Gamebook store

Thursday, 4 November 2010

Creating a character who can surprise you

As a lead-in to a run of Empire of the Petal Throne scenarios and features we have coming up, here is an abridged version of an article by Tim Harford that originally appeared in the Oxford University RPG Society newsletter. The subject under discussion is the choice between Design At Start or Develop in Play for player-characters. The prelude game that Tim refers to, based on my Falesa island campaign setting, was achieved by having the players actually run through their own experiences during the 25 years of recent history described in the campaign notes. You can read Tim's whole article here.

The Generation Game
by Tim Harford

After years as a referee, I’m getting involved as a player in two major campaigns. The tortuous births of Baron Alexei Alexandrovich Karenin and, on the other hand, the Energetic and Modest Kotáru hiShathirin have been with me throughout the writing of this article. The experience has highlighted the choice between character generation (largely random) and character design.

Generation - Kotáru hiShathirin

Kotaru was "rolled up" for the world of Tekumel. A central plank of the character generation system is a D&D-like process of simply rolling attributes.

It’s easy to knock holes in the random design system. It’s not "fair". It’s not "logical". In defense of randomization, a few throws of the dice by Lady Fate do a lot to thwart the forces of cliché. Kotáru has Cleverness* 20 and Reasoning 3 - an absolute academic failure with a brilliant scheming mind! We lack literary models for this, and a good thing too. In Kotáru’s case, this is a stimulating challenge for me.

Another advantage of the generation system is that I was able to produce Kotáru from scratch in about ten minutes, knowing nothing about the Tirikelu system.

Design - Alexei Karenin

The Baron Alexei Alexandrovich Karenin was designed for a game of Traveler using GURPS. GURPS has a full-on design system: attributes, training, status, appearance and contacts all have a price, with a budget set by the referee.

Now, it’s all very well putting detailed design rules in, but one of the reasons it had been so long since I’d designed a character in GURPS was that I simply didn’t have the energy - I’d let whoever was refereeing do it for me. Now, with both a campaign and a character I could really get excited about, I’d been looking forward to designing Karenin.

Karenin took me eight hours to design - and this given ten years’ familiarity with the system and a very clear character conception to work from. But that wasn’t the only problem.

In fact, the firm preconception I had was itself a real obstacle. The aim of character design systems is to allow you to design exactly the character you want, within reason. But this only works if the character fits the system. Superficially, GURPS allows anything. In fact, there’s an envelope of efficient designs which, while broad, excludes many perfectly reasonable character concepts.

The GURPS point system pretty much requires that you choose particular complementary abilities. We see a lot of sneaky, acrobatic warrior types, because if you buy high Dexterity to boost your combat skills, you might as well spend a few coppers to become an all round Olympic gymnast. You can get a similarly fearsome bruiser by making him strong and tough, but the fringe benefits of being able to shadow people, escape from handcuffs, cheat at cards and ride horses just don’t come packaged in. Shame.

We also see wizards, engineers or net-runners who also happen to be splendid tacticians, diplomats or doctors. As I say, it’s a broad envelope but there are plenty of perfectly fair designs that just don’t fit.

The GURPS points costs, then, aren’t well conceived. But is there a more fundamental problem? Alexei faced a tremendous expense in acquiring the fencing and motorcycle skills he wanted, more as an affectation as anything else. In a fantasy campaign, being able to fight and ride as well as the Baron would be formidable: in a science fiction campaign, it was mere color, unlikely to be more than chrome. GURPS aspires to price up characters for any background, and the project seems doomed to fail on those terms**.

Certainly, any system which prices abilities based on how difficult it is to acquire them, rather than on how useful they will be, will fail to create "fair" characters – which presumably was part of the aim. One wonders whether we wouldn’t be better off using Paul Mason’s Outlaws Light system*** :
"Write down what your character is like. Then go through and write a bonus number for each area in which they differ from the average. The number can be positive or negative to indicate aptitude or ineptitude, and can range from 1 to 5 (although most areas should be 1 or 2)."
Design and authoriality

Depending on one’s view of the nature of role-playing, character design has a serious flaw, or great benefit. In the Black Corner, we have White Wolf and the Storytellers. Boo! In the Corner of Iridescent Holiness, we have me, Paul, Dave and all right-thinking folk.

The bout is to decide whether the players and referee are combining to tell a story - the Black Corner’s viewpoint - or whether plot is something which only exists in retrospect, having arisen from the interaction of fictional personalities.

The Black Corner holds that the character is a storytelling tool. The player should wield this tool to help the other players and the referee produce a good story. We might call this an authorial approach.

The good guys suggest that the pursuit of a good story is self-defeating. The best way to play, we suggest, is to treat your character as a person and think yourself into the role. You may skew what the referee had in mind, but perhaps she should have been more flexible - with a good referee and a mature bunch of players, the outcome will usually be a great story which surprises everyone.

Anyone who has encountered this old chestnut will have their own opinion. Suffice it to say that a self-conscious process of character design is highly authorial and so it plays into the hands of the Storytellers.

In creating Alexandrovich Karenin, I could have been a storyteller and designed a character with a view to helping the referee. I’m glad I didn’t, because I am sure the result would have been insipid. Instead I turned myself into the referee for a while. I thought about how he should be played, the kind of political interests he would have, his contacts, his allies, his old lovers and his mortal enemies. This has certainly provided the referee with some material to go on, but at the same time gives him a real headache. This kind of conflict is inevitable in a character design system.

Kotáru grows up

Let’s return to Kotáru. I said that Kotáru took about ten minutes to generate, but this was disingenuous of me. Kotáru’s character actually took eight hours, just the same as the Baron Karenin.

As I’ve said, while creating Karenin, I was effectively referee for the day. The result was good. The eight hours spent generating Kotáru hiShathirin was better. Having taken a few minutes to put together the basics about Kotáru, and to think a little bit about what he was like, I joined the other players and the referee in a joint prelude, spanning many events in our lives rather than focusing on a single epiphany. That prelude turned out to be a fantastic session.

As Kotáru grew up we found out about life on Falesa Island, the Hlüss under the pile of stones on the rock, my unexpected enemy, when Kotáru’s betrothed Kala turned the air blue with her curses and whether she was justified, how Kotáru’s strange younger cousin Kishonu hiLanaka is insatiably curious, and what the friendship between Kishonu and Kotáru is really based on.

When the game itself started, we understood how weird the Nom was, and how great the sacrifice of Chondrek hiLanaka, and why we owed the Stranger great honor.

Much as I love Alexei, I wonder which was the better use of eight hours.

A modest proposal

In his White Dwarf article**** “Origin of the PCs”, Pete Tamlyn set the industry a challenge to produce a game with many ways of producing a character, from a full-fledged design system to a method which would allow fairly detailed characters to be generated with a few quick rolls of the dice - either to allow play to commence quickly, or to allow the referee to produce NPCs.

We haven’t seen too many examples of that, but here are a few suggestions. From the anti-authorial viewpoint, a full-fledged design system is not what is required. Instead, we need to think more about the process which produced the person represented on our creased and coffee-stained character sheets. What is her history? Where did he grow up? How did she spend her time when she was young? Who trained him to hold a sword like that? Why is he so nervous of left-handed men? Why won’t she go to confession? And what does all this mean for the numbers on the piece of paper the player holds in his hand?

The player needs a framework to think about these questions. This could be anything from a simple list of points to consider to a programmed prelude along the lines of a gamebook - make choices about your character’s youth, and note down the skill bonuses along the way. An example:
1 When you were very young, did people say about you:
"He’ll be a great warrior, like his father." (2)
"Such a sickly thing! A miracle he survived." (3)
"He’s got such a sparkle in his eyes." (4)
"Always getting into trouble, that child." (5)

Each answer will affect skills and attribute modifiers, and some may spin the character down unforeseen paths. Later on, the questions may be a little less mystical. These would be questions about how you spent your adolescence:
78 Now that you are a squire, what do you do with your spare time?
Flirt with the ladies of court (+1 Bard, +1 Etiquette)
Run errands for a castle craftsman (+2 to a Craft skill)
Spend time around the stables (+1 Riding, +1 Animal handling)
Keep training (+2 Sword, -1 Charisma)
Get up to mischief (+1 Stealth, +1 Pickpocket)

It’s possible to go into a lot more detail, but that is probably not appropriate. It may well be appropriate to elaborate, but that is probably best left to the player’s imagination - or even better, some prelude time with the referee present. The thing that appeals to me about the gamebook format is that it can be made fractal - you can zoom in or out depending on the level of detail you require. To speed the process up, sections of decision-making can be bypassed and given summary statistics:
Train as a squire: +2 Sword, +1 Etiquette, +1 Riding, +1 Strength.
Or, at any point, a one-line summary can be made into a play session lasting anything from a couple of minutes to a whole afternoon. Just how successful was that flirtation with the court ladies? You picked up some skill points, perhaps, but what else: A jealous rival? A reputation? A social disease? A bouncing baby boy? An adoring new friend?

To allow for these fractal possibilities, the generation system needs to be modular. For example, character development could consist of:
  • Potential at birth: attributes, rolled randomly perhaps, or chosen using a point system;
  • Childhood: roll on a table, play it out, or use a programmed scenario;
  • Adolescence: as above;
  • Apprenticeship: as above. Childhood, adolescence and apprenticeship could all be combined in a summary template to save time, if that’s what required - back to the old "character class" system;
  • Careers. Traveler pioneered the idea of describing the training a character had received in terms of one or more careers. Each career, again, can be described by summary statistics or the highlights can be played out.
This process builds a tremendous amount of background detail. It also allows for a significant degree of player choice while avoiding the absurd mathematical trade-offs inherent in a point-based design system.

It has its disadvantages, too. It’s a lot of work for the referee and for the designers - but they love that. It’s also quite incompatible with the "generic" vision of GURPS: character generation is inextricably bound up with the details of the game world. Some people will find that an objection, but for me it’s a tremendous way to introduce a game world and a type of campaign.

I’m struck by how little has changed in the hobby since Pete Tamlyn’s article. I think that a prelude-heavy form of character generation is a step forward, despite the practical objections. Perhaps these ideas are most likely to be carried forward with electronic resources. Until then, the young heroes Alexandrovich Karenin and Kotáru hiShathirin march forward into a bright future.

* The best explanation I have seen is "Kirk has cleverness, Spock has reasoning."
** There are patches for this, of course. Frazer Payne suggested a simple multiplier. The referee spends a few minutes - or hours, more likely - going through the book and applying multipliers to each advantage and skill. There’s a point break: Primary (x 1), Secondary (x 2/3) and Peripheral (x 1/10). Alexei’s strength is secondary, as is his Baronial status. His theological studies and his fencing are peripheral. On the other hand, his ability to pilot a starfighter is fairly central to the game...
*** You can find Outlaws Light in
Imazine #33 
****White Dwarf #72 (Dec 1985).


  1. I like the idea of a generation key listed above as a way to encourage players to think about their characters' histories and abilities. Do such things exist in the public domain?

  2. Not that I'm aware of, Mike, but as Tim points out, it's only necessary for the referee to run the pilot session of the campaign as a growing-up episode and the players will not only be encouraged to ponder those things but will get to experience them first hand.

    It's always going to be more effective to actually live through the events than to put the character together from a menu of options.

  3. I don't know about "public domain", but by combining several systems (Rolemaster & MERP for the background of cultures and apprenticeship, Warhammer fantasy for the careers and a zest of DW6 for some characters like Barbarians or Assassins), it was possible for me to generate non-playing characters with a detailed background.

  4. Mechwarrior, 3rd edition (now known as the Classic Battletech RPG) had a system similar to this, in which your character went through Life Paths: one for youth, next for adolescence, then higher education, then you got your choice of paths for adulthood which you could do up to 4 of before your character was done. Each path would give you certain attributes or bonuses, and then a random roll would give you further attributes or bonuses and possibly allow you to go to different paths or force you down less favourable ones. It wasn't a completely random system because you would still spend points on attributes, skills etc. after going through the life paths. It was a bit clunky because different character attributes were treated differently in the generation process. Traits (like "Glass Jaw" or "Wealthy") acquired through random rolls would apply directly to the character, but skill points ("Rifles" or "Zero-G Operations") you picked up had to be put on a special scale to determine what your 'actual' starting skill level was. Thus if you accumulated 10 points of rifle training your character would only end up with a +3 bonus in the skill. Most oddly, bonus or penalties to Attributes (STR, DEX etc.) in the life path stage would not directly affect your character's stats, they would only reduce or increase how many points it cost to buy a high score in that attribute with the points you were given in the next phase of character generation. Thus your character's random rolls might suggest he/she is a genius but you don't actually have to give them genius-level INT score, it's just that you could if you wanted to. A clever idea in theory but in practice, implemented a bit counter-intuitively.

  5. Instead of a lifepath you might want to take a look at the random character generation system in REIGN. It is built on top of the designed character generation system but it allows the rapid creation of nuanced and interesting characters.

    And as for your complaints against GURPS and other designed systems, Tim is entirely wrong. (Oh, yes you are!) You can have as many areas of expertise as you like... as long as you can afford them. The way that designed character systems prevent (deliberately prevent) omni-competent characters is to give you a budget. With a thousand points a GURPS character can be whatever you want. But you are likely either to be some sort of superhero or so socially powerful as to be... Well probably the Seal Emperor himself! Which would be dull.

    My current ideas about how best to roleplay on Tekumel specifically involve either REIGN or FATE. As is the case with all Tekumel enthusiasts I am taking my own good time in finalising the rules...

    1. I'm not very familiar with FATE, so my comments might be completely off-target, but I don't feel that's the right kind of system for Tekumel. One of the things I like about Tekumel is that it's not a moral universe in which PCs are acting out preordained stories. Events transpire as in the real universe, without sentiment towards living beings. The way Professor Barker put that was "there are no NPCs on Tekumel" and "the dice don't lie". Of course, he also said that players are free to make whatever Tekumel they like, so there's that, but it just feels to me like there ought to be some gaming worlds that are sacrosanct against fate points and all that narrativist wankery.

      (See earlier disclaimer about unfamiliarity with FATE. Ahem.)

      To put Tim's comments about GURPS into context, he was talking about 3e, which made it quite hard to take a great archer who wasn't also a great swordsman and a great sneak-thief, at least in embryo. It's why our group came up with a 7URPS variant (described in earlier posts) but we abandoned that when 4e addressed a lot of those concerns -- and very well too, I should add.

      In the latest editon of Tirikelu I included Steve Foster's rules for non-random character generation, which are like a stripped-down GURPS points system only without the dopy mental disadvantages. Ah, but I mustn't get into that so early in the day. It's bad for my blood pressure :-)