I've heard a lot against random character creation on the Forge and from other sources and I was wondering just how far these feelings ran. To be honest, I love random, and, to a degree I think my gaming troupe does too.

I remembering getting a copy of the "classic" Traveler game, and just loved rolling out my tours of duty, to enlist and muster out. This love of random continue on with Cyberpunk 2020 (and to a lesser extent, Teenagers from Outer Space and Mekton and Mekton Zeta) as well as King Arthur Pendragon.
My group and I have spent whole sessions watching characters unfold through random rolling and trying to turn the rolls into a playable, realized concept.

One of the reasons why I ask is that almost every game I've ever designed started off with an intensive, random character creation method. I find them fun and inspiring.

Any thoughts? Feedback?

Hi Nolan,

Two things. First, you wrote,

I've heard a lot against random character creation on the Forge

I think random vs. nonrandom character creation depends on two factors of the game in question:

1.  Is it important not to suck?

2.  Does the random creation system tend to create characters who suck?

If the answer two both is yes, then random isn't the way to go.  If the answer to one is yes and the other no (it doesn't matter which one), random or nonrandom is a matter of taste, perhaps best left up to the players themselves.  If the answer to both is no, then why not just go random?  You have nothing to lose, and unknown to gain.

Your right about the "a lot" thing. I guess is that what I mean to say there is that I sense a general trend in game design over the last few years to equate certain things, like random character creation (and classes, and levels and plenty of other concepts) with bad or old fashioned game design. It was wrong of me to imply that the forum as an entity or the forum's members as a community held these views over other views.

As for examples of random character creation, I'm hard pressed to state any real examples because the amount of time thats passed since I last used those techniques, but I'll list the last example that I can think of:

A friend of mine was heading out to Germany, which he does every summer, so we sat down to play a one shot of Vampire: the Masquerade (V:tM) I didn't want character creation to take all night, so I thought I'd inject some random inspiration/creation to get things going.

First, I had each player roll randomly to determine their clan (rolling 1d6 with each number corresponding to one of the current Camarilla clans).
Next, each player rolled 5d10 for each of the backgrounds, with each success (roll above 7) determining the characters level in that background.

This had a number of positive effects.
1) It got players away from repeating their usual troupes, in that players who consistently played the same concept/clan had to adapt to playing something different.
2) It allowed backgrounds that were both interesting and complex as well as being more diverse and well rounded than is normally possible with White Wolf's point by system. (Like the fact that most characters had a least one dot in every background, which seems more than reasonable to me.)
3) It got outside of min/max that point buy allows and sometimes seems to demand.

Also, to clarify what I'm looking for (and this may all be to vague, so ask and I'll get more specific):

I would like to hear about other peoples experience and thoughts with randomization in order to look for possible pit falls of the method and to try and look for solutions to problems that people have with preexisting random character creation methods that have come up in order to try and find solutions to such problems.

As to the sucking and not sucking thing, in systems I've created I've come with various answers to these problems, all of which I see as being generally successful though there were specific complaints the players in those games presented. I guess the reason why I would rather deal with randomness in general in the course of role playing games rather than randomness in the specific instance of a single design I've created is that I want to understand (at least one community's) feelings on the subject in order to be able to better implement future designs.

Also, to make it clear, I know theres tastes and that people have differences in taste and I'm not at all interested in creating or participating in evaluating the values of peoples opinions, on the contrary, I'd just like to share in what people have to say about random character creation in order to better understand it/peoples thought on it.

Damn good points, and as moderator, I'd like to say "thanks" for your final comment. That very principle is a big part of why the Forge is what it is.

So, what to consider?

1. What is randomized, as opposed to what isn't. This falls into two categories, as I see it.

a) Straightforward content .... Scores, concept-based classifications, categories which reflect the character's role during play (e.g. class/clan), character history (life-paths), behavior-profile sets. You can see that I'm not distinguishing between quantitative vs. qualitative, or I suppose those can be broken into (i) and (ii) if one wants.

b) Option-sets ... Which set of resources you can work with, which set of skills you can choose, which other player will be antagonistic to your character's interests, and other stuff like that. Damn, this is hard to describe. I'm thinking of The Great Ork Gods, in which there's a random process in which one picks which God(s) one will be playing, subject to the opportunities raised by chance. Which God(s) you run determines adversity for other characters and favor for your characters (the plural is intended; those who've played know what I'm talking about).  I'm also thinking of the games in which one rolls for a given pool of points per "set" of material about the character, as in Fifth Cycle - you roll for how many points you get to distribute among attributes. I know there's a bunch more interesting options too, often in the Fantasy Heartbreaker games. It's no surprise that these games feature many attempts to make randomization fun, as one major goal of such a game is to "fix" D&D. The cool thing is that they sometimes succeed.

Either way, a big part of this, it seems to me, is what is available during character creation that's not randomized, and how that relates to the randomized stuff. Like the skills in the Dying Earth, which get chosen by the player no matter what, but with more points if he or she chose randomization for one or more attributes. Or, in a different way, like the numerical features of Cyberpunk (attributes, skills), which if I remember correctly are only mildly if ever affected by life-path results. You can make up the most bad-ass cyber-ripper street samurai you want; if the life-path says he had a bad relationship that went south over jealousy, well then, he did; deal with it. In fact, both conformity with the chosen features (profession, skills) and contrast with them are equally fun, but how much you get of each is random because the life-paths are random.

2. The challenges and adversity that can arise from the randomization itself.

a) Getting out of the comfort zone, just as you describe. A fine thing for 1990s games, which in my view tended to stay very, very close to their roots in the Hero System.

b) Contradictory features, which is to say, how to play a character who is both hunchbacked and highly charismatic, and so on. In many games I can think of, one might make up such a character non-randomly as well, but the random factor has a way of jogging one's enthusiasm to do it, for several reasons. One of them is the handicap principle, or the creative difficulty one might experience in reconciling the results, that one can turn around from difficulty into inspiration. Of course, a charismatic hunchback! He's a ... (and go on from there).

c) Springboard material, whether back-stories or specific content that goes with some category or skill or something. This is when I roll up a series of life-path results in any of the relevant games, and holy shit, this guy is looking back at me from the page. I totally have to play this guy! That feeling. It can be experienced by anyone at the table about any character, too, which makes it especially cool.

d) Just plain logistics; for example, in the card game Hearts, the person who has the 2 of Clubs always leads the first round of play with that card. There are strategic certain reasons for this (the 2 cannot beat any card; clubs are neutral in all variants of the game), but one virtue is that play always gets going with no delays or fuss. The randomness of who gets the card plays into it relative to what other cards are in your hand; sometimes the 2 is advantageous and sometimes it isn't. I think RPGs might benefit from some features of this kind, to avoid the "so whaddaya you want to do, Marty?" problem that can crop up.

I think it's fair to say that these challenges and adversity are a big part of the appeal. I mean, that ain't any kind of insight. But it does make the next point really mysterious.

3. History: why the pain and suffering? Why, given all that opportunity, do so few modern RPGs utilize it? Hell, I didn't incorporate it into game design until my fourth major game, It Was a Mutual Decision. Looking back, I kick myself that Elfs character creation isn't full of randomization. Hey! Anyone! Invent a cool randomized Elfs character creation system, and I'll give you a prize (uncool entries will not get a prize).

Well, obviously, we have to look to the various permutations of D&D, but especially the early ones in which there was simply no choice except to cheat in order to get a character who was neither wimpy-to-average nor almost-awesome-except-for-the-6. The funny thing is, back in those days (first D&D in 1972, D&D by J. Eric Holmes 1977, AD&D by Gygax sort-of, 1978-80), one's character's characteristic scores didn't really matter all that much anyway. The quantitative penalty for boring rolls, as opposed to truly shitty ones, was missing out on a jump-start on spells. The qualitative penalties seemed like the worse part, though ... my friends and I really wanted cool characters, and the dice wouldn't give them to us. And oddly, as an adolescent, I tended to identify heavily with my characters, so if I rolled up a dumb guy, it somehow meant I was dumb. Go figure. The really painful part was the randomized starting hit points. Shit, you could roll all 18's and still roll 1 hit point ...

Anyway, let's look at those hit points again, because what's going on there, is that extraordinarily severe consequences arose from that single, flat-lined roll. If I were playing a fighter, with my little 1d10 for my hit points, there was no power on this earth that could make a higher value more likely than a lower one. This was supposed to be "fair." In practice, it was simply suckage. In practice, it meant that a DM literally had to cheat the first-level party through all conflicts in order for them to survive. (You might be surprised to learn that I never even imagined starting at a higher level. Yes, I know, people did it all the time. It never occurred to me nor to anyone I played with. Can you imagine how many magic-users bit the dust at our hands? And I didn't even play all that much D&D.)

The same goes for saving throws and for many aspects of D&D resolution of that day and age, but those features were swiftly corrected by other games in the late 1970s. Oddly, randomized character creation was not. You did it essentially the D&D way or you found an alternative, typically point-allocation.

Maybe that history was just too powerful for anyone to overcome for twenty-plus years. I've written extensively about the subcultural power of that particular game, such that agonizing or at least problematic issues one experienced in playing it literally became, themselves, fetishized rather than critiqued. Did all those social things like bragging-rights and teasing-rights based on random characteristics results really hurt that much, such that game design shied away from finding a way to make "roll 3d6 for strength!" fun?

The Fantasy Heartbreaker games offer some insight into this, because they almost universally preserve randomized characteristic generation ... yet, in order for it not to suck, their resolution mechanics do not use the characteristics at all. Instead, the characteristics are used to make secondary scores which are actually used in play, and those secondary scores are so summarized and so average-ized from the characteristics that their range is extremely low. Basically, roll whatever you want for the characteristics; your game-relevant numbers won't vary much from character to character. What I'm saying is that the authors of these games desperately wanted to preserve the practice of rolling the scores, but basically rendered their precise values irrelevant. D&D 3.0/3.5 is similar regarding the 3-18 range, because there are really only about five attribute scores ranging from -2 to +2, most of the time.

Whew! It's getting late, and this post was intended to throw the kitchen sink at the issue, in hopes of prompting memories and getting lots of others' experiences into the thread. I hope it was interesting.

Best, Ron

P.S. If anyone wants to do the Elfs character creation thing (and multiple people can get prizes! it's not a competition), then post in the Adept Press forum, not here.

I think Ron really helped to clarify some of things that can be discussed. I'll leave it for others to add to for awhile after throwing in a couple more experiences from my own play environment that I think really got me hooked on this whole random character creation thing that I didn't mention above.

1) Dragonlance, saga edition (the one with cards): In this game you got a hand of cards, just enough cards so that you could place one in each stat. Each stat also corresponded with a given suit. Your number was your raw ability and how closely your suit matched the given statistic determines your ability with a relevant skill. Though this was highly entrenched in D&D thinking (which was relevant, considering the source material) it was a great mix of randomness and strategic play that almost always created viable characters.

2) There was another random element that I almost always add to my White Wolf games that we call "Coterie Charts" its become so popular with my group that when people are asked what they'd like to play they'll often answer "Coterie Charts" without actually naming which White Wolf game they'd like to do it in. The whole idea of a Coterie Chart comes from Vampire. The idea is that you'd draw a sort of diagram showing how character A felt about character B and so on... How is this random, well, in most of my WW games I run, the first thing I do is have players fill in Coterie Charts of how they relate to other characters without knowing which relationship is going where. Once the whole chart is filled out, its revealed, and the players now have interesting and often somewhat self contradictory feelings and relationships with one another.
These random connections have really brought character interaction to life in game and serve as a sort of launching off point for discussion and development...

Of course, a charismatic hunchback! He's a ... (and go on from there).

Here's something else to consider. The first game I designed was something akin to a Fantasy Heartbreaker, though it didn't go on long enough that it actually broke my heart. Years later, I revisited the setting, trying to determine if I could make something of it and I realized that I had a very large, very complex world that would take a long time for anyone else to learn. I really think random character creation would help to reduce this learning curve, here's how:
Lets say you have 20 or more cultures in a setting, all with rich, detailed backgrounds and you want these cultures to have a real impact on the way a characters behaves, their abilities, etc... If you ask the players to choose they can either make and uninformed choice or devote a lot of time in advance learning about the various cultures. Now, if everyone rolls for their culture, the learning time is reduced significantly, as any player would only be required to really understand their own culture, which might be explained well enough in a few pages of texts. The same thing could be true for a number of other subsystems that are accessible character by character.

For starters, I don't find anything *wrong* with randomized character creation. It can be done well (add to the fun), but I think it's a matter of having the *right aspects* of the character creation process becoming subject to the dice. I'm glad Ron brought up Cyberpunk 2020. This was one of my favorite games when I was in high school, and I spent a lot of time simply rolling out the Lifepaths of characters I wasn't even going to play.

I think the reason I liked it is because Cyberpunk's lifepath didn't affect the Attributes and Skills you had bought (or were going to buy, if you did lifepath rolls first). Okay, there were maybe 2 lifepath results that could affect things (a -1 Attractiveness from a scar, or a +1 combat skill for finding a teacher), but this didn't come up too much. Instead, it rolled you through past romantic relationships, and told you about your family background. In other words, it provided campaign hooks for the GM to mess with. "What? You have an ex-lover who is a scientist? Ex-cell-ent!" Said ex-lover was almost sure to be kidnapped, killed, threatened, or made into an enemy in a future session.

So, hooks = good.

I hated D and D's random attribute thing. Not only did it mean unequal characters at the start, but it meant that a particular character might not fit in the character *role* the player wanted to play.

In other words, completely random character stats = bad.

I think that players like to retain the power of making choices about their characters. They don't mind working Hooks into their characters, especially if said Hooks are interesting. I have yet to meet a player who doesn't want an interesting background. Also, I think it's perfectly fine for lifepath rolls to *affect* character stats, but only to a degree that it wouldn't completely derail a player's character concept. You can still play an assassin if you lose an eye (eyepatches add a certain "badass" mystique, and won't affect perception rolls all *that* much)... but playing an assassin who is also a pacifist is a bit of a stretch.

I went so far as to create a lifepath system for Factions, and put it in the Player's Guide. My goal was to give the player plenty of decisions to make for their character at each stage in their character's life, and to use random tables that reflected these decisions. Example: player decides on being a criminal for their early teenage years, so their random rolls reflect a life on the dangerous end of the social strata. That way, if they end up shot in a drive by, they know that they had accepted the risks. Also, it's pretty hard to actually *die* in such encounters, before game play officially starts.

My playtesters seem to like it. It *does* make for some interesting turns, though - I have to warn newbies that they are giving up a certain amount of control. The downside(s) are that it adds to character creation time to include these levels of detail, *and* the lifepath system itself eats up tons of room in that Player's Guide: 45 pages. I intend to have my website guy write a program that does all the rolling for you, to try to minimize the pain, but that's another story.

Lastly, I've found that I like the *option* of randomized character creation... but sometimes want a point system. This is both as a GM and as a player. If you as a designer want to make it so that taking the risks associated with random creation comes with the possible reward of more points or whatnot, that's a-okay. But sometimes, as a player, I have a very specific character in mind, and want to be able to make him without having to face the wrath of the dice. So consider a build/choice system, as well, even if using it slightly handicaps the player's point totals in some way, for the payoff of getting to pick and choose where said points end up.

At any rate, Choice = Good.

-a-

I have two comments on what Adam just said.

1) I don't feel that random statistics are all that bad. In Pendragon, for instance, when using fully radomized characters, you get a huge amount of variation on character power. Most characters, though, have enough strong points as to be fun to explore with. Since I see the goal of a Pendragon game to be to explore character personality and how they interact with an Arthurian setting, I don't see player capability in any one field to be that significant. Is there a story to be told about the failure of a mercenary surviving on the fringe of Arthur's court, sure, just as much as their is to tell about a knight who could almost rival Lancelot.
So I guess the question of how randomness is used really involves the question of stance. The values I see in randomness in a game like Pendragon really seem to favor the actor stance (stop me if I'm getting this wrong), while point buy or a similar purely decision based process would be more in tune with Director stance. How do other feel about randomness and stance?

2) I've found that in general having multiple systems or a set of optional systems is a bad thing. I know this may sound dumb, but I think the introduction of multiple methods of rolling ability scores, for instance, killed all the fun. When I first began playing D&D, we played "Basic D&D" and rolled stats straight down. I remember that those rolls some how brought a person to life, maybe that person was a fool with a lot of common sense, or a superman, or a total wimp, but he was, for me, for a long time, a person as soon as those numbers hit the page. As soon as we started using the other methods, or expectations on what constituted  a good character got higher and higher, until, as a DM, I couldn't convince players to use the "worse" methods and as a player felt cheated when I was denied the most powerful option. Now, as a player and DM for 3.5 I find that most players feels cheated if they don't have stats far better than those presented in the book. Maybe this is a very local problem, I don't know.
I think different options that generate (or potentially) generate, widely different results are bad (for me?) because they create a descrepency between what player A expects of the game as opposed to what player B expects. In my mind this equates to incoherence (correct me if I'm wrong).
Note, though I haven't player Dying Earth, it doesn't sound like the randomness option presented in their character creation system falls under this category, because its more a more fully integrated part of the game.

Fair enough, fair enough. And I know my opinion is just one of many.

Regarding your Pendragon example, I don't know much about the system. I'd worry, though, if the random rolls could have a bigger impact than the difference between a failed mercenary and a legendary knight. Here, both characters seem to fall into the same *class* (physical combat). Yes, I know, the stories told about both would be vastly different. And the kinds of players I tend to have would be happy playing either. What they tend to balk a bit at is when the choice of *class* is taken away from them. If your random rolls determine attributes, and being a mercenary or knight or other kind of of "fighter" relies on good Strength, Dex, and Wits (or what-have-you), the dice could just as easily produce a character who really isn't cut out to be a fighter of any kind. Usually, players I'm used to come to the table with, in the very least, a class in mind. Also, I'm trying to use the word class here as something other than the way games usually treat the word class, where it determines your feats list, available skills, etc. I mean class as the sort of role the character will likely have in the group.

You are right on the troubles of player expectations. As long as I'm fair as a GM, and make everyone use the same method of character creation, the grumbles are kept to a minimum. However, I worry that your concern for equal power level across the board for your players will be cheated, not preserved, by using random rolls. What's to keep Player A from rolling out a "god" character, while the rest end up with comparative cannon fodder? Nothing at all.

As for my mention of optional random character creation, with a build system as the fall back position, I think I see where the confusion lies. My fault. In either case, the players will end up with very close end totals, when it comes to build points. The randomized parts of the optional system have a lot more to do with Hookish stuffs. They build in character history, but don't determine what your final effectiveness as a sniper will be (for example).

It goes like this:
You roll for family background. This generates information regarding socio-economic status, number of siblings, that sort of thing.
You spend some points into Attributes.
You chose from a list of "Programs" that are available to you. Your options are wider if your family is better off (no elite tutors for the poor). The Programs you pick show you which Talents and Skills you can buy.
You look at the Programs you picked, and then roll for random life experiences based on your environment for that life period (life on the streets, life in academia, etc).

Then you move on to the next phase of your life. This is a simplified version of it, but as you can see, players will still be spending points into their character, and not rolling for hit points or whatever. Sure, if a character gets really lucky (high family standing, does really great on all the random rolls), they will end up with slightly more points to spend, but I doubt it's more than 10%. You're right that a huge margin would equal cause for uproar; it's not just you!

Well, the class thing is a good point in that the way Pendragon works works for Pendragon (poor verbiage, anyone?) because in Pendragon everyone is a knight or would be knight. There are essentially no other types of characters.

I still think the idea of a "backup system" undermines the integrity of a game. I guess we might just be in disagreement here. I just find it to be a form of redundancy that confuses the game feel. I think each RPG should be a unique thing, focusing on one goal and one design principle, the more things a game tries to do at once, the worse it becomes at each of the individual things its trying to do... or at least thats how I see it. Of course this might be short sighted when the reality is that no matter how gelled a given gaming group is, there will be variations on wants and needs from player to player, so optional rules might be the only (or a good or consistent?) way to help maintain group cohesion at an at-the-table level.

I think the life path idea your talking about makes sense. Another reason why I tend towards the whole idea of life paths is that they add a strong simulationist element to the game, in that, in real life and in the back story of most characters, they don't actually decide the details of their past, etc... while at the same time, in real life, decisions definitely shape the way you turn out.
I guess the real question is this:

1) Is it possible to make an abstract model (for the purposes of an RPG) that simulates the push and pull of choice v/preset environment.

2) What would the rewards be? Would they be worth the cost?

I know the second question has a lot to do with design goals, so, the answer will change depending on what each person wants.

Also, when I was talking about the negative consequences of multiple systems of random generation or random v/point buy, I wasn't worried about a problem of power between players, but between methods, which seem like they would be (very?) hard to avoid, and that over time the tendency of a lot of players would be to favor those methods that offered their characters the most power in game as opposed to favoring that method which might produce characters more suited to the more immediate concern of "whats best for the game" (whatever that means).

Some random character generation was good, and some was bad:
Traveller was great - I liked the idea of trying to come up with a reason "what happened that he got cashiered out so early?"
I didn't like the Cyberpunk system - I would spend a whole game session building a lifepath - and then get shot in the head and need to start all over again.
D20 was a question of which house rules we used to cheat up our starting scores :) - I wanted Heroic Fantasy, not Moorcock.

On the other hand, point buy systems , Hero in particular, sometimes took so long that we would spend hours wading through the design system before actually playing - and then sometimes we'd get shot in the head and have to do it all over again.  Sometimes it was so important to design the character right the first time - If I felt that my current character wasn't happy with his chosen career, there was no way to change it without an unrealistic expenditure of XP nuggets -> Try turning a GURPS heavy-armor warrior into a scout sometimes.

To me, random generation is good when it determines the Tools your character uses to succeed, and not the chance of success.  But then, I'm a big believer in games where you don't Suck at the start and have to buy Effectiveness as you play.

The last two games I played in regularly were Reign and Artesia AKN. Both offer randomised character creation and in both cases we used it. There was one specific difference which made Reign work out very, very well and Artesia not quite so.

In Artesia, you roll a life-path from birth to “game starts”. It’s pretty complex and you can get nearly any character imaginable. You still have a lot of stuff to choose yourself (your skills, specifically), but you already get a pretty complex picture. My character, for example, ended up being a nobleman with a guilt binding and some experience as a warrior, all complete with Spring Queen heritage, from a small woodland castle in the Erid Wold. I fleshed out his skills pretty much by “what this guy would know”. I added some flavor in making him a good-hearted and spiritual fellow for a change, but there wasn’t all that much of me in that character. And what’s more, he did not fit very well with the other PCs, so they kind of, y’know, stuck together but no one really knew why. The game was still alright, but I didn’t engage in it all that much.

Reign, on the other hand, not only allows but demands for you to make something of the results. You can interpret them in totally different ways. And if you cooperate with your fellow players, you can create loads and loads of connections between the characters.

In Artesia, the randomised character creation is a lot of bookkeeping, and sometimes you go, “hey, that’s neat”, but after that in play, it doesn’t pay off that much. In Reign, you get some pieces of your character and your task is to fit them together and make them make sense. That’s challenging and fun, and it pays off in actual play. If you and I roll the same results for an Artesia character, we’ll be playing the same guy, more or less. It you and I roll the same results for a Reign character, we’ll most certainly end up with radically different guys.

P.S.:

"I'm bored, so I'll make a character for fun" (optional addition - Wife: "What are you doing, honey?", Me: "Nothing!")

As it happens, last session I played Reign, I have been faced with the disadvantages of the one-roll chargen: My character simply sucks at everything, sad fact. I like him, but his stats are crap. He’s got a little bit of everything and too much of stuff you rarely need. And even with that stuff, like “plead”, he’s at 6 dice max which still doesn’t make for a very good chance of even rolling a set at all, let alone a better set than the other guy. And he really sucks at anything that has to do with fighting, which is not going to improve fast. But Reign’s a lot about fighting.