PDA

View Full Version : [Game Theory] Caps in a point buy system?



Saph
2013-01-07, 11:03 AM
So, I'm about two-thirds of the way through designing a point-based urban fantasy system, and there's something I've had trouble deciding on for a while. I figured some of the more experienced GitPers out there might be able to help me!

The question is whether, in a point buy system, you should have caps/limits on how good at one thing a character is. A good example is Mutants and Masterminds. A M&M character has a 'power level' determined by how many points they were built with, just like character level in D&D, and they can't have a total bonus to any skill greater than their power level +10. So if they're Power Level 10, their maximum modifier to any skill is +20.

An example of a system that doesn't use caps is Shadowrun. In Shadowrun, your skill is measured in d6s – the better you are with a skill, the more six-sided dice you roll. It's possible to gain a very large amount of dice on a single skill by stacking attribute dice with skill dice and then getting bonus dice from magic, cyberware, bioware, special qualities, etc.

Arguments against skill caps

Having no skill caps gives you maximum freedom in designing your character, which is part of the appeal of a point buy system.
Skill caps are typically quite metagamey – there's no good in-universe reason why a M&M character shouldn't have really exceptional ability in one skill.
There's an argument that putting a hard limit on numerical bonuses just encourages players to find ways to boost their character without using numerical bonuses.
Arguments in favour of skill caps

Skill caps prevent massively overspecialised characters (the gunslinger who can shoot the legs off a fly at a hundred metres while blindfolded but who can do absolutely nothing outside of combat). Shadowrun is famous for these sort of characters (the pornomancer is the one that usually gets brought up).
Having no skill caps can create an 'arms race' situation where characters have to spend more and more of their resources to be able to compete.
Skill caps encourage characters who aren't complete one-trick ponies. On the other hand, there's an argument that if a player wants a one-trick pony, then that's their decision.
The rule I'm considering is something along the lines of M&M's approach – your total bonus in any one skill is limited by your overall number of character points, effectively meaning that you can't spend more than X percent of your points on one skill.

Any thoughts appreciated!

Friv
2013-01-07, 11:24 AM
As a general rule, I'm in favor of caps. They help give me an idea, as a player, of what constitutes "total focus", "partial focus", or "minimal focus" in a given concept, and help to keep people on a roughly even keel.

Stubbazubba
2013-01-07, 12:23 PM
The biggest reason that I am pro-cap is because of intra-party disparity. Sure, the Rogue has a +27 on his Stealth, so you think having the party sneak past the demon with the +18 Spot bonus would be OK, but the Fighter is too decked out and taking major penalties - he can't succeed even on a 20. So, you reverse gears and choose something the Fighter can sneak past; well now the Rogue's Stealth is excessive - he can't fail on a 1 - and would have been better put in something else. This DM can't win either way.

When the assumption of your game is that parties face challenges together, it only makes sense to mechanically anchor a party of a certain level around a certain balance point X, allowing each individual character to pump resources into being closer to or further from X, but all within a given range that keeps challenges meaningful.

Because of this, in addition to maximum caps, I sometimes have minimums, as well; across-the-board bonuses, not unlike 4e's half-level bonus, in order to keep everyone within the right range while allowing the specialists to still get better.

The primary consideration here is how big is your RNG's spread, how much of that is reasonable, what should the odds of success be for a character of level X who is [very bad], [bad], [baseline], [good], or [very good] at that proficiency. Caps, both upper and lower, should be based on that. And then the DM can choose a level-appropriate challenge without screwing someone over by default.

That, to me at least, is the purpose of caps - to keep the party on a relatively even keel.

Saph
2013-01-07, 12:28 PM
Stubba: I can see how that makes sense, but the examples you're using are all level-based systems, like D&D, rather than point-buy systems. Most point-buy systems don't have a 'level' or 'challenge rating' system.

CarpeGuitarrem
2013-01-07, 12:38 PM
I'm not sure what's metagamey about caps. Everything has a limit, a "carrying capacity", something which is not transcended without some sort of other aid. The human body has limits. Certain types of gear help to take them past those limits...to a point. Additional enhancements can continue to take them past those limits...to a point.

Ultimately, though, you can never go past a certain point. (On the other hand, perhaps the specialized characters in Shadowrun really could be that plausibly good? Either that, or the bonuses are too big.)

I would cap base attributes, and also perhaps adopt some form of 4E's feat-stacking paradigm: bonuses from certain sources don't stack, they overlap. (For instance, you could have a bit of gear and a magical enchantment that affect various things, and one of those things that they both affect is your speed. However, you only count the higher bonus of the two, because they're enhancements.)

Stubbazubba
2013-01-07, 12:41 PM
D&D 3.5's Skill System is not level-based, it's a point-buy system with level-based caps.

So, if your M&M campaign is set to PL 20, and one character has a +27 on his Stealth, but another only has +5, then they're the same as the Rogue and the Fighter from before; there is no encounter that is possible for both of them to simultaneously fail or succeed. For simplicity's sake, make it a Climb check; there is no DC that both the Fighter and Rogue here have a chance to both fail or succeed.

If your game has an inherent power increase as play continues, whether its vertical (level-based) or horizontal (point-buy), you need to consider what kinds of challenge to throw at the party, and the game should give you a good indication of a challenge's suitability for what parties. But it's true that not many do.

Siosilvar
2013-01-07, 12:43 PM
Stubba: I can see how that makes sense, but the examples you're using are all level-based systems, like D&D, rather than point-buy systems. Most point-buy systems don't have a 'level' or 'challenge rating' system.

...that is, except for the fact that the points themselves supposedly serve as a decent measure of personal power. Theoretically two 150-point characters have the same power, just in different areas. It doesn't always work out that way (multiplicative effectiveness, specialization vs. generalist, diminishing returns, whatever), but the points serve just as well as more granular levels.
Even if the points aren't an effective measure of power (and why bother with them if they're not?), the basic premise that it's possible for the difference between two people to exceed the randomizer still holds.


There are a few options between uncapped and capped advancement.
Savage Worlds uses a soft cap, where increasing a skill past your attribute costs twice as much.
...something that I can't remember (GURPS, maybe?) uses diminishing returns; every point in a skill costs points equal to the new skill level.
Both options slow advancement beyond a certain point and encourage generalization without explicitly disallowing overspecialization.

Grod_The_Giant
2013-01-07, 12:48 PM
Stubba: I can see how that makes sense, but the examples you're using are all level-based systems, like D&D, rather than point-buy systems. Most point-buy systems don't have a 'level' or 'challenge rating' system.

M&M is point buy; the PL isn't really the same as character level, as all it does is limit certain numbers.

Personally, I like the caps in M&M. It's awkward and metagamey, yes, but, as Friv pointed out, they're good for giving guidelines on how good a given score is. They cut down on a lot of numerical abuses possible in an uncapped system. And, at least in M&M, they promote horizontal growth over vertical-- most of the time, characters gain new abilities instead of upgrading existing ones, improving versatility instead of raw power, again helping to keep things under control.

Saph
2013-01-07, 01:02 PM
So, if your M&M campaign is set to PL 20, and one character has a +27 on his Stealth, but another only has +5, then they're the same as the Rogue and the Fighter from before; there is no encounter that is possible for both of them to simultaneously fail or succeed. For simplicity's sake, make it a Climb check; there is no DC that both the Fighter and Rogue here have a chance to both fail or succeed.

True, but I can hardly think of any system that doesn't have the "one guy's worst beats another guy's best" thing. Low-level D&D, maybe, but that's only because the d20 has so much variance. FATE, M&M, Shadowrun, L5R, White Wolf all make it near-impossible for a specialist to lose to an incompetent.


...that is, except for the fact that the points themselves supposedly serve as a decent measure of personal power. Theoretically two 150-point characters have the same power, just in different areas. It doesn't always work out that way (multiplicative effectiveness, specialization vs. generalist, diminishing returns, whatever), but the points serve just as well as more granular levels.

I'm not sure if they even work that way theoretically. For systems like GURPS, experienced players will generally tell you right at the start that points do not equate to personal power – the whole power-level way of looking at it just doesn't really work, because specialisations vary so much.

I do like the soft-cap idea: it discourages overspecialising without completely ruling it out. Not sure how to price it, though. At the moment I'm using arithmetic-progression costs (going from rank 2 to rank 3 costs 3 points, going from rank 3 to rank 4 costs 4 points, etc).

valadil
2013-01-07, 01:07 PM
I prefer caps because they can be raised to simulate levels.

Here's my issue with capless skill systems. If you play a specialist in a capless system, you start the game with a skill close to where it will peak. Systems like this usually have diminishing returns when it comes to buying more skills. Even if the cost doesn't increase, you reach a point where extra points aren't beneficial. If you're at the point where rolling a critical failure is the only way to fail, what's the incentive to buy more points?

Anyway, what I've seen happen is that the specialists get more generalized. If you're at that point where your specialty isn't worth investing in, you end up buying other skills to supplement your character. In all the GURPs games I've played, the PCs who started out specialists end up generalists because they've got nothing to do but shore up accessory skills.

Why is this a bad thing? Because it takes your character's niche away. If I want to play the stealth ninja of the group I can easily start out as the stealthiest, but once everyone else invests in the skill I may have the highest score but I'm no longer really a specialist. The disparity in skill level between my character and the next most competent will decrease because the other character can catch up more quickly. IMO watching your specialty become less special isn't fun.

Caps are the solution to that in my book. Raise the caps over the course of the game and your specialists will have room to grow.

GolemsVoice
2013-01-07, 01:19 PM
@Stubbazubba:

But what would caps accomplish in that scenario? A rogue-y character will always try to get the maximum in skills like sneak, and a warrior type, possible in heavy armor and other gear with penalties to sneak will likely disregard such skills.

So it doesn't really matter if the warrior has a bonus of +5 and the rogue has a bonus of +18, because 18 is his cap, the warrior will still not be very good at sneaking.

The real reason why the warrior-type will fail here is that it's a roll that is usually only part of a special skillset (like sneaking, hacking, opening locks or deciphering ancient writing) yet the whole group has to make the roll in this case.

Edge of Dreams
2013-01-07, 01:26 PM
I'd like to propose a third option that isn't being discussed here: no-cap (or very very high cap) with diminishing returns.

Classic example of this? D&D 3.5, 4e, and Pathfinder point-buy for character creation. Yeah, you're capped at 18, but with the way the costs scale, going over 18 wouldn't be very viable anyway.

Diminishing returns allow someone to choose to specialize heavily if they choose to, but discourages over-specialization, because you'll be losing even more of your well-roundedness.

CarpeGuitarrem
2013-01-07, 01:34 PM
World of Darkness is actually probably the best example of a "diminished returns" point-buy: you invest XP into a skill equal to 3 times the next rank in order to advance it. (e.g., if the skill is at 3, it would cost 3 x 4 = 12 XP to advance it to 4; you're also not allowed to skip straight to 5...if you want to advance it to 5, you have to advance it to 4 first, so advancing it to 5 would cost that 12 XP plus an additional 15 XP--27 XP total)

But I agree with what valadil notes, and to that, I'd propose something...

Implement a diminishing returns system, but also provide specialists with the ability to purchase something to make it easier to buy skills up. For instance, if you have a skill at Rank X or higher, it costs Y less XP per rank to advance it. (Or maybe offer a "feat" or other equivalent that can be purchased once you have that rank in the skill, call it "Specialization".) This makes advancing the skill still viable.

Morty
2013-01-07, 01:36 PM
World of Darkness is actually probably the best example of a "diminished returns" point-buy: you invest XP into a skill equal to 3 times the next rank in order to advance it. (e.g., if the skill is at 3, it would cost 3 x 4 = 12 XP to advance it to 4; you're also not allowed to skip straight to 5...if you want to advance it to 5, you have to advance it to 4 first, so advancing it to 5 would cost that 12 XP plus an additional 15 XP--27 XP total)


It should be noted that the upcoming rules update the God-Machine Chronicles, will introduce a new XP system that, if I remember correctly, will no longer work this way.

Saph
2013-01-07, 01:47 PM
Hmm . . .

Okay, regarding valadil's point, the current cost system I'm using works like this for the primary combat/magic skills:

{table=head]Rank | Next Rank Cost | Total Cost
1 | 3 | 3 |
2 | 2 | 5 |
3 | 3 | 8 |
4 | 4 | 12 |
5 | 5 | 17 |
6 | 6 | 23 |
7 | 7 | 30 |[/table]

. . . and so on. Would that qualify as diminishing returns, or is the cost increase too shallow?

valadil
2013-01-07, 01:47 PM
But I agree with what valadil notes, and to that, I'd propose something...

Implement a diminishing returns system, but also provide specialists with the ability to purchase something to make it easier to buy skills up. For instance, if you have a skill at Rank X or higher, it costs Y less XP per rank to advance it. (Or maybe offer a "feat" or other equivalent that can be purchased once you have that rank in the skill, call it "Specialization".) This makes advancing the skill still viable.

Acceptable! To continue with WoD, if I could buy a trait or something that lets me up a skill at 2x current instead of 3x current, that would probably be enough to keep my specialist special.

CarpeGuitarrem
2013-01-07, 02:07 PM
*snips handy chart out of quote*

Would that qualify as diminishing returns, or is the cost increase too shallow?
It depends, I'd say, on how hard it is to earn those points, how many you get over the course of play. It's definitely a diminishing returns scale, the bigger question is this: does the amount of points you spend overcome the diminishing returns? If having 25 points in a skill is trivial (working with GURPS numbers, for instance), then the diminishing returns would be negligible. On the other hand, if 10 points is the most you could muster for a skill, the diminishing returns impact you strongly.

What I'd do is this: figure out a basic loadout, assign points to that, and then let the players at it. For example...

I decide that I want players to have (on average) two skills at 5, four skills at 4, and seven skills at 3. That's 34 + 48 + 24 = 106 points, and depending on how generous/mean I'm feeling, I might round that to 100 or 110. That becomes the total number of spendable points.

So you could theoretically grab two skills at 7 and have room enough for a handful of lower skills (3s and 4s), for instance.

Acceptable! To continue with WoD, if I could buy a trait or something that lets me up a skill at 2x current instead of 3x current, that would probably be enough to keep my specialist special.
Yeah, likely something like this...

• Skill Expertise [skill of your choice]
Requires: • • • in that skill
The cost to advance the skill is 2x per rank, instead of 3x per rank.

You'd take it for each skill you wanted to have expertise in.

Though, it also looks like Saph isn't so much talking about advancing skills during the course of play, but just at character creation.

Saph
2013-01-07, 02:18 PM
It depends, I'd say, on how hard it is to earn those points, how many you get over the course of play. It's definitely a diminishing returns scale, the bigger question is this: does the amount of points you spend overcome the diminishing returns? If having 25 points in a skill is trivial (working with GURPS numbers, for instance), then the diminishing returns would be negligible. On the other hand, if 10 points is the most you could muster for a skill, the diminishing returns impact you strongly.

The current baseline I'm working with is that characters start with 150 points and gain more relatively slowly: a character with 300 points would be very experienced and high-level.

The bit I'm wondering about is the possibility for players to build characters with a single rank 10 skill, at a cost of 57 points, which even for a starting character would be expensive but not impossible by any means. It would make them incredibly good at one thing, but mediocre at everything else. I'm not sure whether that's a bug or a feature.


Though, it also looks like Saph isn't so much talking about advancing skills during the course of play, but just at character creation.

Both. I never really liked the White Wolf approach of having character creation work differently from character advancement, as it rewards players for min-maxing their attributes at creation and penalises them for starting with average stats (it's very expensive to buy Dexterity 4 up to Dexterity 5, but very easy to just start with Dexterity 5 to begin with). Instead I'm going with the L5R approach, where it costs exactly the same amount to buy a rank 5 skill at character creation as it does to advance it to rank 5 later.

valadil
2013-01-07, 02:29 PM
Hmm . . .

. . . and so on. Would that qualify as diminishing returns, or is the cost increase too shallow?

I would definitely call that diminishing returns. It's not as severe as WW or Deadlands, but it's still there. Keep in mind that I felt this way about GURPS 4, where there is an upper limit on the price of a skill point.

Siosilvar
2013-01-07, 02:35 PM
I'm not sure if they even work that way theoretically. For systems like GURPS, experienced players will generally tell you right at the start that points do not equate to personal power – the whole power-level way of looking at it just doesn't really work, because specialisations vary so much.

I do like the soft-cap idea: it discourages overspecialising without completely ruling it out. Not sure how to price it, though. At the moment I'm using arithmetic-progression costs (going from rank 2 to rank 3 costs 3 points, going from rank 3 to rank 4 costs 4 points, etc).

I wouldn't expect any game to have a perfect correlation between points and capability, but in theory one point buys you the same amount of usefulness (in general) no matter where you put it. To have it another way... kind of defeats the point of points. In practice, it doesn't work perfectly, because of specializations, temporary stuff, and so on, but point-based systems are generally made so that two equal point-valued things are about as useful as each other. At least, that's my impression of it.


As far as diminishing returns on points: Any sort of progression that increases cost should work, subject to playtesting of course. Arithmetic is good, 2n is fine, n2 works. It depends on how useful another rank turns out to be and how strongly you want to encourage horizontal growth.

You may or may not want to include a bypass for the soft cap - like a 4 point specialization bonus (that can only be taken once or twice for a skill). So instead of buying combat up to 5, a character could buy a +1 bonus with swords only for cheaper.


Or, you could go another way entirely (this one just came to mind): cap skills at their attribute. Attributes follow diminishing returns, skills may or may not depending on how you're feeling. Attributes don't add to skills, only calculated stats (if you have any).


EDIT: Also keep in mind the randomizer you're using. If you're roll-under on a bell curve like GURPS, there's already diminishing returns for higher levels of skills built into the dice once you get past the median.

CarpeGuitarrem
2013-01-07, 02:36 PM
The current baseline I'm working with is that characters start with 150 points and gain more relatively slowly: a character with 300 points would be very experienced and high-level.

The bit I'm wondering about is the possibility for players to build characters with a single rank 10 skill, at a cost of 57 points, which even for a starting character would be expensive but not impossible by any means. It would make them incredibly good at one thing, but mediocre at everything else. I'm not sure whether that's a bug or a feature.
That makes sense. I'd definitely be concerned by the one-trick-pony aspect there, but for me, a lot of it would depend on what each rank of the skill signifies. If 10 is the pinnacle of human ability, it seems fair game to me. If something like 7 is the pinnacle of human ability, I'd not allow it to be surpassed during character creation. Instead, you'd have to purchase some sort of supernatural power that increases the skill, allowing it to go past the "normal" cap.

RPGuru1331
2013-01-07, 02:53 PM
@Stubbazubba:

But what would caps accomplish in that scenario? A rogue-y character will always try to get the maximum in skills like sneak, and a warrior type, possible in heavy armor and other gear with penalties to sneak will likely disregard such skills.
In games with a more granular system where another +1 matters more (Which is n ot MnM 2e or 3e skills for sure), those caps are relevant, is the thing. MnM's caps tend to matter in battle, too.


Both. I never really liked the White Wolf approach of having character creation work differently from character advancement, as it rewards players for min-maxing their attributes at creation and penalises them for starting with average stats (it's very expensive to buy Dexterity 4 up to Dexterity 5, but very easy to just start with Dexterity 5 to begin with). Instead I'm going with the L5R approach, where it costs exactly the same amount to buy a rank 5 skill at character creation as it does to advance it to rank 5 later.
Yeah, the WW system has serious problems when your group has a disparity between how hard some people will minmax the system.


The bit I'm wondering about is the possibility for players to build characters with a single rank 10 skill, at a cost of 57 points, which even for a starting character would be expensive but not impossible by any means. It would make them incredibly good at one thing, but mediocre at everything else. I'm not sure whether that's a bug or a feature.
If that would actually be a problem for the game, just veto the concept.

Saph
2013-01-07, 02:56 PM
Or, you could go another way entirely (this one just came to mind): cap skills at their attribute. Attributes follow diminishing returns, skills may or may not depending on how you're feeling. Attributes don't add to skills, only calculated stats (if you have any).

Having the skill cap based off attributes is an interesting idea. I might use that one, thanks!


EDIT: Also keep in mind the randomizer you're using. If you're roll-under on a bell curve like GURPS, there's already diminishing returns for higher levels of skills built into the dice once you get past the median.

I'm using something more like the d20 system – higher is better, and your total is dice roll + attribute + skill. So there's no built-in limitation, except insofar as you eventually get to the point where you're passing all normal DCs automatically.


If that would actually be a problem for the game, just veto the concept.

Doesn't fly, I'm afraid. I'm not just running a campaign, I'm designing a system.

tensai_oni
2013-01-07, 04:17 PM
Doesn't fly, I'm afraid. I'm not just running a campaign, I'm designing a system.

Does fly. To give an example - M&M is far from balanced, despite the caps. It's very easy to create a very broken character if you know what you're doing, BUT the book also states that the GM has rights to ban broken or cheery characters. As opposed to DnD's approach of "if it's RAW it's legal".

Saph
2013-01-07, 04:24 PM
Does fly. To give an example - M&M is far from balanced, despite the caps. It's very easy to create a very broken character if you know what you're doing, BUT the book also states that the GM has rights to ban broken or cheery characters. As opposed to DnD's approach of "if it's RAW it's legal".

"It doesn't matter if the system is broken, because the DM can always fix it" is really bad game design. Yes, the DM can (and will have to) make adjustments on the fly, but if something as simple as putting lots of ranks into a skill will break the game, then the system is really badly designed.

BRC
2013-01-07, 04:42 PM
I guess a lot of it depends on how many different skills it takes to fill a particular role.

Let's say somebody wants to be a dude with a sword. If there is a "Swords" skill, and that is all he needs, then he'll sink everything into that, and you'll have an unstoppable sword-god who can't do anything else.
However, if he also needs to put skill points into being good with armor, then that changes thing. He could put all his points into Swords, and be an unstoppable sword-god who crumples as soon as he is hit.

So then he splits his skills between Swords and Armor. If he wants to be more versatile, it would benefit him to put some points into, say Polearms, or Bows. If the game assumes characters are versatile, then he is penalized for ignoring other types of weapons.

Basically, the more components there are to filling a specific role, the more acceptable it is to put caps on points you can spend, since it allows more versatility. Maybe a player understands that overspecializing in Swords will make him vulnerable in other ways, and therefore specifically chooses to do that.


If you put your caps too low, characters often cannot fulfill their full intention. I may want to play an armored knight, but once I've hit the skill cap on Riding, Swords, and Armor, I'm left with tons of points left over and I'm only marginally better with a sword than the party thief who randomly threw a few spare points towards swordsmanship.

Put your caps too high, and you get Players overspecializing, forcing GM's to build adventures that tailor to those specializations, or else face a bored player.

I recall some edition of Shadowrun had the idea of "Skill Groups", several complimentary, or related skills that you bought together for cheaper than buying them all individually. I like that, because it encourages versatility, at least in some regards.

CarpeGuitarrem
2013-01-07, 04:47 PM
"It doesn't matter if the system is broken, because the DM can always fix it" is really bad game design. Yes, the DM can (and will have to) make adjustments on the fly, but if something as simple as putting lots of ranks into a skill will break the game, then the system is really badly designed.
Yeah. On a usability front, such a system has poor usability for the GM, which really means that giving the responsibility for this facet to the GM is too much to ask.

Good game design helps the GM make better games. (Though I can see why M&M said "screw it, no way we'll be able to figure out all the combinations".)

The_Snark
2013-01-07, 04:55 PM
My brief two cents: I prefer having caps of some kind, whether hard or soft. I remember being introduced to Mutants and Masterminds and the HERO system at around the same time. Both are point-based, level-less systems originally designed for playing superheroes, with a wide array of available powers and options. Mutants and Masterminds features a cap on skills (and attacks, and defenses), as people have pointed out; the GM sets a power level for the group and that determines the cap. The HERO system features no such caps; you can raise any stat or power as high as you want, provided you can pay for it (and the costs were linear). As a new player, I found this hard to deal with; I had no idea what an 'average' enemy's attack or defense would look like, and had to ask the GM whether my character was liable to fold with one superpowered punch. I had the distinct impression that this problem never goes away entirely; the GM and players have to be very careful not to introduce too much variance. To a certain extent this is true of any game, but in HERO it could very easily happen accidentally. Essentially, each group has to make up their own caps as a social contract, or accept some imbalance between characters.

(Arguably, the variance was intentional in the HERO system. If you have the Hulk (dumped all his points into hitting really hard and being absurdly tough) and Batman (generalist with solid but still human stats, lots of skills and talents) in the same group, then realistically anything that can scratch Hulk will splatter Batman, and anything Batman can hurt with a punch will fold after a single Hulk-punch. It's realistic, and as a comic book it has potential, but it's not great for a tabletop game in which these two are expected to fight crime together.)

Bottom line: I think having some kind of cap is a good idea. It can be a hard cap or a soft cap, but either way you want to give players some guidelines as to how much they can/should invest in any given skill.

BRC
2013-01-07, 04:58 PM
"It doesn't matter if the system is broken, because the DM can always fix it" is really bad game design. Yes, the DM can (and will have to) make adjustments on the fly, but if something as simple as putting lots of ranks into a skill will break the game, then the system is really badly designed.
I think most games should have a note somewhere that says "If you encounter a situation where you think the rules should be changed, change them". That doesn't necessarily make the system better, it just encourages better DMing.

A solution to a problem isn't really a solution if it requires you to go outside the rules/is hidden from all but the most in-depth scrutiny.
Spoilered for pointless rant.

I remember, in years past, complaining about how frequently monsters in DnD were immune to sneak attacks, and how it sucked to play a rogue when your main ability was ignored by tons of monsters.

Somebody responded

"That's not a problem. All you have to do is put points into Use Magic Device, then own Spell Compendium, notice the spell "Gravestrike", buy a wand of it, then own the Rules Compendium, look through that for an unmarked errata saying that the casting time for wands is the casting time for the spell, rather than a standard action (as is indicated in the SRD), allowing you to use a swift action to cast gravestrike before each attack."

Just because a workaround exists does not mean there is no flaw in the system.

Saph
2013-01-07, 05:23 PM
Useful stuff

Thanks, that's helpful. At the moment I'm leaning towards some kind of soft cap, which allows players to specialise if they really want to but which prevents the Hulk/Batman problem from getting too extreme.


Spoilered for pointless rant.

I remember, in years past, complaining about how frequently monsters in DnD were immune to sneak attacks, and how it sucked to play a rogue when your main ability was ignored by tons of monsters.

Somebody responded

"That's not a problem. All you have to do is put points into Use Magic Device, then own Spell Compendium, notice the spell "Gravestrike", buy a wand of it, then own the Rules Compendium, look through that for an unmarked errata saying that the casting time for wands is the casting time for the spell, rather than a standard action (as is indicated in the SRD), allowing you to use a swift action to cast gravestrike before each attack."

Just because a workaround exists does not mean there is no flaw in the system.


Don't forget that you need a spare hand to activate the wand! Which rules out TWF, unless you get a Wand Chamber, which is in Dungeonscape, which requires . . . :smalltongue:

erikun
2013-01-08, 12:10 AM
Stubba: I can see how that makes sense, but the examples you're using are all level-based systems, like D&D, rather than point-buy systems. Most point-buy systems don't have a 'level' or 'challenge rating' system.
Fate/Fudge has a cap of Legendary (or Epic, if the GM allows) and is not level-based, so there is that. You aren't likely to see many skills that high at first, so they are limited to high-level characters.

Burning Wheel doesn't have any skill level caps, but the increasing requirements for each skill level (third the skill level in challanging rolls, half the skill level in near-impossible ones) pretty much guarantee that progression will slow down or stop after a point.


Hmm . . .

Okay, regarding valadil's point, the current cost system I'm using works like this for the primary combat/magic skills:

{table=head]Rank | Next Rank Cost | Total Cost
1 | 3 | 3 |
2 | 2 | 5 |
3 | 3 | 8 |
4 | 4 | 12 |
5 | 5 | 17 |
6 | 6 | 23 |
7 | 7 | 30 |[/table]

. . . and so on. Would that qualify as diminishing returns, or is the cost increase too shallow?
That looks pretty shallow, at least for what you seem to be wanting. I guess it really depends on what you expect the "normal" skill rank value to be. If most people will be around rank 3, then yeah, costs quickly add up over time.

On the other hand, there isn't much difference between 6 skill points or 7 skill points or 8 skill points. If you expect most characters to be around rank 6 or so, then the costs to advance two ranks isn't really much different than advancing one rank twice.

RPGuru1331
2013-01-08, 01:10 AM
Doesn't fly, I'm afraid. I'm not just running a campaign, I'm designing a system.
You want to make a one-size fits all solution, hardcoded into advancement rules? Good luck with that, I'm sure that will both turn out fine and not drive you insane in the design process.

erikun
2013-01-08, 01:27 AM
While a character creation system that allows such singlely-focused characters may not be a concern to a good GM - who can just review characters and veto the obvious abusers - I think that it should be a concern to a good system designer. After all, the designer will be handing the system over to both good GMs and bad GMs, and should have both in mind. Plus, a good system design will mean less work for the GM to tackle while reviewing characters (even if they still do so anyways).


Now that I think about it, I might recommend limiting how much one character can focus on a specific task. Either keep them from spending too many points on a one ability ("No more that half the points at character creation in one skill.") or do something like the skill-pyramid I've seen in Fate, which requires buying more low-level skills for a higher-level skill.

The reason for this is because, even with diminishing returns, there will still be players who insist on maximizing the value of a particular roll. Yes, they may have given up ranks in a dozen other skills for that +1, but they still managed to have a higher roll than anyone else! The old one-eyed, one-toed, deaf illiterate crackshot sniper is still a possibility if you allow characters to dump all their points into any one skill.

Fortuna
2013-01-08, 01:51 AM
It really drastically depends on what you want the 'normal' character to look like and what exactly you're trying to achieve for borderline characters, as others have said. If your goal is to ensure that specialization is costly, but doable, then diminishing returns are your friend. If you expect every character to have a wide range of skills, then the aforementioned skill pyramid will help you out.

However, there is a point to consider about diminishing returns - there are two ways to implement it. On the one hand, there's the method that's been discussed in here for the most part - scaling costs as you invest more into a skill. But there's another way; you can also create diminishing returns at play time.

The major reason why diminishing returns systems and skill caps tend to be so important in many character creation systems is the existence of increasing returns in play. It's what I call the Law of Gaming Numbers - 2X is better than X+X, pretty much every single time. If you deliberately avert this in building your system - whether by creating synergy bonuses or requiring multiple skills to function effectively in an archetype, or simply by making beneficial results take more and more investment to reach when you come to roll the dice - you can keep player-side rules simple and get the complexity out of the way on your end.

Stubbazubba
2013-01-08, 01:58 AM
True, but I can hardly think of any system that doesn't have the "one guy's worst beats another guy's best" thing. Low-level D&D, maybe, but that's only because the d20 has so much variance. FATE, M&M, Shadowrun, L5R, White Wolf all make it near-impossible for a specialist to lose to an incompetent.

By all means, make it near-impossible. The problem I'm describing is that it is straight-up impossible for these two people to have run the same challenge in the same party. Someone isn't even rolling, because thought has not been given to how far different characters in the same party can be from each other in a single skill.

If you're going to allow characters in the same party to have Stealth scores that are more than the entire RNG apart, you are automatically saying that someone doesn't really get the same experience as everyone else when they play Stealth challenges, either the expert or the incompetent, because their outcome is determined without rolling while everyone else gets to roll. I don't know why you would be OK with that. I can't see what possible advantage it brings to allow such disparity except playing to some notion of complete character-building freedom which is, IMO, not nearly worth the price paid at the table later.

Knaight
2013-01-08, 02:31 AM
I can't see what possible advantage it brings to allow such disparity except playing to some notion of complete character-building freedom which is, IMO, not nearly worth the price paid at the table later.
To come back to the forum briefly to comment:

I could see it as a matter of extreme specialization, where simply not rolling works for pure mastery in a skill, particularly in games that emphasize most action being outside of a characters specialty. Combined with troupe play (where everybody has multiple characters), it would be something that only shows up rarely, wherein the strongest characters available to the players that are less frequently played do not need to roll under most circumstances within their chosen field. Say you have a group of a few largely normal peasants and poorer tradespeople, a well traveled guard or two, and an important scholar with a high enough position that everyone else seems like a retinue. If a court document is found, everyone but that scholar should probably be rolling to read the thing at all, whereas it simply doesn't make any sense for them to have to do so. Similarly, if another lead character is along and it is some glorious titled knight, they might not have to bother with defense rolls in combat at all, as they are never actually threatened between being way better armed, way better armored, and way better trained than anyone else there.

Tehnar
2013-01-08, 07:00 AM
I propose a cap at character generation. That way you won't get monster PC's out of the box.

During the game there shouldn't be caps, especially if there a diminishing returns on invested points. The biggest part of design is not to allow gear to add direct numerical bonuses (at least not comparable to the characters skill), make it add other benefits.

Another way of balancing things is in the way the game runs. Games which allow multiple skills to be used will discourage overspecialization.

DigoDragon
2013-01-08, 07:51 AM
My group is in a Shadowrun 4e campaign right now.
I'm in favor of caps because it helps push players to diversify. My spellcaster has pretty much maxed out her dice pool on spells due to the cap, so now she's moving into improving her Charisma based skills, becoming better at Ettiquete, Con, and Diplomacy.

The melee adept can pretty much crush anything with hand-to-hand skills, but as he's hitting the cap, he's now investing in some ranged combat and can cover others in a gun fight.

Saph
2013-01-08, 08:16 AM
On the other hand, there isn't much difference between 6 skill points or 7 skill points or 8 skill points. If you expect most characters to be around rank 6 or so, then the costs to advance two ranks isn't really much different than advancing one rank twice.

Yeah, that's pretty much what I was concerned about. Ideally I'd have some sort of shallow exponential curve, but that's difficult to do neatly with such small numbers unless you break out the fractions.

I'm currently thinking of saying that after a certain point – rank 7, say – costs increase from 1x the new rank to 1.5x the new rank. Rank 7 is also where one or two other things change in the system, so it's a natural break point. That would make 5-7 ranks in a skill "professional" level, and 8+ ranks only for those who really want to specialise.


However, there is a point to consider about diminishing returns - there are two ways to implement it. On the one hand, there's the method that's been discussed in here for the most part - scaling costs as you invest more into a skill. But there's another way; you can also create diminishing returns at play time.

What sort of thing are you thinking of?


My group is in a Shadowrun 4e campaign right now.
I'm in favor of caps because it helps push players to diversify. My spellcaster has pretty much maxed out her dice pool on spells due to the cap, so now she's moving into improving her Charisma based skills, becoming better at Ettiquete, Con, and Diplomacy.

That's interesting. How does the cap in your campaign work? I'm guessing it's a houserule since I don't remember seeing any caps in the Shadowrun book.

erikun
2013-01-08, 01:32 PM
Yeah, that's pretty much what I was concerned about. Ideally I'd have some sort of shallow exponential curve, but that's difficult to do neatly with such small numbers unless you break out the fractions.

I'm currently thinking of saying that after a certain point – rank 7, say – costs increase from 1x the new rank to 1.5x the new rank. Rank 7 is also where one or two other things change in the system, so it's a natural break point. That would make 5-7 ranks in a skill "professional" level, and 8+ ranks only for those who really want to specialise.
One nice thing about the design phase is that the number and math can be as ugly as you want them to be. Put together a set of numbers that does what you want it to, and then check if things work in play as you'd like. You can smooth things out later, once you get the numbers that work. (or make a table, if necessary)

Also, during the design, you can make the system as granular as needed. It could cost 200 points to get to rank 2, then 300 to rank 3, then 450 to rank 4, then 675 to rank 5, and so on. If this produces the natural curve you are looking for, look for something that mimics it with smaller numbers. (Fibonacci
sequence, maybe?)

Arbane
2013-01-08, 01:34 PM
Fate/Fudge has a cap of Legendary (or Epic, if the GM allows) and is not level-based, so there is that. You aren't likely to see many skills that high at first, so they are limited to high-level characters.


There's also Fate's "Skill Pyramid": If you want to have a skill at rank X, you need to have at least two skills at rank X - 1. So even the world expert at basketweaving is also going to be good at some other things.

warmachine
2013-01-08, 02:06 PM
I don't see a problem with lack of caps as long as the tasks vary in category. In a modern, investigative campaign, for example, players need facemen, wire tappers, researchers, drivers, burglars etc. as well as gunmen. Everyone can work out how their skills can advance the plot and get their turn in the spotlight as long as there are various ways to complete an objective. Even if that gunman can't do anything but shoot well, he can still fend off guards whilst the others ransack an office for evidence.

Fortuna
2013-01-08, 02:16 PM
I mentioned a couple of methods later in my post. Making synergy bonuses a significant part of the game provides tangible rewards for diversification, and requiring more than one skill to be really effective in an archetype creates a natural form of diminishing returns if you need to choose between them - the example of a godlike swordsman who crumples as soon as he's hit was mentioned earlier in the thread. Beyond a certain point, you will be more effective in your archetype by covering all your bases than by pushing one skill to ludicrous extents.

Building diminishing returns directly into a more traditional system is harder. If all a swordsman needs is the Swords skill, then the system itself needs to make having a high Swords skill less important after a while. Given that you intend to have relatively slow advancement after a point, one way to do that is to build in the assumption that once you push your swords skill above, say, 10, only truly legendary opponents will be able to compete with your blade. It's important that the system be designed with this decision in mind - for example, it needs to be extremely difficult to remain unnoticed while attacking with Stealth, and while being a world-class liar is well and good it can make storytelling difficult if you're a world-class negotiator, so while Bluff fits into such a system Diplomacy doesn't really.

BlckDv
2013-01-08, 03:13 PM
Disclaimer; I did skip some of the posts in the middle of the thread, sorry if they provide this info.

I think a critical question for me to answer this is what kind of story is the system made to support?

If I'm running "The Princess Bride" campaign, My system needs to allow me to dump all my points into my duelist skills including sneaky off hand expertise, mastery of styles, and perhaps let me buy more ranks thanks to my strict code of honor, my burning obsession and my alcoholism (I said perhaps; not all systems need give points for "flaws".) If the system caps my skill so that I am forced to pick up archery as well, I may end up shooting the six fingered man across the room, and no one wants that. I am the master of swordplay, and while I may learn valuable lessons on my journey, my sword skill probably won't improve much beyond where it started. That is good; it fits the story need.

On the other hand if I'm telling "The Hobbit" it is probably good that I can't dump all my points into troll-slaying and end up dumping some points into cooking here, and singing there. My story will show me grow to meet challenges, learning the skills that my trials teach me I need, and letting me find unexpected moments where the marks and mementos of my mundane life suddenly shine through and come into play. ("Wait.. you don't want to eat the dwarf like THAT!" I roll a Cooking check to critique their setup) At the end of the tale I have grown noticeably in many skills, and likely find myself unable to fathom living the life that my starting skills would have let me.

Caps serve the needs of a story that wants to express classic themes of coming of age, fish out of water, or joe six pack standing up for himself, a guy finding his place and growing into it.

Capless systems are the vehicle for Epic Poems (Notice how capless systems almost ensure you have an "Achilles heel" as the cost of your excellence), Superspy tales, and high fantasy tales of the chosen one; a hero potent from the opening words, perhaps lacking knowledge of his potential, perhaps just reluctant to express it.

Now; for myself I will say that I do usually prefer the L5R style over WoD. Systems in which creation and advancement are both point based, but use radically different rules bug me a lot. And don't get me started on the nonsense of WoD's claiming that the older a vampire is the more slowly he learns but then having a system that ensures that your vampire will learn more in the next few years (or even months) than in the past several decades (or even centuries).

warmachine
2013-01-08, 07:20 PM
Pity you're not using bell curve dice rolling like the 3d6 of GURPS. Over 10, the benefit of each +1 diminishes. Generalists might elect for many level 12 skills (74% chance) and specialists have fewer level 16 skills (98% chance). Generalists are better when normal situations demand +0 modifier because they have a greater range of applicable skills but specialists are better when enemy interference require, say, -4 modifiers (26% vs. 74%).

For example, a character with Administration 12, Lockpicking 12, Bluff 12, Stealth 12, Diplomacy 12 and Research 12 can generally track down and meet a previous, scientific team now scattered across the country but it may take a character with Diplomacy 16 and Research 16 for those members who've gone into hiding.

That is, there is no arms race between characters as sometimes high skill levels are better, sometimes medium skill levels are better.

Saph
2013-01-08, 10:25 PM
Pity you're not using bell curve dice rolling like the 3d6 of GURPS.

Eh, I always found the GURPS approach a bit unsatisfying. I like the simplicity of "higher is better" – same reason I prefer D&D 3.0's approach to Armour Class and DCs than AD&D's THAC0.

Cuaqchi
2013-01-08, 11:00 PM
That's interesting. How does the cap in your campaign work? I'm guessing it's a houserule since I don't remember seeing any caps in the Shadowrun book.

There actually are some soft caps presented in the book... Although it seems that most people ignore their presence.

For All Characters
- Floating Stat/Skill Augmentations limited to Natural Value x 1/2
- Skill Rank 6 Maximum without Aptitude Quality (Max 7 with it)
- Maximum 4 IP from ALL Sources
- Encumbrance Penalties (-1 AGI/REA for every 2 points of highest Armour Value over 2x BOD - Custom Armour & Military grade 3x BOD)
- One Specialization per Skill (None for groups)

For Starting Characters
- Only one Attribute to maximum at start
- No more than 1/2 Points in Attributes
- Starting character skill maximums 1 Skill Maxed or 2 Skills at Rank 5; As a result no Groups beyond Rank 4
- Max 35pts in any type of Quality (35 Positive & 35 Negative)
- Starting Gear maximum Availability 12 (20 with Restricted Gear Quality)
- No more than 50pts in Gear (More available with Qualities - Rich+/In Debt-)

The result is that by the end of it a human can roll at the absolute best about 20-25 dice for any skill and other metatypes can move from this point to their own natural limits. 6 (Attribute) + 1 (Exceptional Attribute) + 3 (Augmented Attribute) + 6 (Skill) + 1 (Skill Aptitude) + 3 (Augmented Skill) + 0-5 (Flex Value: Vehicle Handling, Tracer Rounds, Smartlink, etc)

warmachine
2013-01-09, 07:16 AM
Eh, I always found the GURPS approach [bell curve] a bit unsatisfying. I like the simplicity of "higher is better" – same reason I prefer D&D 3.0's approach to Armour Class and DCs than AD&D's THAC0.
It is possible to be mathematically equivalent with DCs or target numbers with skill levels 10 lower and adding 3d6 to match DC 11. Traveller is similar with 2d6 and TN 8.

Mind you, exponentially increasing point cost is similar as well in terms of diminishing returns.

warmachine
2013-01-09, 07:26 AM
I think a pertinent question is: can a really high skill level be a campaign or major milestone win button, rather than merely a win for a task that is part of a multiple task plan? If so, caps are needed. If not, they're merely guidelines for those who don't understand dice probability.

DigoDragon
2013-01-09, 08:13 AM
That's interesting. How does the cap in your campaign work? I'm guessing it's a houserule since I don't remember seeing any caps in the Shadowrun book.

The cap is basically "Your dice pool caps at 20 dice for any test."
This includes all sources of dice; stats, skills, bonuses, edge, etc. Any dice over 20 simply don't count. The idea is to force the players to diversify so that they can handle a better variety of situations the GM throws at them.

Saph
2013-01-09, 08:53 AM
I think a pertinent question is: can a really high skill level be a campaign or major milestone win button, rather than merely a win for a task that is part of a multiple task plan? If so, caps are needed. If not, they're merely guidelines for those who don't understand dice probability.

I'm mostly concerned about it with regard to combat – I can't see it being a problem for anyone to be amazingly good at Athletics, but incredibly high attack/defence skills are more of an issue. I'm leaning towards the soft-cap approach at the moment.


The cap is basically "Your dice pool caps at 20 dice for any test."

Ah, OK. I had the feeling that 20 dice is already a fairly huge amount in Shadowrun, though.

warmachine
2013-01-09, 10:27 AM
I'm mostly concerned about it with regard to combat – I can't see it being a problem for anyone to be amazingly good at Athletics, but incredibly high attack/defence skills are more of an issue. I'm leaning towards the soft-cap approach at the moment.
Good point, combats are longer than other task resolution and a dominating PC could leave players bored. However, it still depends on the setting. In a modern setting, high attack/defence itself doesn't help against an enemy soldier who's sneaked up unseen and empties an assault rifle at short range. A combat monster may dominate fights but he still needs other PCs to cover his flanks. Not so with a medieval swordsman with incredible attack/parry/HP.

So the question reduces to whether team tactics are necessary.

Cuaqchi
2013-01-09, 11:08 AM
20 is a lot but considering that most PC's start at 12-15 it isn't that far apart. Also look at the values for mooks. PR 0/1 ~ 6-9 Dice, PR 2/3 ~ 10-13, PR 4/5 ~ 14-17, PR 6 ~ 18-20. Characters should start around P(rofessional) R(ating) 3 or 4 which is the range from Upper End Corporate Security to Military GI and by the end can reach the echelon of elite Military Task Force (Tir Ghosts). Also note that you need that many successes to actually break some armour.

Below is an example of some of the math.
A Red Samurai (PR 5) can have 15-20 Armour with generally 6-8 REA (2-3 Successes), Shooting at him with a High Power Assault Rifle (8P -2 AP) still needs 6-11 Net Successes for lethal damage. That means between 20-56 dice needed to attack for a lethal result. Also remember that the DV boost from automatic fire is only lethal if the attack itself is lethal. Against high end bad guys (properly equipped) even 25 dice isn't an issue because you are putting everything into stun damage which can be negated or recovered with the right skills and equipment. Also note that said Red Samurai is going to save against the damage with his 20-25 dice of Armour+WIL meaning that unless you rolled really well his 6-8 Successes could keep him standing unless you broke a total DV of 18-20.

BlckDv
2013-01-09, 11:10 AM
After reading more comments and mulling it over, I'm curious if we may be trying to measure the wrong thing. Rather than determining if a cap is needed as an absolute, why not look at what the in play outcome of having a power at rank X is, and giving that info to the GM to back build limits on the play rather than the generation.

Examples:

If System A has a linear 1,2,3 skill system that provides a static bonus to some RNG (I'll use a d6 for ease of example), and a "Normal" difficulty is set at 6, and a "Hard" at 10, then a Gamemaster wanting characters to not perform well above Normal could limit characters to applying 3-4 ranks to a given check, while one who wants characters to have a real shot at making Hard checks from a base roll without other bonuses might allow characters to apply 6-7 ranks.

If System B gives a larger generation pool, you can cap the pool instead of the power, for example in a Westwood Star Wars/Shadowrun style you get an extra die per point, using the same d6 as before, if you have a "Normal" difficulty" set at 15 and a "Hard" at 30, a GM who wants PCs acting in the normal range could cap the pool at 5-6 dice, a GM who wants PCs who can reasonably try a Hard task might cap the pool at 9-10 dice.

This switching the cap to the Outcome instead of the Input means that a player can elect to stop buying at the level where he can perform (I'll buy 3 points in Aim and 3 in Handguns.. I'm only allowed to use 6 so I'm done) but then circumstances can give him trouble, (-2 for bad weather.. this shot will be rough, My gun jammed, I have to make a Smarts and Handguns roll instead of Aim .. crud) where as another player went and bought 2 Aim 7 Handguns... shooting away in base conditions he gets the same 6 dice and doesn't outperform his buddy, but man, even in a bad storm at night he keeps up that same level of accuracy, he is a machine! And his Gun knowledge is very strong, he can repair handguns even of types he is just exposed to, and can easily adapt scavenged accessories from one model to another, etc.

To me this may be a better solution than hard stat caps. If the GM allows the cap to scale with play, it can also reward front loading. (When you reach 100 points, the cap goes from 6 to 8, etc.) Now that "9" I've been treating as a "6" in shooting jumps to an 8 at the 100 point level even if I elect to spend my new points to pick up that skill in Ancient Farsi the group needs so much.

Dimers
2013-01-12, 07:14 AM
Eh, I always found the GURPS approach a bit unsatisfying. I like the simplicity of "higher is better" – same reason I prefer D&D 3.0's approach to Armour Class and DCs than AD&D's THAC0.

So do the same thing to GURPS that 3.0 did to AD&D. Untwist the definitions until high numbers are good in all cases. I am, in fact, doing so myself, writing my own game system ...

Example: Let's say a good skill level is in the 12 to 18 range. If I want someone with decent skill to have a small chance of success, and someone with amazing skill to have a pretty good chance, the difficulty rating would be set somewhere around 26. With 18 skill, you succeed on an 8 or better on 3d6 ... with 12 skill, you have to roll at least 14.

You get the bell curve without the "YES! All sixes! Oh, wait -- that means I died, doesn't it?"