Linear CharGen/Geometric Advance Needn't Be Inherently Bad

[Ancient History]

I didn't want to derail any existing threads with this, but I've heard this often enough...and it's worth going into a little bit.

What I'm Talking About
In a game like GURPS, character generation and character advancement are essentially the same seamless system: you spend points according to a set schedule and get the same increases, regardless whether or not the points are spent before or after character generation. This is linear generation coupled with linear advancement[/i], where the costs remain the same throughout. It is instinctive and easy to follow system; it is also generally fairly long and tedious as the point costs creep up there, as it can happen in very detailed systems.

In a game like Vampire: the Masquerade, character generation and character advancement use different systems. For chargen, you are given set amounts of points which are to be allocated to specific parts of the character sheet in discrete steps; and after chargen, further advancement happens by buying up additional points using XP, usually at a geometric rate (i.e New Rating times X). This design method simplifies chargen considerably (fewer points, fewer choices) and leads to relatively slower character growth (because the higher your numbers, the more points you need to raise your numbers).

Both systems have their strategic character-building advantages. Linear chargen/advancement offers great detail and improvement is relatively easy, quick, and can be fast if the gamemaster is generous with the points. Linear chargen/geometric advancement tends to lead to front-loaded character building, as it is generally cheaper to raise your numbers at chargen where prices are fixed than it is to raise them later on during the game, where the cost of advancement increases.

More rarely, you might see a geometric generation/geometric advance. This generally happens when you want to keep the advantage of slower PC advancement after chargen and a large amount of detail at chargen. So systems like Shadowrun4e's KarmaGen give you a shitload of Karma and then tell you to use the regular Karma advancement rules to build your character.

I've never seen a geometric generation/linear advancement system, and it's hard to think of a situation where that would be useful - unless advancement was extremely limited in certain ways, in which case it's more important to how you build your character than improve them. I could see this working for some kind of Call of Cthulhu heartbreaker, given how skills are so shit and hard to advance...but I've never actually seen such a system implemented.

Why Is All That Bad?
It doesn't have to be, but it often is. Having linear costs at chargen and geometric costs for advancement, for example, can lead to a situation where two players build characters using the same method, but one character comes out with a character which is objectively worth more points post-chargen - usually by deliberately choosing options that have a fixed cost at chargen but an exponential cost at character advancement.

For most White Wolf and Shadowrun games, this generally boils down to Attributes, which are used in nearly every dice pool roll and are expensive to raise during character generation. It makes sense when building a character to try and maximize your dice pool, which generally means mathhammering out the ideal combination of high attribute and high skill(s) for whatever you want your character to do. Especially in games where opposition is measured by how much better their numbers are than your numbers, having the bigger numbers is important - and if the situation is lopsided enough, then some character options become bad because you're spending points on something that doesn't actively increase your numbers puts your character at a disadvantage.

That segues into an entirely separate conversation on trap options and poorly-designed skill systems that don't do what you actually want them to do. World of Darkness skills, for example, are notoriously stupid. What does it mean to have Thanatology 5? What does that do for you? What target numbers are you rolling against/how many hits do you need? Most of the skills in the system are worse than useless, but some of them - Melee, Firearms, Dodge, Occult - get much more attention because of their place in important combat/magic subsystems. Those are the skills you care about...and most of the time, you still try to avoid rolling them because the systems are terrible and you're likely to fail.

You can't force every character to be equally useful. Not with linear chargen, not with geometric chargen. Any time you force a player to spend points, that's an opportunity cost, and not all opportunity costs are equal or predictable. That doesn't mean you shouldn't try to identify the obviously crap options and eliminate them, or to address imbalances to make the range of creatable characters something you can actually play the game with. Which brings us back to...

Linear CharGen/Geometric Advancement
Which is basically what I went with in Space Madness! - inspired more by Shadowrun than World of Darkness, and for about the same reasons. Adding points together at chargen is a lot quicker and simpler than just handing the players a big pile of points and some limits on how to spend them; it also helps control the dynamic range of the characters that can be created. In Space Madness, you pick two backgrounds and one order - the possible base range of combinations from that is (10 x 10 x 7 = 700)...not counting character specialization, Certs, or Equipment choices. Now, you might not want to play a Belter Terran Ordo Penares with a robotic forearm in place of your genitalia, but that is an option...and from a mechanical standpoint, while probably not the most optimized character, it will still be a viable character.

700 base characters isn't bad. It isn't tremendous, but it's more than sufficient for most purposes. You look at D&D3, just the basic book, there are 11 character classes and 6 races - that's 66 basic level 1 characters before you start to factor in alloting attributes and skills, selecting feats, and actually making the "build" decisions of your character.

But of course, you don't tend to see an even distribution of characters in D&D. There are a lot more Elf wizards than there are Half-Orc wizards. This is because Elves have wizard as their favored class and Half-Orcs have an Int penalty. The Half-Orc character has to spend more resources to get at where the Elf starts out...and the Elf still has an advantage on top of that.

If you had enough Space Madness! characters stat'd out, you'd see something similar develop - there are certain combinations of background and order that lead to min/maxing of certain dice pools. More or less. I tried to balance the numbers a little to avoid too many obvious combinations specifically because of that - it's also why each cosmic force doesn't use the same skill for every power. Min-maxing in one skill (like Combatives or Visualization) will not make a character overwhelmingly powerful - but it still provides sufficient incentive for characters that want to do that to specialize their character in that fashion.

And if more supplements were ever made to Space Madness!, whatever balance I tried to strike might be thrown out the window...but I think that by itself, there can be certain advantages to linear chargen/geometric progression. It doesn't automatically have to be the worst of both worlds.

[TiaC]

I could see geometric generation/linear advancement working for a game that has starting characters as non-adventurers/heroes who become heroes over the game. It heavily encourages starting characters to be spread out and not particularly good at anything but lets them specialize easily as the game goes on.

[OgreBattle]

It seems like a soft way to 'deal' with this is to have gameplay that is...

- Rather abstract in particulars
- Fast to resolve

So any weaknesses or strengths in one particular character are handled quickly and then other parts can be emphasized.

Like playing a hobo with bags of candy and social skills while your buddy is a dragon can work... but not really work if the game is very combat mechanics heavy and everything else lite

[EightWave]

Really what you're arguing for is a sane version of the Feng Shui archetypes or Shadow Run priorities system, not linear chargen. Something like distribute 4,3,3 to stats, 4,4,3,3,3,2,2,2,2 to skills, get traits worth 3,2,1, and flaws worth 2,1, a 3,2 background, distribute 24 XP how you like (with a quick lookup table for how much it costs to upgrade) is materially superior to any linear/geometric combination.

it also helps control the dynamic range of the characters that can be created

How? Linear purchasing schedules have larger dynamic ranges than geometric ones by definition (100 points can linear buy a stat up to 100, but geometric buy it only to 10). This is what makes me think you're really arguing for build packages.

I could see geometric generation/linear advancement working for a game that has starting characters as non-adventurers/heroes who become heroes over the game. It heavily encourages starting characters to be spread out and not particularly good at anything but lets them specialize easily as the game goes on.

This doesn't encourage generalizing, it just punishes specializing in the same way linear/geometric punishes generalists. Better to just have linear advancement all the time but nobody can buy better than a 3 in anything at character creation.

Literally the only thing mixing linear and geometric point buy does better than sticking to one purchasing schedule with clear build packages and guidelines is give yet another avenue for someone to fuck up their character.

[Username17]

Literally the only thing mixing linear and geometric point buy does better than sticking to one purchasing schedule with clear build packages and guidelines is give yet another avenue for someone to fuck up their character.

This.

There is absolutely zero excuse for Linear Generation / Geometeric Advancement because it only does one thing and that thing is terrible.

There is no reason to want the One Trick Pony character who advances to have a Second Trick to be flatly superior to the Two Trick Pony character who advances Both Tricks. And that's the only thing that Linear Generation / Geometric Advancement actually does. It mathematically provably does that so consistently and so obviously that I don't have to see the actual system it's attached to in order to make that assessment. And my assessment is that that is the thing it does, and my judgement is that the thing it does is fucking terrible.

Bottom line: You Are Wrong. This is a bad idea, because it's obviously a bad idea. It was a bad idea in the 80s, and it's a bad idea now. And it's a fucking basic mathematical fact that it's always going to be a bad idea and it's never going to stop being a bad idea.

-Username17

[Ancient History]

Really what you're arguing for is a sane version of the Feng Shui archetypes or Shadow Run priorities system, not linear chargen. Something like distribute 4,3,3 to stats, 4,4,3,3,3,2,2,2,2 to skills, get traits worth 3,2,1, and flaws worth 2,1, a 3,2 background, distribute 24 XP how you like (with a quick lookup table for how much it costs to upgrade) is materially superior to any linear/geometric combination.

Same difference. You're still doing additive character building at generation and then exponential later on; the only difference is that instead of having fixed points to spend on select categories (i.e. White Wolf-style) you add another layer of decision making (Shadowrun-style). The general purpose of both approaches is the same: to guide chargen and set caps.

it also helps control the dynamic range of the characters that can be created

How? Linear purchasing schedules have larger dynamic ranges than geometric ones by definition (100 points can linear buy a stat up to 100, but geometric buy it only to 10). This is what makes me think you're really arguing for build packages.

That's assuming that you're using the same points for both creation and advancement, which usually is not the case, except in linear/linear systems like GURPS.

Literally the only thing mixing linear and geometric point buy does better than sticking to one purchasing schedule with clear build packages and guidelines is give yet another avenue for someone to fuck up their character.

Mathematically, you're correct - and yet that's true of linear/linear systems as well, it's nothing unique to linear/geometric systems. GURPS is a hallmark of a game where you can spend a shit ton of points on things that don't matter, and you end up with characters of wildly different effectiveness and abilities despite them having the same nominal points value.

Any system complex enough to have interesting choices is going to have ways to fuck up your character - the ideal is to have a system where the "fucked up" character is still viable enough to contribute to the game and be fun to play.

[Blade]

Any system complex enough to have interesting choices is going to have ways to fuck up your character - the ideal is to have a system where the "fucked up" character is still viable enough to contribute to the game and be fun to play.

The problem is when you can create a fucked up character without knowing it, or when you can have a character who's mathematically worse than another one (exact same abilities but lower stats because of a difference in how it was built).
Guidelines can help limit that, having publicly known optimal builds can help as well.

You might also want to limit the number of ways you can have this or that ability/boost. When building a character is similar to building a MTG deck, this is a bad sign.

But further than that, in a free-form game you can't even get any correct assessment of a character's value. Depending on the adventure, the team, the way the players approach situations, the same character can be overpowered or completely useless.

[EightWave]

Same difference. You're still doing additive character building at generation and then exponential later on; the only difference is that instead of having fixed points to spend on select categories (i.e. White Wolf-style) you add another layer of decision making (Shadowrun-style). The general purpose of both approaches is the same: to guide chargen and set caps.

No. Emphatically no. There's a massive difference between "pick from these two packages, both are worth 30 points now and forever" and "both are worth 30 points now, but one is worth 25 and the other 35 the moment you start advancing as a character". The difference is that one is a good idea, and the other is not.

That's assuming that you're using the same points for both creation and advancement, which usually is not the case, except in linear/linear systems like GURPS.

No, that example was at character creation. I'm saying that in a theoretical linear system where you distribute 100 points among 10 stats one character can have a 10 in everything while another has a 100 in one and 0 in the other 9. That is definitively a larger dynamic range than a geometric system where you can have a 5 in everything or a 10 and nine zeros.

Mathematically, you're correct - and yet that's true of linear/linear systems as well, it's nothing unique to linear/geometric systems. GURPS is a hallmark of a game where you can spend a shit ton of points on things that don't matter, and you end up with characters of wildly different effectiveness and abilities despite them having the same nominal points value.

"Well, sure, tigers are wild animals that can overpower and eat you at the first mistake, but since both they and cats shed there's no real difference between them as house pets". There being bad places to spend your build points is immaterial to the problem of there being bad ways to spend the points. Having a linear/geometric advancement schedule does nothing to curtail the problem of there being trap builds, it just adds one more trap to the pile.

[Omegonthesane]

Mathematically I gotta echo Frank's statements here, there is no thing that is improved by having unlimited-time character generation use linear point buy and later XP growth be geometric.

The only possible advantage to linear chargen/geometric advance is if you really want geometric advance to be the order of the day but also want 5-minute chargen.

Even then you're probably better off just not giving a great many free floating points to people who don't opt in to precision character generation. I'm imagining like, you are meant to be worth a minimum of 5N experience at character gen, so you might have like 10 archetypes that each have just over 4N worth of fixed purchases and can spend N points on top of that to differentiate between one another, and then a "Specialist" who just gets the full 5N because someone's always going to want to play a unicorn.

(The idea there is that you are explicitly rewarded for taking an archetype instead of being a snowflake by being about half a session of advancement ahead of everyone else, because you just ate the opportunity cost of having the lion's share of your character points spent in a very particular way, and also aren't granted much opportunity to completely fuck up if you go that route. Assuming of course that by some miracle all the base statlines are worth a damn.)

[hyzmarca]

Linear chargen and geometric progession does one thing well, and that's encourage hyperspecilization at chargen followed by broadening during the campaign. It works quite well if the characters you're creating are supposed to be the best there is at what they do, and not all that great at anything else, and they they learn other skills as they progress.

It's actually a pretty common trope, where the hyper-focused killer lacks important life skills and has to learn these things to become a functional human being and it's worth emulating. To a lesser degree, many modern special forces do it like this, with each member training in a specialization and then crosstraining with the squad so that everyone becomes passable at every skill.

Geometric chargen will always encourage broad jack of all trades builds because there comes a point there the value of an extra die just isn't worth the cost.

[souran]

The only advantage of linear chargen/exponetial advancement is if you want to guide advancement to certain facets of character design.

So, for say white wolf RPGs, if they had not beend designed badly would have done something like:

A) Attributes are linear at chargen and exponential to advance

B) Abilities (Skills) are linear through 2 or 3 dots at chargen and exponential beyond that point at chargen. All advancement of of Abilities are linear to advance after starting play.

Similar paradigms could be created for backgrounds and powers so that players are encouraged to define a characters major arctypes at chargen and then advance those instead of building hyper specialists at chargen and then adding all the support stuff.

However the evidence is really in that if you do linear -> exponential then you will make everybody build very narrow concepts at chargen and then "flesh out" their characters. It actually mostly sucks.

[Orion]

Linear chargen and geometric progession does one thing well, and that's encourage hyperspecilization at chargen followed by broadening during the campaign.

But instead of incentivizing hyperspecialization at character generation with the promise of future advantages, you could just make it mandatory. There's no reason to say "A starting character can max out 5 skills, or they can start with 10 skills by setting their future on fire" when you could just say "a starting character can only have 5 skills."

[Blade]

There's no reason to say "A starting character can max out 5 skills, or they can start with 10 skills by setting their future on fire" when you could just say "a starting character can only have 5 skills."

I don't know, you always have players who want to make a quirky characters. And sometimes, some of these characters can be interesting.

So I think that saying "We recommend to start with 5 maxed-out skills, doing otherwise is sub-optimal in terms of character progression" is a good middle-ground between saying "here are some points, spend them how you want" and saying "you start with 5 maxed-out skills".

But of course, it all depends on your game. If you want your game to be about specialists, or to be very simple to pick up and play the 5 maxed-out approach is better. If you want a more nuanced game with more options and accept more complexity in character creation, giving the choice while warning about the consequences isn't absurd.

[EightWave]

It's not that restricting player choice is good or bad, but that encouraging a kind of build by setting up mathematical traps for other builds and hoping for people to figure that out is bad.

[Foxwarrior]

Isn't the option to build in a different way just a power now for weakness later choice? At game start, the "quirky, interesting" character might indeed be worthwhile, but after some amount of experience has been gained they will instead be garbage. I guess if you want a game where some, but not all, of the players are encouraged to throw away their characters after a while, that helps.

[EightWave]

Okay, there are two things that really need to be cleared up because they just keep being repeated.

1) linear character generation is not automatically faster than geometric generation. In the very specific case of having to do all of the math in your head it is easier, but once you have more than a dozen places to distribute build points the most time consuming part of character generation is the bookkeeping, not the the math. And even easier than doing math are build packages and lookup tables, which are advancement scheme agnostic.

2) Mixing advancement schemes is a reduction in player choice because it creates mathematically superior build orders. If starting as a kung fu exorcist and then advancing along both kung fu and exorcism paths is mathematically inferior to starting as a master exorcist and then learning kung fu that is an option that has been removed, not a choice that is added.

If your argument falls under either of these two concepts you are wrong and, at best, are arguing a stupid edge case that can be dealt with in a one paragraph sidebar.

[kzt]

It's not that restricting player choice is good or bad, but that encouraging a kind of build by setting up mathematical traps for other builds and hoping for people to figure that out is bad.

I really doubt they usually know it's a trap. The mathematical cluelessness of the average RPG writer is pretty high. Plus they seem to often actually play their game differently than what the rules they wrote says it should be played.

[Orion]

Another huge benefit of sticking with one advancement system is that it makes it easy to audit character sheets or reconstruct missing ones. A beautiful thing about D&D 3.5 is that if you houserule intelligence bonuses to grant retroactive skillpoints, I can look at at a character sheet for a high-level character and figure out whether it has the correct number of skills and feats without needing any life history. In games where people receive xp in small increments and spend it a la carte, it inevitable happens that someone either forgets to record XP they received or forgets to deduct XP they spent, and in those situations it's really useful to be able to just compute the XP value of their character sheet without needing historical data.

If you want to force generalization there are so many ways to just outright force it. Limit how many skills can be capped. Make people level up increase 2 skills at once whenever they increase skills, one skill that's 1-3 and one that's 4+. Outright give out free skill points periodically that can only be used on skills <3. Make people buy skills above 5 with a different currency that's distinct from normal XP.

[OgreBattle]

What outputs do you get from linear vs geometric advancement,

What genre conventions fit one or the other better

[Username17]

What outputs do you get from linear vs geometric advancement,

What genre conventions fit one or the other better

100% of what it does is change the relative value of characters who start one-dimensional and expand into a second shtick versus characters who start with two shticks and gradually get better at them. It makes the second character have progressed less than the first one. That's literally all it does.

Since most genre stories don't really have character advancement in the way that RPGs traditionally does, it's very difficult to even find an example where it would be relevant. The closest I can even think of is Deep Space Nine.

Miles O'Brien started as a random yellowshirt transporter officer on the Enterprise. In Deep Space Nine they expanded his backstory to include him being a badass veteran infantryman in addition to an engineering officer. Essentially that's "character advancement" where a character started with one relevant skill set (engineer) and had a second skill set added on laterally (soldier). Worf started as a soldier and a ship's officer and thus had essentially both ability sets. He was a fan favorite and by the time he was brought into DS9 he had been built up quite significantly and he was able to take on sqads of Dominion super soldiers pretty much by himself. That's a character who started with two tracks of relevant ability and had both of them advanced later in the series.

In Linear Generation / Geometric Advancement there will come a time that if the game goes on long enough you will wish that you had been O'Brien instead of Worf because after a significant amount of advancement is under both belts the O'Brien character has strictly more character points than the Worf character.

That's literally the only thing it does. It's completely indefensible.

-Username17

[Foxwarrior]

I thought OgreBattle was actually trying to derail the conversation, as in "What outputs do you get from linear (advancement, and character generation) vs geometric advancement (and character generation)". In which case the answer has more to do with how specialized you want characters to be, and how much time the players should spend on doing arithmetic. Usually min-maxing leads to more useful characters (well, until you hit fundamental caps like "no matter how hard it is to hit the guy, I will have a 100% chance of success"), so a simple linear cost system will make people min-max out the wazoo. Even geometric actually can do that sometimes, with just a bit of side-dipping; in my experience you have to go all the way out to exponential for focusing on your core stats to be only as good as diversifying.

If you're a bit trickier with your linear cost system though, it's pretty easy to get people to diversify a bit.