think using subjective moral standards for alignment when two opposing views can both see their actions as just and correct is too difficult to implement. Granted, I've unsimplified the factions a bit - from the Copars' point of view they're being invaded and their deity desecrated so their actions are fully justified and righteous.

However, taking the list as a reference point I would propose:

Altruism/Egotism (basically charity or generosity as opposed to greed and self centredness)
Order/Chaos (lawful and unlawful, incorporating obedience and mischief)
Peace/Violence (as is)

This basically allows for (nearly) everything the twelve point alignment system enables, but will be simpler to implement and without the risks of subjectivity. That said from a systemic perspective, I think it could be *too* flexible, but I'm happy to go with it.

Some Examples:

Alignment Axes:
Altruism =1, Egotism -1
Order = 1, Chaos = -1
Peace = 1, Violence = -1

Example alignment traits (note all unspecified alignment axes will be unless combing it so a greedy lawless bezerker would be -1-1-1)
Greed = Egotism -1
Generosity = Altruism 1
Mischievous = Egotism -0.5 Chaos -0.5
Dispassionate/Logical: Order 1
Passionate: Chaos -0.75 to -1 (mix with other traits to define what the creature is passionate about)
Just = Order 1, Peace 1
Unjust = Chaos -1, Peace 1
Honor: Altruism 0.5 to 1, Order 1

Decent and indecent are harder to do because they tend to default to order (what a given society deems orderly conduct) and chaos (disorderly conduct) and varies from society to society. Such an alignment trait would not be applicable across all factions, which by nature have different social norms.

Also the alignment system would need to be implemented at a creature level, not a faction level. This will allow some members of different factions to get on well with each other and coexist peacefully. Where there is a factional bias between faction members that would be handled by the faction attraction modifier, not the alignment attraction modifier.

This means each creature would have a faction and a separate alignment on each of the three axis here. It would have faction coefficients for each of the factions in the game and alignment coefficients for each alignment axis. Based off 4 current factions and 3 alignment axes, you're already talking 7 coefficients per creature, if you had six axes we'd be looking at 10, and that's before bonus modifiers if those get implemented later.

As a comparison the simplicity of the Law/Chaos Good/Evil alignment system is only gives 2 alignment coefficients. Basically I'm in favor of keeping it simple to minimise the background scripting. If we simplified the factions to 3 and used the Law/Chaos Good/Evil alignment system to differentiate them all, we'd be looking at only 5 coefficients (plus bonus) per creature in the initial build. I'm not saying this isn't expandable in the future, but in the interests of getting the game up, I'm inclined to argue for the occams razor system - simplest system to give the desired flexibility.

I'll work on an example so I can show you what I mean... later... maybe after sleep.

[svenskmand]

Okay, the issue I have with the proposed alignment system is that unlike the Lawful-Chaotic Good-Evil matrix, the proposed alignments here are subjective.

To take the example of indecent and decent - how do we define what is decent or indecent? Is an unannounced invasion of another societies land to plunder their resources a decent act? How is carving your name in the flesh of someone you sleep with (to use a recent example from the wiki) a decent act allowing humanity a decency score of 1?

The labes of the axes are always up for discussion, you can even introduce as many as you want , I am only saying that we should keep the simplex-system as opposed to the hypercube-system as the hypercube-system is flawed/broken (you can have extremes that no faction has, and or all factions need to be forced to have alignments that does not suit them, to actually lie in the corners of the cube. The hybercube system also has the nasty property that the number of forces/factions should be a power of 2, which is not flexible :S

We can argue averages or semantics or points of view and intentions until the cows come home - and then we can argue about the ethics of keeping cows for meat as well, but I think using subjective moral standards for alignment when two opposing views can both see their actions as just and correct is too difficult to implement. Granted, I've unsimplified the factions a bit - from the Copars' point of view they're being invaded and their deity desecrated so their actions are fully justified and righteous.

However, taking the list as a reference point I would propose:

Altruism/Egotism (basically charity or generosity as opposed to greed and self centredness)
Order/Chaos (lawful and unlawful, incorporating obedience and mischief)
Peace/Violence (as is)

This basically allows for (nearly) everything the twelve point alignment system enables, but will be simpler to implement and without the risks of subjectivity. That said from a systemic perspective, I think it could be *too* flexible, but I'm happy to go with it.

Some Examples:

Alignment Axes:
Altruism =1, Egotism -1
Order = 1, Chaos = -1
Peace = 1, Violence = -1

Example alignment traits (note all unspecified alignment axes will be unless combing it so a greedy lawless bezerker would be -1-1-1)
Greed = Egotism -1
Generosity = Altruism 1
Mischievous = Egotism -0.5 Chaos -0.5
Dispassionate/Logical: Order 1
Passionate: Chaos -0.75 to -1 (mix with other traits to define what the creature is passionate about)
Just = Order 1, Peace 1
Unjust = Chaos -1, Peace 1

Decent and indecent are harder to do because they tend to default to order (what a given society deems orderly conduct) and chaos (disorderly conduct) and varies from society to society. Such an alignment trait would not be applicable across all factions, which by nature have different social norms.

Also the alignment system would need to be implemented at a creature level, not a faction level. This will allow some members of different factions to get on well with each other and coexist peacefully. Where there is a factional bias between faction members that would be handled by the faction attraction modifier, not the alignment attraction modifier.

Assigning a point for each creature would be a good idea, we have discussed something similar, here each creature has a hyper ball associated to it, so if a keepers alignment point lies inside the ball then the creature will be attracted to the keepers dungeon. Here we can also increase the probability, that the creature will join the keepers dungeons, proportional to how close the keepers alignment point is to the center of the creatures alignment-hyper ball.

I'll work on an example so I can show you what I mean... later... maybe after sleep.

Then please use the simplex system, do not propose something using the hypercube system. The simplex system is more flexible.

I can make a wiki page describing the details of the simplex-system. Then it can also describe the differences of the two systems, as the difference might seem subtle to people not familiar with higher-dimensional geometry.

[Keldaryth]

[
Then please use the simplex system, do not propose something using the hypercube system. The simplex system is more flexible.

I can make a wiki page describing the details of the simplex-system. Then it can also describe the differences of the two systems, as the difference might seem subtle to people not familiar with higher-dimensional geometry.

Well, I'm arguing not from any give alignment system per se, I'm arguing about minimising the amount of code the game has to run in the background and how much we have to write. My point was that the simpler the initial alignment system (whether or not it gets added to in future in terms of number of axes), the easier it is to code and test in the first instance and the less resources it will consume running in the background behind hopefully impressive graphics, which is why I'm happiest with the simplest possible alignment model that gives the desired outcome. If the desired outcome is to be able to position creatures realistically without forcing extremes of good/evil, then yes, the LC/GE matrix isn't the best system, but keeping axis numbers to a minimum (2 or 3) seems best to showcase the concept.

I think we're basically agreeing with each other so I'll shut up now, sleep and work on this later

[svenskmand]

LC/GE matrix? what is that?

I do not think we need to keep the number of alignment axis low, the more we have the more precise we can describe each faction and how the player is, and I think that it is then easier to make the internal story-logic consistent and the game-play balanced.

And from a coding point of view it will not be more hard to implement or use the simplex system.

[Keldaryth]

LC/GE - Lawful, Chaotic / Good Evil.



Anyway. I'll come back with a list and example interactions, although I might put that in a new thread for cleanliness.

[andrewbuck]

Just to respond to an earlier point made about the code, the calculations required for the computer to do even simple calculations are orders of magnitude greater than even the most complex alignment calculations. That being said, I do agree that we should try to keep the number of "axes" in the alignment system to a minimum (i.e. just one or two, preferably two). The reason for keeping the number small is not for the sake of the computer, but for the sake of the games designers and players. If you get too high of a dimensional alignment system it gets difficult to accurately place the creatures in the various axes so the values end up becoming somewhat meaningless. Also, it just gets too hard to keep track of the creatures alignments with more than two axes. Two well chosen axes can provide the vast majority of the categorization we are looking for, without introducing the problems of having 5 or 6. Also, I think the Lawful, Chaotic / Good Evil is a pretty good set of axes. It works quite well for the D&D series of games and there is almost no character personality that you can't fit onto that system. They also just use 3 "positions" on each axes, although I think we should go with a continuum.

-Buck

[svenskmand]

Andrew: It sounds like you are describing a hypercube system again. Please note that it has some major problems: 1) if each faction gets a point as the average alignment of creatures of that faction, and this point does not lie in a corner, then a player could have an alignment which is more extreme than any of the factions, which would give him a poor choice of creatures. 2) If each faction gets a corner point in a d-dimensional hypercube then we would need a faction for each corner of the hypercube, which means we would need 2^d factions, else we end up in the situation described in 1). Having 2^d factions is not flexible if we want say anything else than 1,2,4,8 or 16 factions. I wanted to introduce two more factions: Elves (both light and dark) and Mercenaries, to give more diversity in the creatures available to the player, but this gives us 6 factions which cannot be handled by the hypercube system.

Also why is it a bad thing to have more dimensions? As I see it it only helps to more precisely distinguish between factions.

[Keldaryth]

The purpose of this is not to give factions an alignment. Creatures have an alignment and the keeper has an alignment based off a combination of rooms built, initial alignment (if any) and average creature alignment. Based on an individual creature's attraction to the alignment and faction mixture of the keeper's dungeon, it will be given a probability of appearing in the keeper's dungeon. The total probabilities of all available creatures whose threshhold conditions have been met will then be rescaled out of 100% and creature appearance will be based off this. Since the keeper's alignment is always an average of the alignments of his current creatures (plus modifiers) and can never exceed 1 or negative 1 on any axis, there should never reach a situation where no creatures will appear.

[svenskmand]

The alignment system, as it was originally proposed was based off the single-player campaign, such that based on how the player plays the game he will receive an alignment based on this. The way he builds his dungeon only affects his alignment a small bit, but if he always builds the same types of dungeons then his alignment will start to change from game to game, given that it is different than the alignment that the his dungeon has.

[Keldaryth]

And it would still work that way - the players ending alignment could easily be set as his or her starting alignment for the next campaign, modified as they build their dungeon.

Also, it should be noted that each additional axis used diminishes the impact of each other axis.

[Keldaryth]

Linked from discussion here

The nazis built a better world by increasing the suffering felt by others. To increase the standard of living for few they increased the suffering of others. This is 'evil'.

I agree that it is evil, but the nazis did not think it was evil, that is my point.

Not quite. The LC/GE matrix is not used to describe how each creature sees themselves (or at least it has evolved this way). It suffers from using the terms 'good' and 'evil' which can be subjective. I am arguing for an objective measure of good/evil based on the infliction of suffering, which is measurable, if these axes are to be used. Furthermore, a large majority of fantasy players are familiar with the concepts of good/evil in absolute non subjective senses and can appreciate it as such even if the actual measure being used is hidden from them.

But my point is -- as you also write youself -- that if it where up to each faction they would label themself as good, which would mean that all factions would have the exact same alignment in any system. It is the same with you labels as they are as subjective as mine.

You're mistaking the LC/ GE matrix as 'my' axes. I am using them because there are those who believe that it's the best way to get what we want. I'm currently leaning towards the three axis system I am using, which attempts to be completely objective and not subjective. However, if the dual axis system is to be used, I would like to use a good/evil axis that is quantifiable, as has been done in games such as Black and White (and referenced above in this thread).

Fine then you do not mind that we use the simplex-system as it is much more flexible

Incorrect. I want the simplest possible system to achieve the desired result. Frankly, having to come up with six or more alignment coefficients for a creature strikes me as a complete waste of time. Not only does it force us to track six or more alignment variables as well as a faction variable for each player, we would have to come up with a way to weight the faction and alignment factors so that faction was still as important as alignment in determining creature attraction. Also, with every alignment axis we add we diminish the power of every other alignment axis. Essentially we could quite easily get to a point where they'd largely cancel each other out and we might as well not bother building the system - the results of each would wind up being 1% each way, especially if more axes are added.

I also want axes that can be applied across every single creature in the game that can be attracted to a players dungeon. This would potentially allow us to script in alignment and faction happiness/unhappiness and like/hate relationships between creatures based on their alignment and faction coefficients (with room to manually add more). This would mean that every creature added in or modded in by the community in their own projects would automatically relate to every other creature in game based on its faction and alignment.

I like flexibility, but I want simplicity. Maximum flexibility across a minimal number of axes. I'd like 2, but I'll settle for 3 since I don't personally see a way to get everything we (or I) want out of two. However, some people believe it is possible, and to test the theory I am doubling my workload and testing out both. Once the results come in we'll be able to see how it actually works.

[Keldaryth]

I would like to note that abandoning the Good/Evil axis will make it difficult to incorporate dungeon skinning based off alignment through... could still be fun to play with, but someone else can tackle that.

[svenskmand]

First of all having 3 axes is as arbitrary as having 6 or any other number for that matter. So why would it not be as much a waste a time to come up with 3, why not just use one axes then? According to what you are saying I cannot see why 3 is such a good number?

And if you want objective axes then you cannot use good/evil as they are based on ethics and culture. The peaceful/violent axis is objective yes, but as the goal of the game is to kill the opposing keepers then all factions would be equally violent, none will be peaceful as you cannot win the game by being peaceful. So this axis does not make sense as all creatures and keepers will be violent, as you cannot win the game through peace.

Also I still do not see any benefit in using a 2 axis system for creature alignment and 3 for keeper alignment (or was it the other way around?)

[svenskmand]

Incidentally my initial thoughts on simplex - hypercube notions is that one is based on triangles and the other on cubes. To be perfectly honest given the mechanic I'm trying to use in this thread, I don't see it as being an issue, especially given the attribute of an alignment to the individual creature, not the faction.

Hypertriangles (this is not a real word, but yes a simplex is sort of a triangle in higher dimensions) and hypercubes gives really different properties when it comes to placing points in them (alignment points) compared to the defining vertices of the simplex or hypercube. The simplex system will work both for factions an creatures, and I really think they should be have alignment points in the same system, having more systems does not make sense.

I like flexibility, but I want simplicity. Maximum flexibility across a minimal number of axes. I'd like 2, but I'll settle for 3 since I don't personally see a way to get everything we (or I) want out of two. However, some people believe it is possible, and to test the theory I am doubling my workload and testing out both. Once the results come in we'll be able to see how it actually works.

The simplex systems is very simple and very flexible. You simply give each faction a alignment point on the axes, this defines a simplex, then the player and creatures can have their alignment point inside of this simplex. This system is built upon the assumption that the players alignment is a mixture of the alignment of the factions he serve and the creatures he uses, which is also why every alignment point can then only be placed inside this simplex.

With a hypercube system you would have to assign each of the corner points to different factions, meaning that the labels will need to be carefully chosen such that it makes sense for each faction to belong to this corner. An example using axes good/evil and peacefull/violent then you will have the four corner points
(good, peaceful) which makes sense as they are not contradictory, (good,violent) which does not really make sense, how can you be good and also violent? (evil, peaceful) also does not make sense, and the last is (evil,violent) which again makes sense. So here you get combinations which does not make sense as they contradict each other. Secondly each corner point needs to be assigned to a faction and chances are that the point the faction get will not be very suitable.

[Keldaryth]

First of all having 3 axes is as arbitrary as having 6 or any other number for that matter. So why would it not be as much a waste a time to come up with 3, why not just use one axes then? According to what you are saying I cannot see why 3 is such a good number?

Because 3 is smaller than 6 and thus easier to implement, while largely being able to retain the flavour of the six.

The peaceful/violent axis is objective yes, but as the goal of the game is to kill the opposing keepers then all factions would be equally violent, none will be peaceful as you cannot win the game by being peaceful. So this axis does not make sense as all creatures and keepers will be violent, as you cannot win the game through peace.

Yes and no. Assuming that is the goal in the average multiplayer, yes. However, I would like to see the option for different goals being set that may or may not involve killing the opposing player (although it is often going to be a viable option) and the opportunity for co-op gameplay. Destroying all the hydra, basilisk and giant spider lairs around the map for example, or being the first to bribe the dragon with X amount of gemstones. It depends on how far you push the 'open' part of open dungeons.

Also I still do not see any benefit in using a 2 axis system for creature alignment and 3 for keeper alignment (or was it the other way around?)

Neither. Both keepers and creatures will either have a 2 axis system OR a 3 axis system. I'm simply doing examples for both to see if Buck is right is saying that

Two well chosen axes can provide the vast majority of the categorization we are looking for, without introducing the problems of having 5 or 6. Also, I think the Lawful, Chaotic / Good Evil is a pretty good set of axes. It works quite well for the D&D series of games and there is almost no character personality that you can't fit onto that system. They also just use 3 "positions" on each axes, although I think we should go with a continuum.

(Emphasis mine)

Personally, while I would like to agree with Buck I lean towards the three axis system that I proposed because I think it takes the best parts from your 6 axis system to allow more flexibility while being relatively simple to implement.

In regards to your Simplex/Hypercube example you are exactly correct - IF your assumption that Factions will have an alignment holds true. Under the system I am proposing no individual faction has an alignment and therefore there is no hypercube. My system is closest to your simplex system except once again the faction has no alignment.

Each creature has a point on each alignment axis (either 2 OR 3, depending on which if any of the systems being tested is chosen) and each player is given a corresponding alignment based on the mix of his/her creature alignments and/or any modifiers that may be in play.

Any potentially new creature's willingness to join is then based on how well it regards that player's alignment as determined by its alignment coefficient

It is further modified by the proportional faction mix of the players' creatures and how well it regards each of the present factions as determined by is factional coefficients.

[andrewbuck]

I'm not opposed to three axes but any more than this is getting too complicated. You start to get diminishing returns very quickly when you get beyond a couple of axes. The point that Keldaryth made about being able to fill in a value for all the creatures is exactly what I was concerned about when we get too many axes. You end up having to fill in so many values that half of the ones you fill in just end up being effectively random guesses and they detract from the power of the ones you actually care about. My statement regarding D&D using 2 axes was not an argument that we had to use exactly 2, it was just to show that a small number of axes can accomodate a huge number of creature archetypes. Just a 2 axis system with 3 points on each axes like D&D uses gives you 9 possible personality types which is about as many as you can have without them becoming meaningless. In a 3 axis system you would have 27 so you see that it quickly becomes difficult to come up with representative characters to fill a lot of those in. Since you have a lot of them left empty they are meaningless anyway and you might as well just go with the smaller number.

One advantage that the two axes system does have as well is that you can have a faction representing each corner (1,1) (1, -1)... as the four extreme factions and a 5th neutral faction. This simplicity makes it easy for a player to look at the available choices and see where they want to play. Whether you are in single player or multi-player you can then just pick one of the 5 and that alignment is set as your alignment and then your actions during the game move you around.

Ultimately I think this discussion has gotten way too involved. No one really seems to oppose the idea of having some n-dimensional space where we assign each creature a vector in that space and give the player a similar vector. then a distance function is used to calculate the probablity of a given creature coming through a portal. Given that we all agree on this idea I don't see why we don't just pick something and try it. We can always add more dimensions or delete unused dimensions later. I like that Keldaryth has put together that list in the other thread and I think we should through it on the wiki and try to get everything filled in and placed appropriately. No system is perfect and by trying them out we can see what works and what doesn't in each one.

-Buck

[Keldaryth]

Well, the initial examples are up using a small scale test here and you can see the proposed mechanics and results there. I will try to post a larger scale example to see if the system is still workable with a larger number of potential creatures, and then I will post and elabourate my thoughts on the questions, problems and opportunities the results pose.

At the present time the results are highly similar, which is an argument for using the two axis system, although I expect the three axis system to yield very different results in the larger scale, so we'll see. Still, I like the complexity this relatively simple system builds into the game in terms of creature dynamics, as in effect, every creature in your dungeon will affect the probability of any other creature showing up, and while I've shoved in numbers which I think will allow for factions to be played cohesively, there are a number of potential mixes that can happen - and happen well or very badly for the player trying to use them.

[andrewbuck]

Yeah, what you did with those posts is what I am referring to. We have had pages and pages of discussion about this topic (it had been going on for months even before you got here) and you are the first person to really propose something concrete enough that you can do probability calculations, etc with the rules laid out. Up until now everyone (myself included) has been simply expressing their vision for how things should be but this is about as far as we ever went. I think we should adopt your system as stated and start trying to fill in probabilities, etc for each creature and see where it gets us. This is why I suggested putting up something like a big table on the wiki for the creature probabilities so we can see how they work. We may also want to put together some sort of a spreadsheet to simulate the probability calcs so we can refine the creature alignments more easily. I have to go to work in a few minutes but I will try to whip up a spreadsheet later today, I think I have a pretty good way of setting it up so we can see the effects on the probability calculations.

If it turns out your system doesn't work then we can modify/abandon and move on, but I think this would be better than just continuing to debate here forever about how we should do this.

EDIT: After reading your recent post in the other thread about the simulated pick of 15 creatures, I think the spreadsheet be pretty hard to implement since some creatures "will not" show up and it is no longer just the probabilities determining the draw. I still think your idea is sound though.

-Buck