The Forge Archives

Archive => GNS Model Discussion => Topic started by: Balbinus on December 05, 2001, 05:27:00 AM



Title: Statting, NPCs and GNS
Post by: Balbinus on December 05, 2001, 05:27:00 AM
Hi all, a brief exchange between Mike Holmes and JBurneko got me thinking about this.  Basically, the pre-assigning or calculation of stats for NPCs, is it related to GNS?

Mike said that, even though he is a simulationist, he rarely stats his NPCs.  Rather, he assigns what seems appropriate at the time.

J does stat NPCs, and likes systems like Sorceror in part because they make it easy to do so on the fly.

I almost always stat them, if they are at all important I create them as if they were PCs.

Here's my thinking.  From a gamist perspective statting NPCs is necessary (or at least very important) so as to create a stable challenge, a "fair" challenge where the GM won't just make the NPC harder or easier as the conflict requires.  An objective, predesigned set of NPC stats makes the NPC a fixed quality in the game to be interacted with against which characters can test their own abilities.

For a simulationist, statting is not so necessary.  What is important is the simulation of genre or reality (which is simply another form of genre).  NPCs may be statted or not depending on whether the GM finds that helpful in creating a believable, well-simulated environment or not.  Simulationism per se is not really relevant then.

With narrativism, I would suggest full statting to be counterproductive.  The NPC has a role in the story, hard stats may dictate story-negative outcomes (the NPC master fencer whiffs his parry as written down and does a story-damaging prattfall).  Some idea of the NPCs role is important, but actual full stats may just get in the way.

What do people think, is there a GNS bias to this or at least a GNS consideration?


Title: Statting, NPCs and GNS
Post by: Marco on December 05, 2001, 07:24:00 AM
That breakdown seems pretty fair to me. I stat when I feel like its important to give the players a fair-fight. For superhero games (or, say something like Mortal Kombat genre) it's important for the players to feel that their equals in the story are a roughly fair fight (Gamist).

I consider myself simulationist and I rarelys stat NPC's (the excpetion being members of the player's team which I flesh out a little more just to save headaches).

-M.



Title: Statting, NPCs and GNS
Post by: mahoux on December 05, 2001, 08:57:00 AM
To stat or not to stat...

I fall into that category of "well, it depends".  I do stat all my major baddies and major NPCs, of course.  As far as other NPCs, it really does fall into the realm of what their purpose is in the story.  The posting about a die failure by an NPC who should be moving the story along is a great point.

It's a lot like making chili.  People start with most of the same ingredients, then add the "family secret" and season to taste.  Unless it's Iron Chef chili - Your ingredient is Suckling Pig.  Allez Cuisine!


Title: Statting, NPCs and GNS
Post by: jburneko on December 05, 2001, 11:27:00 AM
I don't know how GNS relates to a planning technique such as stating but I can at least explain WHY I like to stat my NPCs.

First I would like to state that I don't think stating is counter productive for Narrativists.  Narrativists are concerned with story creation, not story preservation.  So, if there's a master swordsman who "fails" a roll all kinds of Narrativist techniques allow you to run with that.  Fortune in the Middle allows you to alter WHY the swordsman failed.  It might be because another character pushed him at that moment.  In this case the "masterness" of the swordsman is preserved despite the failing roll.  Also, this master's one failure might become a key plot point.

Okay, now that, that's out of the way, I'll explain why I personally like to stat NPCs.  It's out of respect for my players.  My players are forced to follow a set pattern of rules and regulations to get their characters to the point they are now.  Why shouldn't NPCs?  I'm not saying I create an NPC from scratch and then improve them like a PC would but rather I like to have my NPCs statistical distribution be believable based on the average PC of similar background and experience.

When I was running D&D3E I found this to be the bigest nusance.  I wasn't content just saying the guy has 5 ranks of diplomacy.  The reason being is that between class level and intelligence the number of skill points is fixed.  And often I would find that if I did the ENTIRE skill distribution in a manner similar to that of an average PC then I wouldn't have enough points for 5 ranks in diplomacy.

A game like Deadlands is a little more flexible because there isn't a set advancement process.  Who knows how much XP this NPC has sunk into Deftness and Shootin'?  So it's a little easier for me just to say I think he has a Deftness of 2d8 and a shootin' of 3 because those values don't effect anything else nor do they progress in a predetermined fashion.

Where a game like Deadlands falls apart is that there IS a set baseline distribution that can't be mentally taken into account very easily.  This is why I'm so thankful for the archetypes.  If I need a baseline preacher I take the archetype and then tinker from there.  So, I'm leary of just assigning a bunch of Traits and Apptitues off the top of my head because I don't know how far off those values are from the base distribution.

Finally, a game like Sorcerer DOES have a base distribution that I can keep track of mentally.  Initially, the three main stats add up to 10.  So, I know that if my NPCs total stats hover somewhere around 10, I'm okay.  A little more or a little less is okay, but if they add up to like 25, I know I'm probably way off base.  But three scores that add up to 13 for a pretty powerful NPC is probably not too far off.

Okay, so you're thinking, fine then just stat up your Major NPCs.  That's really not good enough for me because in my games, at least as of late, you don't know when a minor NPC will BECOME a Major NPC mid-game.  In the last Deadlands game a cultist, a two attourneys, a judge and an entire jury ALL got elevated from minor background NPC to forefront major NPC in the span of about 15 minutes.  I was very edgy because I hate assigning stats on the fly for the above reasons.  I would LIKE for there to have been more die rolling and a little better understanding of these NPCs in that trial but generally if I don't think I can "fairly" stat an NPC I turn to Drama mechanics because I trust my intuitive sense of thematics more than I trust my intuitive sense of statistics.

Had this been Sorcerer I would have been able to stat all of those NPCs in less than 5 minutes.  Oh, and by the way, I do stat both thematically and "in character."  So often the numbers help me get a base handle on the nature of the character.  So the faster I can stat and tinker the faster I can formulate a mental image of the character which means that ultimately I present a better roleplayed presentation of that character.

Jesse


Title: Statting, NPCs and GNS
Post by: Ian O'Rourke on December 05, 2001, 11:45:00 AM
Wow, Jesse, your role-playing life must be one big statting nightmare! :smile:

To be honest I rarely stat unless I envision tha characters coming into combat with the characters, and even then I'll probably not stat as much as I should.

It's all irrelavent ain't it? Why would you need to know the minutia of every NPC's skills. He's a competent psychologist, what's important is his place in the story, goals and 'character' - when is he going to roll his psychology skill?

To be honest, I've never seem to roll much NPC skills (other than combat as I've said). It's all just talking and role-playing.

Of course, this may be criminal.


Title: Statting, NPCs and GNS
Post by: jburneko on December 05, 2001, 12:01:00 PM
Quote

On 2001-12-05 14:45, Ian O'Rourke wrote:
Wow, Jesse, your role-playing life must be one big statting nightmare! :smile:


Yeah, actually, it is.  It's one of my RPG hangups that I haven't been able to get around.  I just have trouble saying I think this 3rd Level Rogue has 5 points in Hide when if you yourself as a player were to create a 3rd Level rogue and distributed your skill points in a "believable"* manner you'd discover that you'd be hard pressed to have 5 ranks of hide.

*By "believable" I mean in the manner the average player would.

Quote

It's all irrelavent ain't it? Why would you need to know the minutia of every NPC's skills. He's a competent psychologist, what's important is his place in the story, goals and 'character' - when is he going to roll his psychology skill?


Yes, and no.  I use dice rolls when I'm wafling on a roleplaying point.  If the appropriate reaction/response for a given character doesn't leap to mind or if I'm stuck between two equally valid plot twists.  I might roll the psychologist's psychology skill to see if he's perceptive enough to pick up on a point that I think might change his behavior but I'm not entirely sure if that behavior change is apropriate at the given moment.  

I do this even when I'm playing as s PC.  When I can't make up my mind if my character does something or not I pick a related skill or attribute, assign myself a TN and roll for it.  

Basically, I use stats often as the tie breaker between two equally valid plot points, themes or decisions.  But like I said, that's just me.

Jesse


Title: Statting, NPCs and GNS
Post by: Mike Holmes on December 05, 2001, 12:34:00 PM
Actually, to the extent that a game is a well written Narrativist game, I can see having the stating done right as being a boon to the narrativist. It should be. Though not absolutely necessary by any means.

For a Gamist audience I try to stat out more, as that increases the feeling of the level playing field. But even then I shy away from it. I play Rolemaster, remember? If I had to stat out even just the most important NPCs it would take days. And then when would I find time to post here. :wink:

Instead, for Gamist games like RM or D&D I take only the pertinent stats and enumerate them, often coming up with them on the spot. Seems wrong to give that 3rd level rogue 5 levels of hide? Then give him less. Is it that you are afraid of violating some development rule? You're the GM, you get to do that. I find that in the end the only thing the gamist cares about is whether or not the opponent is still up and fighting (or bargaining, or strategizing, whatever; I do not mean to imply that Gamism is only about fighting).

And that's all only if you have a gamist concern. As stated, for us simulationists, as long as I can come up with a believable reason for the character to have a particular skill or stat or whatever, who can gainsay me? Of course lord Gobash is an excellent swordsman, he studied under Kandar at the academy. "ooooh", go the players never having heard of Kandar before (and I make a note to flesh out this legendary teacher later). That all said, they seldom ask why an orc has 65 hits in the middle of a battle instead of 35 if I give him such. If they do, I'll probably say something really deep and creative like, "He was big." Goes a long way.

For the Narrativist playing a non-narrativist game like D&D... why are you playing D&D? Play something Narrativist. I've never seen a really Narrativist game where you had to worry about incorrectly assigning stats. Gave out 12 points of stats to a character in Sorcerer? Ron will not come and stalk you down. This is just one kickass opponent.

Even then, if you just make up a number out of thin air in a Narrativist game I think that probably works just as well. Oh, Sorcerer Theo? He's got a ..(scrabble around and pick up some dice) .. four, yeah, that's it, four Will. Descriptor? Oh, uh, he has great concentration. And then begin a beautiful narrative from there. Actually, the idea that Theo has Great Concentration says a lot more about him to me than the four dice of will. These are the things it pays to come up with for NPCs for Sim or Nar. What is the character like. Make stats to fit as is necessary.

I didn't see a single stat on any of the characters in the Art-Deco Melodrama set-up. Would it be that much more usable if they had been included? All the utility was already there.

At least that's my opinion, I could be wrong (he said trying desperately to sound like Dennis Miller).

Mike

[ This Message was edited by: Mike Holmes on 2001-12-05 15:36 ]


Title: Statting, NPCs and GNS
Post by: Ron Edwards on December 05, 2001, 02:04:00 PM
Hello,

I think that this issue is one of those especially-fun ones, in which ANY mode of play (G, N, S) can include a lot of diversity. In other words (infuriatingly), "it depends."

Say I were going to run Champions, and believe me, I know from Champions. Full individual writeups for the mooks? No way. On the other hand, for Sorcerer, Maelstrom, or The Whispering Vault, I do it, and it's a lot of fun. Note that another GM - perhaps one as rabidly Narrativist as I am - might be entirely the opposite with exactly these games.

I think we're looking at a different set of variables from GNS, involving (a) time, (b) personal takes on what is necessary to know prior to play, and (c) various games' degree of layering in terms of character Currency components.

Best,
Ron


Title: Statting, NPCs and GNS
Post by: Gordon C. Landis on December 05, 2001, 02:15:00 PM
Quote

On 2001-12-05 08:27, Balbinus wrote:
Here's my thinking.  From a gamist perspective statting NPCs is necessary (or at least very important) so as to create a stable challenge, a "fair" challenge where the GM won't just make the NPC harder or easier as the conflict requires.  An objective, predesigned set of NPC stats makes the NPC a fixed quality in the game to be interacted with against which characters can test their own abilities.

Having recently emerged from/endured the Gamist/Competition debate thread, I feel the need to point out that "fairness" is only a POSSIBLE desire for a Gamist.  Some Gamist groups/GMs no doubt like things to be as you state, but one can quite easily be a Gamist who makes the details up in order to ensure the challenge is "right" (which may have nothing to do with "fair") at any given moment.  The details of the restrictions on how such details are made up (and for that matter, what type/degree of challenge is "right") may be explicit in the game, or they may simply be a matter of the oft-unspoken "group contract".

My claim would be that having "stats" (which have to be considered in the broadest sense) that are utilized as the challenge plays out is important to Gamism, as that is the field upon which the competition takes place.  But the details about how/when they're determined, when it's OK to fudge 'em, and etc., will vary widely.

The GNS question re: stats is (I'd think) "WHY do we have them?"  They exist in Gamism to foster competition, they exist in Sim to enhance player engagement, and they exist in Nar to fascilitate the creation of story.  If the stats themselves, the way/time you determine them, and/or the way you use them is incompatible with those goals, they're bad - if they support 'em, they're good.

We can probablly find some general trends, like "too many stats can get in the way of good story creation", and that could be quite interesting.  But I'm reluctant to go too far with categorization here - properly used, a wide variety of "stating techniques" can support any of the GNS goals.

Gordon


Title: Statting, NPCs and GNS
Post by: Epoch on December 05, 2001, 02:22:00 PM
Jesse,

Color me astounded.  You find it improbable that a 3rd level Rogue has 5 points of Hide?  So do I, but that's because he could have 6 points of Hide (before stat bonuses, of course), and I've never known a Rogue not to max out his Hide skill.

Anyhow, that's obviously a minor point.

Everyone and everybody,

In a one shot, I don't believe that I've ever statted any non-PC's, ever, regardless of system.  Sometimes, I'll jot down a couple of important numbers (like modified attack bonus in D&D) beforehand.  In a campaign, I'll usually-but-not-always stat major NPC's.  And, hell, my last two games have been light systems.

Just one more data point for the curve.


Title: Statting, NPCs and GNS
Post by: Mike Holmes on December 05, 2001, 02:50:00 PM
Quote

On 2001-12-05 17:15, Gordon C. Landis wrote:
Having recently emerged from/endured the Gamist/Competition debate thread, I feel the need to point out that "fairness" is only a POSSIBLE desire for a Gamist.  

Yep, replace fairness with that very long thing that Gareth replaced it with in his definition. Something about consistent feedback. That's why stating might be important for the Gamist player.

Quote

But I'm reluctant to go too far with categorization here - properly used, a wide variety of "stating techniques" can support any of the GNS goals.


Sure. To the extent that the system fits your goals it is a tool to use in furthering the game. But I find that since many pre-prepped stats go unused and since I seem to be able to make up satisfactory stats on the fly, it's a waste of time to do them before hand. YMMV.

Mike


Title: Statting, NPCs and GNS
Post by: Mike Holmes on December 05, 2001, 02:51:00 PM
Quote

On 2001-12-05 17:22, Epoch wrote:
And, hell, my last two games have been light systems.


Which means that you felt it unnecessary to stat NPCs, or you did stat a lot of NPCs because it was so easy?

Mike


Title: Statting, NPCs and GNS
Post by: Epoch on December 05, 2001, 02:59:00 PM
Which means it would've been easy to stat 'em, but I still couldn't be bothered.

I think this has something to do with me being a lazyist GM.  ("Makes a decision based on picking the solution that requires the least effort.")


Title: Statting, NPCs and GNS
Post by: Ben Morgan on December 05, 2001, 03:54:00 PM
Quote
I think this has something to do with me being a lazyist GM. ("Makes a decision based on picking the solution that requires the least effort.")


Ah, so that's what my style is called.

I usually can't be bothered to come up with full stats for NPCs. Hell, the Vampire game I ran there were over 50 fairly regularly occurring Vampire NPCs - NYC was undergoing a population explosion following the ousting of the Sabbat (a story idea I managed to come up with 5 years before WW did) - not counting mortals passing through the story. For each, I detailed only what I absolutely needed to keep track of: their Generation (mainly to keep track of "family trees"), their actual age, their apparent age, years since their Embrace, and a little descriptive blurb (stuff like appearance, roleplayng hints, etc). Then I tried my damndest to be consistent about assigning them pools for various things: "If Sir Adrian had a Firearms pool of 4 last week, it should probably be 4 this week as well."

For me, I follow a very simple rule as far as how to treat any piece of information in the Game World: Until it's been stated, it's malleable. However, once it's been stated, It's set in stone, until there's an in-game reason for it to change. This goes for NPC stats, motivations, relationships, backstory events, anything.




Title: Statting, NPCs and GNS
Post by: Bankuei on December 05, 2001, 05:02:00 PM
  I personally stat based on the character's importance and likelyhood of interacting with the PCs.  My major NPC's may be so powerful, or so unlikely to come in a direct contest with any of the PC's to ever merit me having to give stats to any of them.  Ditto with mooks or nameless npc's that I don't intend to get into the conflict, I just note what descriptors I need to know and general motivations.

  Generally I stat according to how often the NPC is likely to be in conflict or competition with the players and how narrativist or gamist the campaign as a whole is.  I get some twisted satisfaction from a binder full of fullstated ready to go npc's...who knows why?  Either that or it gives me a good excuse to sketch up another messed up individual to play with the PCs. :smile:

Bankuei

   


Title: Statting, NPCs and GNS
Post by: Epoch on December 05, 2001, 05:07:00 PM
The family tree thing is important.

I just realized yesterday that, if one of my PC's dies, the following are the most likely heirs to the throne:

His traitorous-half-brother-in-exile
His full aunt
His double-half-aunt
His eldest half-uncle
His eldest half-uncle who has an interest in the throne
His eldest surviving half-uncle who is unquestionably legitimate
His eldest half-aunt who is also unquestionably legitimate and is older than his eldest surviving uncle who is unquestionably legitimate
The son of his eldest (now-deceased) unquestionably legitimate half-uncle (who was older than his eldest half-aunt)
His double-half-first-cousin
His half-first-cousin-plus-half-nephew

...is it any surprise I don't have time to stat the NPC's?  I have to work on my mathematical proof that it's not possible to graph the family tree in two-space!

[ This Message was edited by: Epoch on 2001-12-05 20:08 ]


Title: Statting, NPCs and GNS
Post by: lumpley on December 05, 2001, 06:23:00 PM
Hey, would anybody mind if we expand the question?  When is it important to stat PCs, and when can you get away with not?  The same circumstances?  Different ones?

Also, I wasn't counting, but are we seeing the old DIP/DAS [Develop In Play / Develop At Start] split?

Sometimes the quality of discussion around here is so high I can't even blink.

-lumpley Vincent


Title: Statting, NPCs and GNS
Post by: Laurel on December 06, 2001, 09:34:00 AM
Quote

On 2001-12-05 21:23, lumpley wrote:

Also, I wasn't counting, but are we seeing the old DIP/DAS [Develop In Play / Develop At Start] split?


I think so.  I'm a proud member of the DIP squad, and don't stat my NPCs.  I'm 'fair' to my players in the sense that I try to create antagonists that challenge but don't destroy them, and rely more heavily on storytelling than dice rolling to keep things balanced.  

Quote

Sometimes the quality of discussion around here is so high I can't even blink.


I feel that way too and like a little fish who thought she was a big fish til she jumped in this particular little pond :smile:


Title: Statting, NPCs and GNS
Post by: Blake Hutchins on December 06, 2001, 09:42:00 AM
I've adopted a technique borrowed from the NPC write-ups in Tribe 8, namely to identify the highlights of the character, whether they be stats, skills, dice pools, personality aspects, or whatever.  I try to keep them to three for simplicity's sake.  If I have a major scene approaching, I try to map the limits of what I'm going to improvise, more as guidelines to keep me in synch with the story than anything else.  My rule of thumb is to keep any stats or numbers on one side of a one page "cheat sheet," with everything in legible, easily scanned format.

I have a friend who runs DnD 3e and devotes at least eight hours to preparation for a single session.  I'm in awe of that, but there's no way I have the time or inclination to do that amount of work for a weekly game.

Best,

Blake

[ This Message was edited by: Blake Hutchins on 2001-12-06 12:44 ]


Title: Statting, NPCs and GNS
Post by: Ron Edwards on December 06, 2001, 09:52:00 AM
Blake,

I follow that same strategy for Hero Wars and for most other games, when possible. In Hero Wars, it's especially easy because character game-features are a series of descriptive words anyway. In other games (RuneQuest is a good example) it's practically impossible; one is almost certain to violate the system-integrity, which I like to preserve.

After a character has been in play, I then find it very easy to round him or her out with full game-writeups, adding stuff that got improvised in or revising the original idea, or whatever. After that point, the NPC is "cemented" and becomes a solid piece of what's going on, in system terms.

Of course, sometimes I have a really solid NPC all set before play, just because I liked the character and enjoyed working out the details. But I've found that well into the play series, about half of the really solid NPCs were "vague-then-cemented," and about half of them were "cemented-before-play." It seems to go either way.

Best,
Ron


Title: Statting, NPCs and GNS
Post by: jburneko on December 06, 2001, 10:04:00 AM
Quote

On 2001-12-06 12:52, Ron Edwards wrote:

In other games (RuneQuest is a good example) it's practically impossible; one is almost certain to violate the system-integrity, which I like to preserve.


Ah, thank you Ron, that's the phrase I was looking for.  I don't like to violate the system-integrity, either.

Jesse


Title: Statting, NPCs and GNS
Post by: Mike Holmes on December 06, 2001, 10:57:00 AM
Quote

On 2001-12-06 13:04, jburneko wrote:
Ah, thank you Ron, that's the phrase I was looking for.  I don't like to violate the system-integrity, either.


?!?

I go into Champions battles with nothing but a character concept (I remember making up a character, The Ape, after a battle had started, and the bad guys were going down too easily). Interestingly, from a balance POV in battle, any good Champions GM knows that point totals have little to do with the outcome of a battle. It's Active Point limits for powers that keep characters in line. I design powers on the spot to make them challenging but not overpowering. I just make up each stat as I need it.

I guess I'm violating the system integrity, whatever that means. But it's fun. Where is the harmful part? Even when my Players figure out that I'm making something up? Make it up before, make it up during, what difference is there? Mistakes? Well, that character is an exception to the rules (might make for an interesting plot development). If I really goof up...well, oops, have to make a small change. That's the price you pay for making prep time a tenth as long and ten times as interesting. IMO.

Mike


Title: Statting, NPCs and GNS
Post by: Ron Edwards on December 06, 2001, 11:25:00 AM
Mike,

We're talking about personal preferences based on time and creative styles, not general guidelines. This is exactly what I meant way up in my first post on the thread, that GNS perspectives aren't going to provide guidelines for one's preferences in this regard.

No one said a word about what you should or shouldn't do. This "system integrity" thing is no more than a desired personal-constraint that Jesse and I share.  Cool yer engines.

Best,
Ron


Title: Statting, NPCs and GNS
Post by: Ian O'Rourke on December 06, 2001, 01:33:00 PM
I'd actually go as far as to say that having to worry about too much 'system integrity' is what puts me off some games. Basically, third edition D&D has so much relating elements that hang off each other it drives me insane.

I find that game hard to play without exhaustive stats.

Anyway, slightly off topic for the thread so I'll stop :smile:


Title: Statting, NPCs and GNS
Post by: Mike Holmes on December 06, 2001, 02:20:00 PM
Quote

On 2001-12-06 14:25, Ron Edwards wrote:
We're talking about personal preferences based on time and creative styles, not general guidelines.


Yes, this is, of course, about preferences. And I apologize if I sounded didactic. I just want to point out that people can save a lot of time if they just skip the stats step, and it has never, under any circumstances, been a bad thing for me.

That having been said, I like to stat. From a theoretical POV. I don't actually do it much anymore, but I used to spend hours just writing up characters which I had no idea if I was ever going to use. Now that's a waste of time (though a fun one). What it did for me is make me adept at figuring out how to stat on the fly. Which I find infinitely more useful than pre-statting.

Mike


Title: Statting, NPCs and GNS
Post by: Le Joueur on December 06, 2001, 04:22:00 PM
Quote
Ron Edwards wrote:

Quote
Blake Hutchins wrote:

I've adopted a technique borrowed from the NPC write-ups in Tribe 8, namely to identify the highlights of the character, whether they be stats, skills, dice pools, personality aspects, or whatever.  I try to keep them to three for simplicity's sake.

I follow that same strategy for Hero Wars and for most other games, when possible.

Me too.  I quite agree with you Ron.  In fact, that is how I write that gamemasters should do it (not the only way, but a good, non-intuitive, starting point to learn from).  That is also why I like systems that do not have much ‘cascade’ – stats and ability scores resulting from calculations resulting from calculations (etc.) and exclusively write such.

Fang Langford

[ This Message was edited by: Le Joueur on 2001-12-06 19:25 ]


Title: Statting, NPCs and GNS
Post by: Joe Murphy (Broin) on December 07, 2001, 05:54:00 AM

I use three different NPC methods, I think. I haven't yet used much of the GNS in my games, so they may not be all that analyzed.

Whenever I come across a new game, I tend to stat up a few dozen characters. This familiarises me with the ins and outs of the system and sometimes shows me chargen pitfalls my players might encounter.

Often, these test characters become a folder of NPCs I can pull out in-game. As I've spent an hour with each character, I'll have an instinctive feeling for their style, voice, techniques, etc, and they haave fully-developed stats. They'll often have sketches and ties to other NPCs noted.

Less important NPCs (the bartenders and shopkeepers) tend to be devloped completely on the fly. These tend to have two personality traits. One trait may be, say 'Friendly'. The other trait directly opposes that trait, or twists it somehow. The friendly shopkeeper may be paranoid till a bond of trust evolves. The cop may be dutiful but weak-willed. The sailor may be optimistic, but with a tragic past. These NPCs can develop from those seeds quite nicely.

If actual statistics for these NPCs are required, I tend to give 'em a rating in Thinking/Talking/Fighting. These approximations work fine for simple interactions.

Finally, I sometimes have my players develop NPCs. In a recent supers game, I had my players begin the session by scribbling up 'a superhero military team' or 'a villainous team themed around the circus' or whatever. We recorded these alphabetically on index cards.

These NPCs would often later be encountered, or referred to in newspapers, etc. These NPCs have simple, approximated stats as before, and were just typically just background flavor.

Joe.