On The Fly

Posted by on Feb 7, 2008 in Musing, RPGs | 21 Comments

Tabletop RPGs traditionally squander — utterly — the opportunity to make character creation an exciting part of a game’s actual play.

Sure, in lots of RPGs the process of rolling up a character is an entertaining mini-game of its own. It has the benefit that you can do it without the rest of the group around. Think back back to all those Traveller characters you created, just goofing around by yourself. Remember how half of them died in the process? Made the ones that survived feel that much more awesome, didn’t it?

But even for that, think about how few RPGs — and precious few mainstream ones — do much of anything to directly tie character generation to the action of the game itself.

Here’s what I want to know: Why the hell not? And what would be better?

Ask just about anyone with an adult opinion about stories whether character or plot is more important and here’s what they’ll tell you: character is way more important than plot. The reason is simple. Plot that’s worth anything flows from character. Plot without character amounts to porn; at best, you appreciate it on a mechanical level.

That truth — that the only decent plots flow from character — makes the vast and traditional chasm between characters and adventures in RPGs that much more sad. And the chasm is both vast and traditional. It’s vast because the only feature of character typically taken into account when designing an adventure is level-appropriateness. It’s probably traditional for commercial reasons. I mean, it would have been a lot harder to sell a half-gazillion copies of The Keep on the Borderlands if it had to be tailored to your own PCs.

So, what to do?

Come at the problem from a different direction, that’s what. In a novel or film, every character’s introduced as a cipher. Nothing is known in advance of the character showing up and starting to do things.

So why not just start playing the scenario — before creating characters — and let the players define their characters as the adventure goes along? The player describes his character’s appearance the first time someone looks at him, defines his ability to schmooze the first time he has to fast-talk an NPC, describes his ability to fire a gun accurately the first time he pulls the trigger on one.

Would there be rules? Absolutely. But if you think it through, they wouldn’t be as different from your standard RPG character creation systems as you might think at first.

The benefits to this approach could be awesome.

You can start playing right away. On-the-fly character creation completely eliminates the worst RPG start-up annoyance of them all: That you can only start playing once the slowest player is done creating his character.

Indie RPGs have been much better than mainstream games at tying character creation to gameplay. I wouldn’t be surprised to discover some indie release has a system just like what I’m describing here. If you’re familiar with one, please share (and describe!) in the comments.

Everyone has expanded and concrete opportunities to tie their characters’ capabilities and characteristics to the dramatic questions of the game. Each character’s relevance increases because its nature flies out of the specific challenges the adventure poses. And this approach does nothing to limit the GM’s traditional ability to create scenes and scenarios that riff on what the players define about their characters.

Buyer’s remorse is greatly curtailed as a character creation consideration. The most common cause — that players’ choices become uninteresting in the face of the storylines that actually develop in play — is simply eliminated.

There’s at least one obvious objection to the idea I’m proposing: Wouldn’t it allow players to define their characters’ capabilities, in ultra-cheesy fashion, explicitly to overcome the first obstacles the game throws at them?

Yeah, maybe. And so what?

Don’t the first-act obstacles thrown at the protagonist in traditional stories serve mainly to define character? In a tabletop RPG, the fact that those initial obstacles don’t necessarily tell us anything about our heroes strikes me as a much greater sin.

In any case, this objection will primarily be raised by those who value the competition over the story. (The gamists, if you’ll accept my fast-and-loose Forgisms.) For those players — and those kinds of games — this approach is probably less than ideal.

Give it a try. So will I. Show me yours and I’ll show you mine.


  1. Carl Klutzke
    February 7, 2008

    Freaking brilliant. You’ve just inspired a very interesting idea for StoryCards that I’m itching to try out now.

  2. Will Hindmarch
    February 7, 2008

    Oh my God, Jeff, you’re totally blowing one of the big secrets about my in-active-development game, Tomorrow War, in which a huge part of the game is the organic revelation of character abilities during play. The gist of it is that you’re amnesiac soldier-from-the-future remembers his abilities during play, Bourne-like, as he encounters situations and objects with which he was previously familiar.

    Did we not talk about that?

  3. Ben Wright
    February 7, 2008

    I think it’s an excellent idea, with one caveat- it requires the players to be tolerably familiar with the system at the outset. Naturally, what ‘tolerably’ means will vary considerably from system to system. If the game is designed specifically with that in mind it will flow very naturally.

    There’s a case to be made for a system that assumes average ability in all fields until specified otherwise- you can be better than average in a field, but the only way to acquire more ‘points’ is to allow yourself to be hampered in some area. Unlike traditional flaw-buy systems, you only get bonus points for actually taking a penalty here and now. It could also promote deliberately putting a character in the pickle to get the points to give them the skill to get out of it, which has a pleaisng symmetry.

  4. Christian Johnson
    February 7, 2008

    I really like this concept, maybe combined with the Three Dragon Ante creation rules in one of the last issues of Dragon, as a way to start a campaign.

    If I remember correctly, the Greyhawk Adventures hardcover by James Ward has a similar concept used in its Level-0 play appendix. The PCs start with more hps than normal, to signify the luck of the “innocent,” but as they make choices and earn a certain amount of xp, their class (with slight modifications) is determined by their actions.

    I’ll have to look it over when I get home.

  5. Jason
    February 8, 2008

    Amusingly enough, I was just looking at a game like this recently. It was called Persona and now I need to remember where I found it.

  6. John Arcadian
    February 8, 2008

    The concept is definitely interesting, and would work incredibly well in a game that it was built into from scratch. It might come across some issues in a more traditional game, where players might make choices about their character abilities based on how the first session went.

    An interesting concept to look at. Buyer’s remorse probably wouldn’t be a problem for a while, but it might set in when the players realize the game style is following a different path than they had at first anticipated. This happens in more traditional character creation models as well, but would be highly contingent in this model because of the character builds in the first session. Allowing a midpoint retcon might be a good idea. It would allow players to tweak out their characters after they’ve gotten used to their skins a bit.

    Of course this idea sends me on a spiraling path into thinking about a system based completely on changing characters as you go along. You might build your character up with points over the first few sessions, but as you level up you would be given the option to change their abilities out. The longer you go in the game the higher levels the character’s abilities might be able to reach, and you could never lower an ability down to nothing. So you might start with diverse characters who are mildly capable in many areas, but end up with specialized characters who share roles.

    I definitely think I’ll be trying the original concept soon.

  7. TS Luikart
    February 8, 2008

    You know, if this was the blog of any other designers, the following statement would have far less relevance: I’ve tried this on several occasions, it never worked for spit in any presently existent system save one, and in that one, it worked like a charm every single time: Feng Shui.

    Something about the final Archetype that gets chosen for each character evolving through the course of play just works nearly without a hitch. The three times I’ve run it following a path similar to what you discuss, Jeff, everybody started as quasi-Everyday Hero’s with the clearly stated “goal” that their “true” archetype would become clear through play over the first few nights. Anything that didn’t quite make sense was ret-conned, e.g. in one game, one of the characters took so much abuse and kept on coming that everyone unanimously agreed that he was a Big Bruiser, even though the player had been playing him as “small but tough as nails”.

  8. Matt Colville
    February 8, 2008

    “So why not just start playing the scenario — before creating characters — and let the players define their characters as the adventure goes along?”

    Allow me to pretend that was not a rhetorical question. 🙂

    While in a movie, you and I in the audience are discovering the characters and therefore the plot as they unfold, the author is not. And in an RPG, I am the author of my character.

    Better, then, to spend the time making my mans, and then have a first act that finds out what he’s made of. Puts him to the test. Forces him to make tough decisions.

    There were rules in one of the AD&D Greyhawk supplements for how to start your character at Level 0. At Level 0 you and all the other characters were basically identical lumps of PC clay and as you adventured, you could try pretty much anything. You could try to turn undead, read magic, lay on hands and as you did these things you’d eventually get good enough at one or more of them to become 1st level in that class. And you’d get to hang on to whatever stuff you’d tried that worked, but wasn’t related to your final class. So you might end up a wizard who could Lay On Hands.

    It was great because it also gave the GM a great way to introduce guilds and churches and stuff. Your character, in a panic, invokes St. Cuthbert and is astonished to learn that the zombiemans run away. Later, a member of the church shows up and says “Hey I heard about you…” and now you get to play the act of joining the church.

    I’ve seen buyers remorse with characters before, the Knight who spends lots of resources on Ride and Mounted Combat and spends the entire campaign in a dungeon. In that specific case, the player asked the GM at startup whether this would be a bad character choice and the GM assured him that if he made a mounted combat mans, there’d be lots of mounted combat.

    The sin lay in not following up and actually putting that content in the game.

    The burden is on the GM, I feel. You make your mans, and I’ll create challenges for them. That’s a lot of burden for the GM to shoulder, but I feel like robust tools are the solution. Or more well-defined roles, so the GM doesn’t have to deal with a myriad of PC options.

  9. Fang Langford
    February 8, 2008

    This makes very good sense. I’ve always felt that at least some in-game alteration / creation should be allowed simply because of “buyer’s remorse”.

    However, I believe a game set up so that the players will be discovering the entirety of their characters as the game begins will take a fair amount of discipline and maturity from the players.

    One of the main benefits of pre-game creation is the intense buy-in right at zero-hour. The comparison to movies and novels is a false one though. The revelation of character in these media is the mechanism of sympathy or ‘buy-in’.

    I think an alternative to in-game ‘discovery’ of character would be to allow the player to dictate some of their early demonstrative scenes. (And I have seen a few games try this, but I won’t mention them.)

    Fang Langford

  10. David Dunham
    February 9, 2008

    HeroQuest (by Robin D. Laws) has such a system, and I’ve used it to great success to introduce an inexperienced player. You don’t in fact need to know anything (as a player).

  11. Paul O`Connor
    February 9, 2008

    This is a brilliant concept, Jeff. I seem to recall that its been done once before, an indy RPG called “Sandman.” Didn’t last long but the players enjoyed the concept a lot.

    I run GURPS and it would support concept fairly well. The players would know that they have a certain number of points, would choose their appearance and any skills they think they’d need to have and then start playing the character, adding advantages, skills, and disadvantages as they’d play. The only two problems I can see are:
    1) The character becomes very specific to the adventure and the players are unlikely to want to reuse the character for further adventures because they’d want to create new characters for specially created for each adventure. While this isn’t necessarily a bad thing, I prefer to have my players keep their characters and modify the campaign to meet their needs.

    2) The characters would probably use up their character points rather quickly, at least in GURPS. There’d be a strong desire to create characters that excel in the first few encounters perhaps to the detriment of later encounters. It would take some play-testing to see what happens.

    So have you tried the concept yet? If not, are you looking for players to try it? If so, I’m game for a session. If you have tried the concept, how did it turn out?

  12. Drew Baker
    February 12, 2008

    Jeff, the closest I’ve heard to character creation in play is a game called Psi Run. Google leads me to believe it was at the Ashcan Front at GenCon, and made by Chris Moore. I’ve found play reports here and here that I think have enough information in them to be interesting, assuming you didn’t already know about it.


  13. Mike Holmes
    February 19, 2008

    Oooh, David beat me to it. Hero Quest has a character generation method labeled (appropriately) “As You Go.” I, too, have used this to great effect with players who have had no experience with the game system in question at all – to say nothing of experienced players who take to it quite readily.

    Basically there are two principals you can use with any RPG that make it effectively “As You Go.”
    – Any character generation decisions may be delayed until it becomes neccessary to make them due to the use of abilities or narration requires description. For instance, the player may decide to delay describing their character until such time as the “camera” of play turns on them (probably the first scene of play, but not neccessarily).
    – Any decisions made and recorded (on the “Character Sheet” or whathaveyou), are considered provisional until they become “cemented” in play by their use or narration that makes them unavoidably permanent. For instance, a player may have written that his character is an elf, but until he is described as an elf, he can change that to any other race that doesn’t conflict with anything else previously established. Once something has been established in play, it must be properly noted, and perhaps marked as a permanent addition (though this is usually unneccessary – people will simply remember).

    A common “fear” of the use of this system is that players will “cherry-pick” abilities based on situation. If combat arises, they will suddenly choose to have a type of character good at combat, for instance. This can actually be seen as an advantage, however, in that characters will tend to automatically converge with the introductory scenario of play.

    The only real potential problem here is that of “niche protection” where somebody intent on playing a particular sort of character may be beaten to the punch by not having had a chance to establish certain things about their character before another player does. The fact is, however, that this is always a problem in character generation – often it becomes the first player to shout out an idea for his character, or two write it down. This “problem” is always solved the same way, which is that any such choices should be discussed with the group as a whole, and players should find an equitable means of dividing up the cool stuff between them (I could suggest methods for this, but it’s outside the scope of this post).

    This does mean that there is a third principal in effect:
    – Narration that “cements” something in play is only considered binding if/when there is no discussion concerning it. That is, every new entry about a character is subject to a short moment where it can be “recalled” and the narration revised, should players agree on a redistribution of the decisions made about the characters. Only once this “moment” has passed, is the entry considered truely permanent.

    In practice I find that this last part is rarely ever invoked, and that players usually come up with enough variations in character types that it’s rarely neccessary to even have a discussion (much less a redistribution of rights on some niche). But the rule should exists so that players refrain from rushing to acquire their character concept, afraid that they won’t get what they want if somebody happens to hit upon it first.

    With these ideas in play, any RPG chargen system can be made into an “As You Go” system.

    Mike Holmes

  14. Jason Morningstar
    February 19, 2008

    Grey Ranks does this. The game’s divided into ten chapters, and the first chapter is played without fully fleshed-out characters. When it’s over, the other players fill out the rest of your character’s traits based on what they observed in play. Heroquest’s “build-as-you-go” option is a wonderful antecedent.

    I’ve played a lot of Psi Run and you answer questions about your character in play that fill in their back story and the game’s overall arc. You start with those questions already written and the entire game is about discovering those answers. Authorship of the answers isn’t fixed – it might be you, it might be another player.

  15. Jeff Tidball
    February 20, 2008

    Matt, this is interesting:

    “While in a movie, you and I in the audience are discovering the characters and therefore the plot as they unfold, the author is not.”

    True in a sense, but the author’s experience is also — very often, in my experience — to learn about his own characters in the process of writing about them. In fact, the more an author tries to plan everything out about his characters before crafting the story of which they’ll partake, the more stilted and weird the whole exercise usually winds up.

    So, if you buy into Will’s idea that the point of a roleplaying game is to take part in storytelling rather than to tell a story, I think that the premise of on-the-fly character creation holds up.

  16. Fang Langford
    February 20, 2008

    Hi Jeff!

    It’s strange to hear you say that about writing. I’ve gotten a few fiction pieces sold and work with a writing group that includes a couple of published authors and no one I’ve heard of learns their characters as they write. They almost always have a very clear idea of who the characters at the drafting stage and occasionally find their characters ‘gaining life’ by the end. This is unlike changing the important details according to the role-playing game examples given.

    Still, I think this is a neat idea as well as a cool game mechanic. I think it demonstrates the incredible flexibility of gaming as a medium and one of the most unique qualities. It’s just not like writing.


    p.s. My friend Roswell says his father (John Sanford / John Camp) may actually do this. Or perhaps he plans everything in his head. After an editor pointed out a problems with one novel, he rewrote the whole thing from chapter 4 on. YMMV

  17. Jeff Tidball
    February 20, 2008

    I think there’s an extent to which pre-writing assumptions fly away because they’re not useful, and an extent to which elements of character which were not planned out evolve as the story is written, because they’re necessary.

    My “fiction” writing has almost entirely taken the form of feature screenplays (well, other than game writing), and maybe there’s something more rigorous about that form (screenplay compared to prose fiction), because there’s so little room for narrative “flab.” (Christ, could I sound more elitist? That’s not the intent. But the fact is that in a screenplay you’ve only got 30,000 words, or less, and you’re forbidden certain types of expression with which you can run amok in prose.)

    Anyway, I frequently find myself on third and fourth page-one revisions where major factors about characters — their careers, their major life experiences — change dramatically to tie them more organically to the action.

  18. Fang Langford
    February 20, 2008

    Wow! That’s a really good point.

    Forgive me, but I’m a bit confused. I don’t know your writing process (screenplays: teh kewl!). When you say ‘revisions’ are you talking about the first time you draft a page or going back to make changes?

    Thanks for clearing this up!


  19. Jeff Tidball
    February 20, 2008

    A “page-one re-write” is typically understood in screenwriting to mean a major iteration of an entire screenplay. That is, you start on page one and re-write the whole thing.

    So, by “third and fourth page-one revisions,” I mean “the third and fourth times I went back to change everything about a particular script.”

  20. Will Hindmarch
    February 20, 2008

    Now I can’t find the quote anywhere to link to, but Aaron Sorkin has written before about how his characters don’t have backstory or abilities until they appear in an episode. He doesn’t do a lot of deep planning — he let’s the need of a particular episode determine whether this character played basketball in high school or whether that one is allergic to cats.

    If a great deal of information about one’s characters can emerge during the course of writing one’s first draft, let us remember that all RPG stories are first drafts.

  21. Kevin Matheny
    February 28, 2008

    In addition to HeroQuest, Amber Diceless puts a portion of character creation into the game, with the auction for stats. It’s not the same an an emergent process, but it is a way to engage the players with one another in creating characters. I suspect it’s not what you were looking for, though.

    I really like this idea. I think the concern about players cherry-picking the best stuff is understandable but overstated, at least as regards min/maxing. I’d be more concerned that players would wind up with ill-fitting sets of stuff that didn’t tie into a central character concept.

    I think this would work well with GURPS. You could either start with completely blank slates or have the players start with an initial concept but no points spent, play through a session taking notes on the things people do and then follow up with the numbers part.

    Very interesting…


Leave a Reply