on designing roleplaying games

[vasilidor]

Some people think that in building a role playing game one should start from what they think the first level characters should be able to do and go up from there. Have any of you thought about taking the opposite approach, starting from what you wanted the end game to look like and going backwards? like with the fighters of DnD, if the end goal is Thor (or something compatible) then after you get the basic math in place, build thor, if the end goal is conan, build a conan, and then work backwards to figure out the steps to get to thor or conan. you can then build your Merlins or Gandalfs to match them at the final endgame level.
thoughts?

[Mechalich]

There are many difficulties with doing this, but the simplest one is that depending on the mathematical system you're using you may run out of any ability to make small enough numbers and be unable to model basic things.

D&D 3.5, for example, has a well-known 'House Cats kill Commoners' meme, which is a consequence of it being impossible to do less than 1 point of damage with an attack rendering anything capable of dealing damage shockingly deadly at the low end. This is easily solved by adding in fractional damage, but the core mechanic of D&D (and essentially any other RPG using integer-based dice) does not all that option.

Functionally therefore, if you start by building a system attempting to model the most powerful thing in a fictional world you may find yourself unable to use that system to model incredibly important common things, like well, commoners.

Now, it is relevant to consider at the beginning of the design phase, whether or not your game has a functional 'masses' with which major characters are expected to interact. It is entirely possible that a setting does not. In DBZ, perhaps most famously, the non-martial artists simply don't count. Evening the weakest of the main cast can defeat arbitrarily large numbers of ordinary humans (even military personnel) and therefore it is not necessary to mechanically model their capabilities at all.

[Zombimode]

if the end goal is Thor (or something compatible) then after you get the basic math in place, build thor, if the end goal is conan, build a conan, and then work backwards to figure out the steps to get to thor or conan.

But what would be your frame of reference? How can you define what a high level character can do in a vacuum?

This is much easier when you start with a baseline competency. The baseline is the norm. High level is a progression of the bases and as such only in a relation to the base.

[NichG]

From a philosophical perspective I'd say do neither - start big picture with things like 'What sorts of questions is this system asking the players? What sort of ideas does it want to teach, or allow people to explore?' and then the internal progression should be matched with stages of that goal. E.g. if you want players to grapple with 'what does it mean to be human?' then progression is a tool that lets players transition from baseline to fully synthetic beings or other more exotic forms of existence one step at a time so each contrast is explored (versus 'I start as an AI' or similar). If you want players to think through complex reactive martial arts combos at a fine-grained level of tactical detail, then starting from baseline and layering is probably not a great idea since the high-end complexity of play could easily become unstable, so you need to be designing with the whole range in mind.

In practice, what I probably do most often is to start near the mid-range and push in both directions from there. I have some idea of what 'ongoing play' should feel like, and then pushing down details the build-up while pushing up details the climax or destabilization of that long-run zone. For more complex systems, I might intentionally make two or three such zones and link them with breakpoints that invalidate the assumptions which would otherwise define play at the lower zone.

For example, in Mythclad I have a low-end stable zone which is basically sky pirates stuff, where transportation and access to physical locations are all relevant challenges, and threats and tensions tend to be about people, reputation, wealth, politics, threats of physical destruction, etc. Then there's a high-end stable zone that has to do with fighting over the conceptual structure of the world - which stories will be remembered and which will be forgotten, what will people's ethics and values be, what aspects of the fundamental structure of reality are vulnerable, changing, or being co-opted by alien forces, etc. The bridge between these zones of play involves a switch from extensive mechanics (how much damage do you do, what do you hit, etc) to intensive mechanics (what sorts of things can you manipulate or control directly as an act of will; what sorts of effects can you bring about, bypass, or twist; etc). There's also an onramp at low level play where randomness drops away and stops softening mechanics, so success or failure gradually becomes more about understanding how things fit together, and an offramp at high-level play where you're basically a living god and the mechanics sort of step away and say 'beyond this point, its more about free-forming things anyhow since what matters is concepts not numbers'.

[Martin Greywolf]

There are two sides of argument here, and they have nothing to do with each other.

Designing from up to ground

This is a misunderstanding of how a designing a thing works. You don't go "OK, let's make ten classes, and now give them 1st level abilities, then give all of then 2nd level abilities etc". When you design something like this, you first determine what your boundaries are, and go from there.

So, a 1st level character is someone who is an utter noob/already a competent person/a 1st year student at a magical school. That's your lower boundary. Then you decide that your final level character is someone who can sprint faster than light/can sprint at the level of olympic athlete/can sprint fast enough to consistently beat olympic athletes.

Then, you design all the in-between tiers, and all of this is dependent not only on the setting, but also on the genre, and also on the specific story. Just compare what a standard trained mook can do in Witcher to some of the Wuxia stories.

So yeah, you do take this into consideration from the very start, or at least you should.

DnD specific casters vs martials problem

This is not something that is inherent to any design approach, really, it's just that DnD has severe problems in two areas.

First area is player expectations - some people want their fighters to be able to only do things that are possible to do in reality by peak performance humans. Others want them to go full anime, or medieval legends, and be able to toss a dozen spears, jump off of them to reach their target and kill him by stabbing with a sword. What's worse, DnD is meandering on where it wants to be officially, with vanilla fighter being the former, and something like Tome of Battle on the latter side.

Second issue is one of consistency - DnD doesn't have a core mechanic that is used for resolution of a thing. You have attack bonuses, you have skills, you have spell descriptions and spell slots. So, when you want to balance a caster with a martial, you have to essentially balance two completely disparate mechanical systems. Hell, not even Starcraft is this bad, at least the core of that is the same for each race.

Even worse, spells are balanced in part based on their description, so you have one system that is almost pure math and another that isn't, and you need to know all the spells and their descriptions to balance properly. At this point you may as well give up.

Compare that to something like FATE Core where you always roll skill+dice+bonuses (usually no more than 3), no matter if you are casting a spell or stabbing a dude.

[Cluedrew]

I think the starting point should be: What do you want the system to be about? But assuming you have gone through all that and are now just focus on character abilities I think it kind of doesn't matter because you should keep things flexible and move back to old sections as design work in a new area pushes against them. Starting with a couple of fixed points (such as NichG's stable regions or starting and end-game character) might help, but you should never consider them "done" until the entire system is complete.

On and play-testing. Play-test everything you make with different groups, different characters, different builds and different campaigns. Try everything you can think of and try a couple of random things at random and see if you find something you didn't think of. Then get people to play it without you there and listen to what they say and be prepared to make even more changes. I feel a lack of play-testing is part of the story that lead to the problems you are talking about.

A bit hurried but those are at least my initial thoughts.

[Mastikator]

I'd start with the world Thor or Conan lives in to provide context for their power level. Specifically what could challenge them, what would be at a appropriate stake and how much spectacle should there be.

I also wouldn't just try to build back or forward, start with a few milestones, initial, final and a couple in between (depending on the difference between start and end). Do the same exercise as above for these powerlevels. This should give you a ballpark idea of how the experience/leveling/advancement system should look like, because it needs to accommodate all of them, especially the early ones.

Now that I mention it I'd argue that the early game is key, so put all your effort there. I also agree with playtesting. I've only done one playtest for my game and it was a treasure trove of data.

[Man_Over_Game]

Designing from up to ground

This is a misunderstanding of how a designing a thing works. You don't go "OK, let's make ten classes, and now give them 1st level abilities, then give all of then 2nd level abilities etc". When you design something like this, you first determine what your boundaries are, and go from there.

So, a 1st level character is someone who is an utter noob/already a competent person/a 1st year student at a magical school. That's your lower boundary. Then you decide that your final level character is someone who can sprint faster than light/can sprint at the level of olympic athlete/can sprint fast enough to consistently beat olympic athletes.

Then, you design all the in-between tiers, and all of this is dependent not only on the setting, but also on the genre, and also on the specific story. Just compare what a standard trained mook can do in Witcher to some of the Wuxia stories.

So yeah, you do take this into consideration from the very start, or at least you should.

This is what I'd vote as the appropriate answer. On a scale of 1 to 100, estimate what both the 1 and the 100 looks like first. From there, estimate what 50 looks like. And then start dividing it even further at 25 and 75, until it's roughly balanced across all levels.

The reason this fails in some tabletop games is often because developers are only considering one element of play (such as combat), or they're expecting players to build a foundation as you progress from 1 and don't really have an estimate or an expectation as to what "endgame" is supposed to look like.

[Duff]

I suggest this is a bad idea. Your starting characters are a "Fixed point", the high level isn't as much.

Which is not to say you have to create your starting point and then everything else flows up from there without any calibration. The details varies depending on what sort of game you want to produce, but generally I suggest:

Rough out your starting point to be interesting and give characters some ability to specialise and differentiate themselves. Power check*
Run up the experiance tree a bit with characters becoming more different. Are they still interesting? Have they grown more into the character the player wants? powercheck. if it doesn't work, go back
Run further up the tree, powercheck. Flavourcheck ** if it doesn't work, go back
Run up to the very high end. powercheck. Flavourcheck if it doesn't work, go back.

Then do the same again with more detail - features that were " summon an animal companion" now get a list of the types of animals, with stats and rules about replacing animals that die etc

* Powerchecks compare the different types of character for balance. This should be done subjectivly. Can each character influence the outcome of events in a reasonable number of situations? Do their choices and die rolls matter? May be unimportant in some flavours of game, but that's rare.
** Flavour checks are about whether a character is growing into and evolving a theme. Do they fit the type of stories the game is going to tell? Can the player still use the concept they had at the start or have they been forced into a different path? Have their choices made the character more interesting?

[vasilidor]

I may have been confusing design styles. in some game types, the end is designed first and then they work backwards from there, I am referring to video games done by professional studios here. an example is the legend of zelda games: in order to reach ganon, you need to accomplish A, in order to accomplish A you need to do B Etc. Role playing games do need the variety of play levels, or benchmarks, in order to make comparisons now that I think of it.

[Mastikator]

I may have been confusing design styles. in some game types, the end is designed first and then they work backwards from there, I am referring to video games done by professional studios here. an example is the legend of zelda games: in order to reach ganon, you need to accomplish A, in order to accomplish A you need to do B Etc. Role playing games do need the variety of play levels, or benchmarks, in order to make comparisons now that I think of it.

Isn't that more in line of how to design a plot/story/quest/module though? In that case I think it's valid but to a lesser degree in table top RPGs where PC actions tend to be unpredictable.

I think you should clarify here what you're asking. Are you asking "how to design a RPG ruleset/framework" or "how to design a campaign setting" or "how to design a module"? These are all very different questions.

[Man_Over_Game]

Isn't that more in line of how to design a plot/story/quest/module though? In that case I think it's valid but to a lesser degree in table top RPGs where PC actions tend to be unpredictable.

I think you should clarify here what you're asking. Are you asking "how to design a RPG ruleset/framework" or "how to design a campaign setting" or "how to design a module"? These are all very different questions.

Agreed with Mastikator. It works in video games, because you're generally playing those by yourself, and most things you do don't actually adjust the storyline all that much. People play TTRPGs because they don't have those same limitations.

At the same time, though, you can't plan around the end. You kinda can, but it has to be build open-ended from the very start.

Consider 5e's module of Curse of Strahd. It revolves around taking down a Vampire Wizard from his dark castle while inside of his own pocket dimension called Ravenloft. It's designed around letting the players choose what direction they want to go for specific relics or allies to assist them on their mission while Strahd and his minions taunt the players and get in their way, before heading towards Strahd's castle for one big showdown. In terms of the overall pacing, it's made almost exactly like Breath of the Wild's main campaign.

It works, because the only thing that the players are locked into is the beginning AND the end, with everything in the middle being a big mush of player choice. Strahd is usable as a villain in this way, because he finds ways of constantly inserting his relevance and personality on the players. Ganon works in BotW, because you can constantly see him and everyone constantly alludes to him. You can't just have a named BBEG, it needs to be one that the players are constantly reminded of, and something the players look forward to stopping.

[vasilidor]

it was a thought experiment of applying one design concept to another type of design. I needed a sanity check, it doesn't help that I do most of my design work at 12am-3am because I cannot sleep.

[Man_Over_Game]

it was a thought experiment of applying one design concept to another type of design. I needed a sanity check, it doesn't help that I do most of my design work at 12am-3am because I cannot sleep.

You were onto something. Starting from the top and working down is a good way to ensure that all of your players end on an equal note. But you also want them to start on equal ground, too. The point is that the only wrong way of doing it is by ignoring either side.

[Telok]

Two things I'm coming to embrace in game design are "do we care" and "that sounds like fun".

At certain points you look at something and ask if you care about it, if the game cares about it, and if the players care about it. Champions has a "problem" in that normal humans can fall off a 4 storey building and usually walk away with stun damage that goes away in an hour or so. It's not a problem because nobody really cares. Champions is about supers punching each other through walls and blowing up the scenery, the normals are going to splatter like the rest of the scenery at that power level. On the other hand D&D as a game really cares about individual creatures movement speeds, to the point that you need an entirely separate variant encounter system just to have a foot race that isn't predetermined before it starts. It really becomes a problem when someone is fighting, someone else is trying to race, and another person has some unusual movement mode to use. The game cares, the players care, and it isn't handled real well (better than some past editions but that's nearly "better than nothing" levels). So, problem.

Look at parts of a system and ask who cares and why. If you still care, do it well.

While testing your system try characters, actions, plans, and general stuff that sounds like it would be fun. Make that stuff fun. StarFinder has an example of what not to do. Fully automatic weapons fire in that game is pretty bad, we're already starting off in a hole here. There are several 4 armed species in the game available as PC choices. There is a feat that specifically allows a 4+ armed character to emulate full auto weapon fire by using 4 pistols at once. That manages to make full auto fire an even worse choice than before. It is almost, like 99% of the time, better to shoot people with regular full round double taps of a pistol than to use that feat.

[Duff]

I may have been confusing design styles. in some game types, the end is designed first and then they work backwards from there, I am referring to video games done by professional studios here. an example is the legend of zelda games: in order to reach ganon, you need to accomplish A, in order to accomplish A you need to do B Etc. Role playing games do need the variety of play levels, or benchmarks, in order to make comparisons now that I think of it.

Writing stories often starts with the ending and then work back. But when you design an RPG, you aren't telling the story, you're building the engine for someone else to tell the story. I don't know, but I suspect the engine on a computer game is built for beginer's mode and then added to as well (with an eye on the end goal at every point, but still, start with basic mode)

[KineticDiplomat]

As a contrast, I have found that the most consistently enjoyable RPGs solve the problem another way - they limit the power range to a much tighter band, usually starting at “quite reasonably competent for the power level of the system” and for all intents and purposes cap out “exceptionally capable, but still within mostly the same league.”

D&D (and several of the systems using it as inspiration) has the issue that you want to have “a pretty good second grade t-ball player” and “babe Ruth who is also magic” playing the on the same ball field and you need to model them using something that can be broken into say, less than a dozen numbers on a character sheet, using analog dice usually only have a range of 1-20, or 1-100 at the long end. I would say that theres not much likelihood any RPG with realistic constraints on design is going to manage that scale, no matter what order you design the classes in.

And that’s before D&Ds many other flaws kick in.

[vasilidor]

ok, thank you for your feedback. glad to know that i am not as nuts as i thought i may have been. i will keep the advice you have given me in mind.

[LibraryOgre]

"What do you want this game to do?"

Are you trying to make a game that will easily fit into multiple genres? Are you trying to specifically emulate a given setting? Do you want things to stay street level? Do you want to be gonzo, balls-to-the-wall, guns and spells? Do you want to be Gonzo, the oddest Muppet Baby in a setting of Muppet Babies?

Because moreso than the mechanical framework of the game, you have to know what you're trying to get that framework to do. You CAN use a screwdriver as a hammer, but if you're picking tools to take with you, it's a lot better to use a hammer.

[LordCdrMilitant]

Some people think that in building a role playing game one should start from what they think the first level characters should be able to do and go up from there. Have any of you thought about taking the opposite approach, starting from what you wanted the end game to look like and going backwards? like with the fighters of DnD, if the end goal is Thor (or something compatible) then after you get the basic math in place, build thor, if the end goal is conan, build a conan, and then work backwards to figure out the steps to get to thor or conan. you can then build your Merlins or Gandalfs to match them at the final endgame level.
thoughts?

Given that the vast majority of playtime is spent in the early and mid game, and by definition only a little bit of the game is spent in the endgame, you probably shouldn't put what the players are expected to play as at the top of the tree.

The top end of the mechanical character development should be the pinnacle of the character concept's ability and what the character concept aspires to be, not really what you're playing as, since you won't be playing as it for very long compared to the rest of the game.


This is my problem with Deathwatch and Dark Heresy. At characters creation, you're supposed to be veterans of your organizations... but your're about as effective as an IG Conscript, if even. The starting package for Adepta Sororitas in DH2e is particularly egregious: a catsuit and a knife as a order novitiate; you don't get your Power Armor and Bolter that comprise your basic kit until like halfway through the game at earliest.At character creation, there just isn't the ability to realize what your character concept's starting place is when they exit your organization. The Inquisition doesn't staff it's ranks with Whiteshields, Novitiates, Interns, Footpads, and Mall Cops.


The one time I've done this [develop my own RPG system], I took a holistic approach and considered what a defined character archetype should look like at the beginning of their career, the middle of their career, and the height of their career, and that lets you lay out what progression should look like.

If you start at the top and work down, you'll wind up with characters always being below and working towards their basic concept. Starting at the bottom and working gives you at least the ability to formulae what your character is supposed to be, though you may have coherency issues when then grow.

[vasilidor]

one of the things I am currently trying to do is create a classless pathfinder setup. where you spend experience to gain abilities. I currently need to edit what I have posted in the homebrew section, but the basis for it is there. I also have a campaign for shadowrun that I am building for a 3rd edition game.

[MoiMagnus]

Some people think that in building a role playing game one should start from what they think the first level characters should be able to do and go up from there. Have any of you thought about taking the opposite approach, starting from what you wanted the end game to look like and going backwards?

I'd say the following.

Step one: Have an ad hoc system that run and is fun for the "main power level". (So in D&D5, that would be levels 5-9). Your system don't need to have level up or any kind of formal progression yet. You probably have only one or two group of players testing the game for this phase.
Step two: Formalise what you have, and design the first levels that lead to your "main levels" in a progressive way, allowing new players to be added to testing.
Step three: Work on higher levels. You will be faced with the issue of scaling (some low level effects will become OP at high levels, while others will become useless), which will probably force you to rework everything you made before, except that now you have a good idea of how the game should be played.

In parallel to those three steps, you should be constantly alternating between removing entire parts of your RPGs that "are too complex for what they allow to do" and adding new mechanics that create more depth. [If you work with peoples creative enough, you won't be missing ideas and technical solutions. So don't be shy on putting most of them in the garbage bin if they don't work as well as intended].

Step four: Declare the RPG to be "acceptable state" and focus on all the details (like balance). Do not go back to redesigning major parts or you will be lost in a infinite loop.

[vasilidor]

thanks for the advice so far. I think my next step is going to be to revise what I have so far in the development and then try and find a way to test it.

[Jay R]

Ironically, the game that most directly applied this anti-D&D approach was Dungeons and Dragons. It was designed by two miniatures players, and upper level characters were expected to recruit a bunch of followers, clear out some wilderness, build a keep, and settle down to have miniatures battles with armies that attacked the keep.

We didn’t play it that way, because that was Gary’s plan for high-level characters, not ours. D&D instantly recruited a bunch of fantasy gamers who were more interested in their characters than in armies.

But there was a clear plan.

[Xuc Xac]

*puts on sunglasses and uses Morpheus voice*
What if I told you that you don't need levels at all? If you want to play Thor, don't figure out what abilities he needs at 20th level and make the players start at level 1. Just let them be Thor.

Thor doesn't start as a level 1 fighter and go up. He starts as Thor, super strong flying God of Thunder and that's where he ends up. All of his character development involved changes in personality, equipment, and appearance. He tones down his arrogance, learns to respect others, develops trust, loses his hammer, falls in love, gets his hammer back, loses it again, gets a haircut, replaces his lost hammer with an axe, resolves a lot of toxic family issues with his parents and siblings, gets his hammer back, puts on a lot of weight, gets depressed, and learns that the hammer was inside him all along and the real lightning was the friends he made along the way.

Superheroes (and a lot of other fictional characters) "level up" once in their origin story because "zero to hero" is a two step process and not a level 1-20 progression.

[vasilidor]

*puts on sunglasses and uses Morpheus voice*
What if I told you that you don't need levels at all? If you want to play Thor, don't figure out what abilities he needs at 20th level and make the players start at level 1. Just let them be Thor.

Thor doesn't start as a level 1 fighter and go up. He starts as Thor, super strong flying God of Thunder and that's where he ends up. All of his character development involved changes in personality, equipment, and appearance. He tones down his arrogance, learns to respect others, develops trust, loses his hammer, falls in love, gets his hammer back, loses it again, gets a haircut, replaces his lost hammer with an axe, resolves a lot of toxic family issues with his parents and siblings, gets his hammer back, puts on a lot of weight, gets depressed, and learns that the hammer was inside him all along and the real lightning was the friends he made along the way.

Superheroes (and a lot of other fictional characters) "level up" once in their origin story because "zero to hero" is a two step process and not a level 1-20 progression.

i want to be able to have it show a range of abilities.

[Composer99]

*puts on sunglasses and uses Morpheus voice*
What if I told you that you don't need levels at all? If you want to play Thor, don't figure out what abilities he needs at 20th level and make the players start at level 1. Just let them be Thor.

Thor doesn't start as a level 1 fighter and go up. He starts as Thor, super strong flying God of Thunder and that's where he ends up. All of his character development involved changes in personality, equipment, and appearance. He tones down his arrogance, learns to respect others, develops trust, loses his hammer, falls in love, gets his hammer back, loses it again, gets a haircut, replaces his lost hammer with an axe, resolves a lot of toxic family issues with his parents and siblings, gets his hammer back, puts on a lot of weight, gets depressed, and learns that the hammer was inside him all along and the real lightning was the friends he made along the way.

Superheroes (and a lot of other fictional characters) "level up" once in their origin story because "zero to hero" is a two step process and not a level 1-20 progression.

This really seems to be missing the point.

[Xuc Xac]

This really seems to be missing the point.

I don't think so. The OP seems to have the typical "fantasy heartbreaker" mentality of "I've only seen D&D and assume all games are D&D, but my new game will be so much better because it's... well, D&D... but one thing is slightly different!".

[NichG]

Iterative refinement work tends to be more successful than jumping off into totally difference spaces, though it's less fun and you learn less from doing it, so I don't think 'I'm going to make a fantasy heartbreaker tuned to my tastes and the tastes of my particular gaming community' is necessarily a mistake. Ignoring ideas from outside of the starting point that could be useful is an error, but if you look at those ideas and say 'that doesn't meet my design requirements' then that's completely fair.

[False God]

I'm more of the mindset that games should be more open-ended. So you can't really design from the end if there isn't one.