Topic: Quick CharGen Question
Started by: Kester Pelagius
Started on: 11/5/2002
Board: Indie Game Design
On 11/5/2002 at 6:58pm, Kester Pelagius wrote:
Quick CharGen Question
Greetings All,
I am in the process of revising sections of my Pixies rules find original thread here and was wondering whether character generation should be point allocation or random die generation?
If the latter what sort of die would you prefer to see used?
Kind Regards,
Kester Pelagius
Forge Reference Links:
Topic 3659
On 11/5/2002 at 8:37pm, ethan_greer wrote:
RE: Quick CharGen Question
I'm not a really big fan of either of those options. Why not go with freeform char-gen?
Alternatively, You might consider first determining the trait limits for characters (highest and lowest possible trait level, and how many total points, etc.) Once you know how you want a character to look numbers-wise, you can tailor two char-gen systems - one which uses random roll, and the other which uses point allocation. Then let the individual group decide which to use.
If you want a straight answer (what a novel concept...) then I would say go with point-allocation. Given the choice between random and non-random char-gen, I'll take non-random every time. But with a game about Pixies, a random char-gen potentially has stylistic benefits.
Would you like some syrup with your waffles? :)
-e.
On 11/6/2002 at 12:38am, Kester Pelagius wrote:
RE: Quick CharGen Question
Greetings silkworm,
Thanks for the input.
silkworm wrote: I'm not a really big fan of either of those options. Why not go with freeform char-gen?
...
If you want a straight answer (what a novel concept...) then I would say go with point-allocation. Given the choice between random and non-random char-gen, I'll take non-random every time. But with a game about Pixies, a random char-gen potentially has stylistic benefits.
Would you like some syrup with your waffles? :)
No syrup, but if you could expand upon what sort of "freeform" chargen system you'd prefer to see that would be most insightful.
If it helps any the system is essentially Skill driven. Very minimalistic on the number of Traits/Abilities.
On 11/6/2002 at 3:02am, ethan_greer wrote:
RE: Quick CharGen Question
Well, my current favorite set of char-gen rules is here (towards the middle of the page) as part of Pollies. It's what I had in mind when I mentioned freeform char-gen above. I don't know how well it would work with a set skill list, but it works fairly well as written.
-e.
On 11/6/2002 at 3:14am, Ron Edwards wrote:
RE: Quick CharGen Question
Hi Kester,
I suggest looking at the other elements of the game in order to answer this question.
Random-feature character creation has a spotty history in role-playing and I don't think we've seen many games in which it's played a constructive role (but I do think that's possible). If you use it, you should consider what specific elements of play are enhanced by its presence. Look at the resolution system, the reward system, and other things - for instance, if Doing X relies on a particular feature, and Getting Better relies greatly of Doing X, then I as a player will be mighty pissed if the particular feature might start very poor due to a random roll.
On the other hand, if Doing X has lots of viable ways to be accomplished, and a random roll gives me a "starting profile" for how I Do X, then I as a player am intrigued and interested in how I might start, secure in my knowledge that whatever comes up (possibly with disadvantages as well as advantages), I will still be able to Do X.
The best example of such a system that I know of is the magic system in Forge: Out of Chaos, in which one's spell features (range, etc) are determined randomly, and the entire reward system for magic-guys is based on increasing one's number of rolls and re-rolls regarding one's spells. I have always wanted to play this game long-term to see how this plays out.
Best,
Ron
On 11/6/2002 at 11:34pm, Kester Pelagius wrote:
RE: Quick CharGen Question
Greetings silkworm,
Nice page you have there with Pollies.
silkworm wrote: Well, my current favorite set of char-gen rules is here (towards the middle of the page) as part of Pollies. It's what I had in mind when I mentioned freeform char-gen above.
Interesting, sort of backward to what I had in mind... which was have Character Generation provide the descriptors for the players to use to describe their characters.
silkworm wrote: I don't know how well it would work with a set skill list, but it works fairly well as written.
At the moment I am just curious to learn what method of Character Generation members of The Forge prefer, and why. I want to try and make the game one which people will want to play. To do that I am looking into what players currently find most enjoyable. That said I appreciate your feedback.
Kind Regards,
Kester Pelagius
On 11/6/2002 at 11:48pm, Kester Pelagius wrote:
RE: Quick CharGen Question
Greetings Mr. Edwards,
I appreciate your feednack but, at the moment, I am really just exploring possibilites. Doing a bit of "brainstorming" if you will.
However I am a bit confused by the following:
Random-feature character creation has a spotty history in role-playing and I don't think we've seen many games in which it's played a constructive role (but I do think that's possible). If you use it, you should consider what specific elements of play are enhanced by its presence. Look at the resolution system, the reward system, and other things - for instance, if Doing X relies on a particular feature, and Getting Better relies greatly of Doing X, then I as a player will be mighty pissed if the particular feature might start very poor due to a random roll.
It sounds as if I might have given the impression that I don't have a system in place and am wanting to generating statistics on the fly. That is far from the case. I do have a number of statistics in place, but I want to explore what sort of methods for generating the statistics (meaning the range of numreical values) are currently prefered by active gamers.
IE: point allocation systems, set number of die rolls (and thus what sort of statistical scales are currently in vogue), and the like.
Pixies is intended to be a simple game, perhaps using tiles to generate a random board, so play should be the quick and easy. I really don't want to weigh the game down with unnecessary rules.
Thys the free-form approach does begin to sound interesting. But wouldn't that be too much work for the players? Which makes me wonder whether or not I should have asked whether players prefer "fixed" Stats (like those found in AD&D) or Stats which are generated in some other manner?
Kind Regards,
Kester Pelagius
On 11/7/2002 at 5:09pm, Mike Holmes wrote:
RE: Quick CharGen Question
Kester Pelagius wrote:He's not saying that at all.Random-feature character creation has a spotty history in role-playing and I don't think we've seen many games in which it's played a constructive role (but I do think that's possible). If you use it, you should consider what specific elements of play are enhanced by its presence. Look at the resolution system, the reward system, and other things - for instance, if Doing X relies on a particular feature, and Getting Better relies greatly of Doing X, then I as a player will be mighty pissed if the particular feature might start very poor due to a random roll.
It sounds as if I might have given the impression that I don't have a system in place and am wanting to generating statistics on the fly. That is far from the case. I do have a number of statistics in place, but I want to explore what sort of methods for generating the statistics (meaning the range of numreical values) are currently prefered by active gamers.
What he's saying is that, usually, randomization serves no purpose, other than, well, randomization. There is some sentiment that a randomly generated character is perhaps more of a challenge to play, or seems more "real" because of it's arbitrary generation, but even if so there are problems. Mostly what Ron's saying is that if you allow people to randomly generate characters, and they suck, then players will be dissapointed. If you then "fix" that bu making it less random, then the question becomes "why randomize at all?"
Further, most systems that randomize only radomize part of the character, and leave the rest to choices. This rarely makes sense. At best it gives a feeling that the parts randomly generated are due to Nature, and the rest due to Nurture, or character chioce. But that's pretty illusory.
The big downside to random generation is that players won't get to play the character they want to play, often. They will instead end up playing, "What they got." This is really offputting to many players.
In general, random generation as it's been done in the past is pretty passe. The other thing Ron is saying is that it doesn't have to be. There are likely ways to make it work for you. But not going with the old style version. You have to ensure that randomized abilities are all as interesting as each other. The randomization only makes them different.
IOW, if by random generation you mean roll up stats, then I think that most around here are going to come out against it. I strongly do.
That said, there are other alternatives to points. For example, mixed templates.
Mike
On 11/7/2002 at 5:33pm, Ron Edwards wrote:
RE: Quick CharGen Question
Hello,
I just remembered, delayed I'm sure by advancing senility, the perfect modern example for functional randomized character generation mechanics: The Dying Earth. One of the distinctive features of this system is that the random-method is (a) optional and (b) slightly quantitatively superior to the choice-method. Another distinctive feature is that effectiveness is not randomized, but descriptors and modes of the player's effectiveness are.
Kester, in general, "poll" posting doesn't tend to work well at the Forge. That's not a matter of policy, just an observation of what's happened in the past. I think most people here, myself among them, are not very interested in a slew of single-posts of "what I think," or "Yea-Nay" kinds of responses. I also think that such a thread's utility to a designer is dubious at best.
Best,
Ron
On 11/7/2002 at 5:51pm, Marco wrote:
RE: Quick CharGen Question
Point based--all the way (of the two). Mixed templates ... various "point pools" (something we do in JAGS to a degree), or other limiting factors can go a long way towards mitigating some of the precieved problems with "Here's a buncha points--go at it--" that GURPS and Hero can get accused of.
Something that I have enjoyed is dice for "back-story" (Mekton). When done right, rolling for your arch-nemesis, your relationship with your ex-lover ... I mean, in Mekton it was funny and outrageous--in the right place I could see that being a cool, random char-gen feature.
-Marco
On 11/8/2002 at 8:18am, M. J. Young wrote:
RE: Quick CharGen Question
Multiverser's design is freeform; that is, character creation is about defining the character in game terms, not about creating the character through game mechanics. As it's usually played as an I-game, it comes down to trying to represent yourself as accurately as possible in the game; but in those cases where you are playing someone else (or the referee is creating non-player characters) it is the same process: understand who the character is, then put him on paper as you've imagined him.
But against the overwhelming hue and cry, let me make the defense for randomized systems.
There is inherent in the point-based approach this unstated assumption: all characters are created equal, not in law but in fact. It suggests that the star football player and the valedictorian and the president of the student government association are all equally endowed but have strengths in different areas. I'm reminded of a joke made by a balding professor at the beginning of a speech, just after being introduced by the amply bearded long-haired host: "I figure we all get the same measure of hormones, and if he wants to use his to grow hair, that's his choice." The fact is that we are not all created equal, even when you account for differences. And when you look at the heroes of our stories, you realize that they are not equal, either. No one can argue that the nine companions of the Fellowship of the Ring are equally gifted. Clearly Gandalf and Aragorn are superior to the others. Take into account the enemies, and it is obvious that the gifts are not equally distributed. Luke, Leia, Han, Chewie, 3PO, R2, Ben, Lando, are not equals, nor are they equals with Vader, the Emperor, Darth Maul, the federation representatives. If these characters were equally gifted from the beginning, the story would be much less interesting.
Point based systems tend to create characters who are equally gifted at the start, except in those cases in which players either don't know how to use the system or intentionally create characters weaker than they might. Randomized systems capture at least some of this aspect of different strengths and different levels of overall strength, making some characters inherently better than others. That's more realistic; it also makes for better stories.
Obviously if the randomized system is geared such that it can create completely worthless characters, it's flawed. This is the real problem with most randomized systems: they create some characters of no value. They also tend to create mostly average characters. Fixes to this run the gamut from preventing any below-average characters at anything (a weakness itself, because sometimes the most interesting characters are the incredibly strong brute idiot and the brilliant weakling) to throwing out characters who don't meet certain minimums (which doesn't prevent a proliferation of average characters) to creating mostly superheroes in every area (again, eliminating the interesting balances between strengths and weaknesses). But there could be fixes that work.
One is a combined randomized/point based system, in which the range of points to be made available is established but each player must roll to determine where in that range his character falls. This gives the variation in total ability while giving the players control of where the weaknesses lie.
One which I think has been suggested worth pursuit is a sort of template and points system. If I understand aright, this would be a matter of selecting a basic character type for its specific strengths, and then spending a much smaller pool of points tweaking it--such as getting a stupid brute, and then deciding whether to make him a bit less stupid or a bit more brutish.
One is to provide for different die roll choices for different aspects of the character. For example, you might say that there are six scores which can be any number from one to twenty; the player is allowed to roll, once each:
• d20
• 3d6
• 3d4
• 4d4
• 2d10
• 2d4+12
and arrange them in any order desired. This gives a lot of different curves, assuring that at least one score will be at least 14, at least one will be not greater than 12, and so forth. It is still very random, but it has a bit more stability in the differences between characters. Clearly you will not get one character with all twenties and one with all twos.
Now, there are probably ways to design point-based systems that overcome this concept of equality; and there is probably a contingent of gamers who like the idea of starting on an equal footing (a very gamist idea, although not all gamists value it). Personally, I think that if you could find a way to get characters to define a character concept and identity and then translate that into game mechanics, that's the best way to get what you want--but it's a lot harder to do in a fantasy setting, because it's tougher to get the benchmarks for what's appropriate.
--M. J. Young
On 11/8/2002 at 2:58pm, Le Joueur wrote:
Not All It's Cracked Up to Be
M. J. Young wrote: But against the overwhelming hue and cry, let me make the defense for randomized systems.
There is inherent in the point-based approach this unstated assumption: all characters are created equal, not in law but in fact.
Actually, this myth is based on the assumption that all point-based systems require equal (or any) point cut-offs. If you simply all players to take however many points they desire, without limits (perhaps even without guidelines), this 'problem' disappears. What that means is that it is the assumption that all characters will be made with equal points which is both the assumption and the indicated flaw.
When we began designing Scattershot, point cut-offs was the first thing to go, for exactly the reasons you list. That resulted in a question from our play testers; what are the points for? That took a while to figure out. Ultimately, it resulted in the creation of the Sine Qua Non Persona Development Technique, a way of clarifying 'start with what is important.' When we took that back to the point system, we realized that the points then became an efficacy management tool and, when coupled with Sine Qua Non, an indicator of 'what the player wants to do.' Frankly working with both of those 'on the table' makes gamemastering a group whose Approach is inherently Self-Sovereign much easier. (You give everyone what they want and use the points to adjust spotlight time to suit player desires.)
M. J. Young wrote: Point based systems tend to create characters who are equally gifted at the start, except in those cases in which players either don't know how to use the system or intentionally create characters weaker than they might.
I've always argued that due to the unpredictability of the situations the characters will find themselves in, even experienced players don't always have as effective of characters as their points might indicate. A number of dysfunctional play-styles I have seen are founded on players leveraging the point-based system to force an unfair imbalance in spotlight time.
That's why I prefer to use points as an indicator of efficacy (can they do this?) and interest (it's what they want to do) rather some arbitrary measure of 'balance.' My typical example is Lois and Clark: the New Adventures of Superman. Only one of the characters has superpowers, far and away a point-based 'imbalance.' To spite that, the show is centered on a newspaper bullpen, a place where superpowers don't give you any edge. This causes spotlight time to be apportioned, not by point values, but by the contents of their characters.
M. J. Young wrote: Randomized systems capture at least some of this aspect of different strengths and different levels of overall strength, making some characters inherently better than others. That's more realistic; it also makes for better stories.
I believe it is a fallacy that realism "makes for better stories." I think 'stories' are made better when the participants receive a share of spotlight time to their liking and are empowered by the things they find interesting in their characters (not to mention the crisis-climax-resolution cycle and coloring on a theme to create a message via metaphor, but that's just personal bias).
M. J. Young wrote: But there could be fixes that work.
One is a combined randomized/point based system, in which the range of points to be made available is established but each player must roll to determine where in that range his character falls. This gives the variation in total ability while giving the players control of where the weaknesses lie.
I'm not so sure. I believe that would capture the worst of both systems. The dice can confound what a player may want (both by starving them of points and by being excessive). Keeping the point cut-offs magnifies this because it almost never comes out even. (Ever have those occasions where you're finally getting a character idea but you must spend almost as much time 'fine tuning it' to make up those last few points, either over or under?)
As I've discussed, I think the problem is alleviated when you throw out the cut-offs with the understanding that characters will definitely be unequal. If you place 'fairness' control in some realm other than point levels, you effectively avoid this problem.
You do edge close to a way to work a "randomized/point-based system:"
M. J. Young wrote: One which I think has been suggested worth pursuit is a sort of template and points system. If I understand aright, this would be a matter of selecting a basic character type for its specific strengths, and then spending a much smaller pool of points tweaking it--such as getting a stupid brute, and then deciding whether to make him a bit less stupid or a bit more brutish.
We're testing this: randomly determine which template to use. Each template narrows the choices of abilities (upon which any number of points may be spent so long and any with high degrees of efficacy show up in the retroactively generated Sine Qua Non). The resulting characters are not equal, but some level of niche protection (arguably what class-based systems are tailored to) is maintained.
M. J. Young wrote: Now, there are probably ways to design point-based systems that overcome this concept of equality; and there is probably a contingent of gamers who like the idea of starting on an equal footing (a very Gamist idea, although not all Gamists value it). Personally, I think that if you could find a way to get characters to define a character concept and identity and then translate that into game mechanics, that's the best way to get what you want--but it's a lot harder to do in a fantasy setting, because it's tougher to get the benchmarks for what's appropriate.
What you've described is the fundamental basis of Scattershot's Sine Qua Non Persona Development Technique. We describe initial Persona Development as a 'recipe' approach (take all these ingredients and combine them in whatever order), but in most cases the playtesters have gone Sine Qua Non to Persona Write-Up. The "benchmarks" rise explicitly out of Scattershot's Genre Expectations Technique. All players expect a fair (but not necessarily equal) share of spotlight time and 'chance to shine,' not because of point-based equality, but because the points are an aid to moderating 'who gets to go.'
I'd be interested in your reactions to the above mentioned Techniques, but to avoid 'thread hijacking' I warmly welcome you to post them in the Scattershot forum.
Fang Langford
Forge Reference Links:
Topic 2009
Topic 1662
Topic 2043
Board 22
On 11/9/2002 at 5:26pm, Kester Pelagius wrote:
responding slowly, but surely...
Greetings Mike,
Thanks for the input. So “totally random” character generation is out then? ;)
Mike Holmes wrote: The big downside to random generation is that players won't get to play the character they want to play, often. They will instead end up playing, "What they got." This is really offputting to many players.
In general, random generation as it's been done in the past is pretty passe. The other thing Ron is saying is that it doesn't have to be. There are likely ways to make it work for you. But not going with the old style version. You have to ensure that randomized abilities are all as interesting as each other. The randomization only makes them different.
IOW, if by random generation you mean roll up stats, then I think that most around here are going to come out against it. I strongly do.
Actually I didn’t quite mean randomization of this sort, then I could perhaps have been more clear, or at least less vague.
Of course any system that relies totally on random rolls will seem, random. What I was thinking of was more the sort of system where a part of character generation, namely the core stats, are generated by some random (dice rolls) or directed (point allocation) method by the players.
Of course, in a game, the idea that a character “might suck”, in my opinion, is a flaw within the rules themselves. I’m sure we can all instantly think of at least one system of this sort. It’s how the ideas of “Munchkinism” and “Minimaxing” came about. Such rules deliberately set out to provide massive positive modifiers for *high rolls* while pointedly penalizing (or not even taking into account) *low rolls* made during the character generation process.
I assure you Pixies will not have anything of the sort. There wont even be the concept of “advantages vs. disadvantages” you will have Traits, which will either be beneficial or not, dependant upon how the player has their character apply them during actual play.
Of course, since there are *only* Pixies in the game, I thought I’d re-examine the character generation I started to put into place. It felt, to me at least, like it might be too complicated; for all that I am designing the game to use nothing more elaborate than standard six-siders that can be found in any Monopoly or Parcheesi game.
Though that may change, too. Which is why I thought to inquire about what sorts of systems are preferred. I would like everything to fit together seamlessly. (insert background music of Aerosmith’s “Dream on” here)
That said, your input has been much appreciated.
Kind Regards,
Kester Pelagius
On 11/9/2002 at 5:43pm, Kester Pelagius wrote:
responding slowly, very slowly, but surely...
Ron Edwards wrote: Kester, in general, "poll" posting doesn't tend to work well at the Forge. That's not a matter of policy, just an observation of what's happened in the past. I think most people here, myself among them, are not very interested in a slew of single-posts of "what I think," or "Yea-Nay" kinds of responses. I also think that such a thread's utility to a designer is dubious at best.
Greetings Ron,
Didn’t intend for my post to sound like a poll. Though, upon reflection, I suppose it might seem that way since I am doing the proverbial “fishing for input” bit. But, in all honesty, it never occurred to me that all I might get was a bunch of “Yea-Nay” responses from my fellow denizens. In fact, so far, your post included, I’ve received some pretty good informational feedback.
Next time I’ll preface my query with a “In which direction would you like to see (name of game/system) developed.”
That said, thanks for the input.
Kind Regards,
Kester Pelagius
On 11/9/2002 at 9:33pm, Kester Pelagius wrote:
still responding slowly...
Greetings Marco,
I believe this is my first post to you, nice to meet you. Hope all is well in your part of world this fine day.
Marco wrote: Point based--all the way (of the two). Mixed templates ... various "point pools" (something we do in JAGS to a degree), or other limiting factors can go a long way towards mitigating some of the precieved problems ... Something that I have enjoyed is dice for "back-story" (Mekton). ... in the right place I could see that being a cool, random char-gen feature.
Yes, I never did understand how the numerical starter values were arrived at in certain games, GURPS and James Bond 007 (old Avalon Hill RPG) among them. A lot of old game systems just sort of provided you with template systems that, while great starters, didn’t really make any effort to explain the how’s and why’s. I never really liked that.
A mechanic for rolling “back story” (or just plain story) sounds intriguing. I am not very familiar with Mekton. Could you possible point to a link with a good overview of this mechanics essentials?
Looking at the JAGS basic rules PDF file… so, there are core stats, from which secondary stats (Enhancements, Defects, and Traits) are derived… already that sounds a bit more complicated than what, in my opinion, would be optimal for the *feel* of the sort of game that Pixies should be. Though depending Traits from core stats might work.
Ah… players purchase Skills. I overlooked that in my initial question, didn’t I? There are CharGen systems that use a set # of random dice, point allocation, and then systems which sort of tie stats into acquired traits/skills which determine the actual numerical value of the stats during play. Interesting, but perhaps a bit too much for Pixies to have to worry about. Still something to think about.
Kind Regards,
Kester Pelagius
On 11/9/2002 at 9:38pm, Kester Pelagius wrote:
RE: Quick CharGen Question
Greetings M. J. Young,
That was a very eloquent and informative introductory post. So fine a post was it that, in my humble opinion, I think it could be excised from this thread and expanded upon to form an article cum guide to the basic principles of character generation.
A grade-A excellent effort!
Apologies for truncating it so badly.
M. J. Young wrote: I think that if you could find a way to get characters to define a character concept and identity and then translate that into game mechanics, that's the best way to get what you want--but it's a lot harder to do in a fantasy setting, because it's tougher to get the benchmarks for what's appropriate.
You make many interesting points...
Alas, and this in no wise is meant to malign your post, I do not think all your points fully apply to Pixies, at as the game stands at the moment. But only because there currently is only the one character type extent; the Pixie. For instance your remark, “There is inherent in the point-based approach this unstated assumption: all characters are created equal, not in law but in fact.” seems a bit extreme. I realize that most players of our modern fast-food generation want everything two seconds ago and characters that are all that and a side of fries but, from what I recollect of most point allocation systems I have seen, wasn’t it more a matter of setting a common difficulty standard?
Of course some systems probably do leave the proverbial door wide open for potential abuse. Then again if the system is designed to be a “combined randomized/point based system, in which the range of points to be made available is established”, as you point out, is perhaps a better approach. Too, I am intrigued as to why you feel that a CharGen system “worth pursuit is a sort of template and points system” for Pixies?
There is always room for expansion, especially in the ‘creation of non-player character’ types department, but that is perhaps a discussion best left for another thread.
Kind Regards,
Kester Pelagius
On 11/11/2002 at 1:34am, M. J. Young wrote:
RE: Quick CharGen Question
I'm pleased that my post was of some help. I can respond to most of your comments by saying that at no point was I intending for anything said to apply specifically to Pixies, but rather to be a consideration of the strengths and weaknesses of point-based versus random character generation in the abstract. I perceive that you did ask for opinions related to the game, but save Ron's none seemed to consider that (and Ron's was that other elements of the game should be considered in that connection).
You wrote: there currently is only the one character type extent; the Pixie. For instance your remark, ?There is inherent in the point-based approach this unstated assumption: all characters are created equal, not in law but in fact.? seems a bit extreme.
In this world there is only one character type extant: the human. Humans are equal in law, but would any argue that Stephen Hawking, Nolan Ryan, Donald Trump, President George W. Bush, Reverend Billy Graham, and Angela Lansbury are equal in fact? And even if after comparing strengths and weaknesses we decided that their abilities all came out to the same amount, would we then be willing to include ourselves and our friends on the same list? If all pixies are the same, then you don't need character generation--you just say, "Here's a one each pixie character sheet, please write your character's name at the top" and you're done. The game is made interesting because characters have differing strengths and weaknesses, but also to some degree because they have different total strength.
The idea of the template and points system was also "worth pursuing" in general. There are more games in progress at the Forge at any moment than I can fully learn, so I try to keep my investment in each limited to a very general understanding of what is being attempted and my comments to very general concepts. However, what a templates and points system allows you to do is devise character concepts, whether as strong as classes or as weak as primary areas of ability, and use them as a starting point. Thus you could have templates for the pixie who tends to use force, versus the one that tends to think, or negotiate, or outmaneuver. Each of these could represent a core set of abilities, possibly associated skills (don't remember whether Pixies has skills) and a splash of points that would be used to customize them. But I'm not at all convinced this is "right" for Pixies; it's just an option tossed out to help you think of how to balance equality with inequality (which is in essence what point based seems to be about, Scattershot excepted).
Again, I hope this is helpful.
--M. J. Young