D&D Next: How About We Stop the Increasingly Granular Skill Rules?

I've been thinking a lot about what I want from the next edition of D&D lately.  Not just because of the recent announcement that there will be a new edition of the game, but that has contributed without a doubt.  Largely, I've been thinking about it because of the D&D Encounters group I play with every week.  I've been running encounters for over 2 years now, and I noticed a significant change in how the game played when Encounters adopted the "Essentials and Newest Book" only stance.

The game is smoother.  There's less analysis paralysis. The players role play more.  The players think of their characters in terms of personality more than as a list of powers.  The addition of themes has helped this, but so has the return to more archetypical character design and the clear articulation of the powers associated with a given thematic build.  Character builds in the post Essentials world have been more "thematically" oriented than the calculated and Hero System-esque min/max affairs I had been seeing before.  It's been a nice change, and it got me thinking how the addition of Feats and a scaling skill system have made D&D less creative.  In early D&D, if a character wanted to be a blacksmith, all the player had to say was that their character was a blacksmith.  If a player wanted a character to be a blacksmith, poet, wizard, tumbling acrobat, scientist, engineer, the DM might have told the player that was ridiculous...OR the DM might have said "sure, why not" it doesn't affect the mechanics of the game and might make for interesting stories.  In the post Hero/Gurps D&D era where skills are all articulated, if the player doesn't have a skill then that character is up Illithid Creek without a helmet.

Image by Jody Lindke

I'd like to see D&D Next return to an era where there is no scaling skill system and where characters can be good at a lot of things that are associated with their character's "core strengths" without needing to allocate skill points each level or without the ridiculous results of the 4e skill system where a 10th level character might have a higher DC to kick down the same door as a 1st level character.  If anything can be learned about how to implement a robust skill system -- inspired by the 3.x and 4e skill systems -- the Dragon*Age role playing game should be our guide.  The DCs for difficulty are fixed early.  They are a fixed point in space.  Difficult is difficult.  Challenging is Challenging.  An average task in that game can be achieved on an 11 or better on 3d6 -- an exactly average roll.  To that roll an ability bonus (+1 for average ability or +4 for amazing) and a "focus" bonus (if the character possesses a relevant focus) of +2 are added.  An average person, with a focus in a relevant area, would add +3 to the roll making the chance of success 8+ on 3d6 on an average difficulty task.  Rolling an 8+ on 3d6 is fairly easy -- with an 83% chance of success.

There's no need to create DCs that are impossible for 1st level characters, at least not if you want free-wheeling games where low level characters take huge risks, you can just make them supremely difficult.  All you need to do is make it so that the main difference is "trained" and "not trained."  Throw level out of the skill system -- except maybe to add some "focuses" that match a character concept as the character increases in level.

Let's just have a look at how the skill system in D&D has evolved to get a look at where we might want to go to mix the D&D legacy rules with what I think is a wonderfully innovative Dragon*Age skill system -- I might even recommend using Robin Laws'  Gumshoe system for mysteries, but that is an entirely different discussion.

In the White Box of D&D there are no skills.  Period.  There are abilities and they affect very little about game play except the rate which a character advances in level.  Eventually, with the addition of the Blackmoor and Greyhawk supplements, the ability scores have more mechanical effects, but there is still no skill system per se.

In Moldvay/Cook -- pre-Gazetteer/Compendium era -- edition of the rules added the "There's Always a Chance" rule on page B60 where you roll a d20 and compare it to a statistic (adding a -4 to +4 modifier to the roll based on difficulty) to account for all of the things a character could do outside of those abilities specifically granted by the character's class.  This rule might also be in the Holmes Basic, but I couldn't find it.  This is the root for what became the eventual early skill system of D&D.  The D&D Gazetteers, starting with Aaron Allston's Grand Duchy of Karameikos, gives characters a number of skills based on their level.  Success is determined by rolling equal to or less than an attribute on d20 -- modified for difficulty.  Players can spend an entire skill choice to gain +1 to an attribute for the purpose of a skill check.

The AD&D "non-weapon proficiency" system introduced in Oriental Adventures and implemented in both the Wilderness Survival Guide and the Dungeoneer's Survival Guide is very similar to the Allston system -- and pre-dates it.  This system comes after the broad AD&D "background skill" system in the Dungeon Masters Guide which gave characters background skills, but had no mechanical adjudication for those skills.

Until the 3rd edition, the Allston/Non-Weapon Proficiency system was the core skill system in D&D.  It always seemed a little wonky to me.  A character could use one choice to get a 55% chance to succeed at a skill (assuming an 11 stat), but had to spend an entire extra choice to get an additional 5%.  That never seemed quite right.  Additionally the difficulty of a task was entirely controlled by a character's full stat, rather than by the bonus, so an 18 in a stat meant a 90% chance of success for most tasks -- minus any modifiers.  It just seemed to make good characters too good.

The 3rd edition of the game learned from the 90% is too good example and tied skill checks to the stat bonuses -- which reflect the natural standard deviations of the attributes on a 3d6 curve -- and set them against difficulties.  A character with an 18 stat added +3 to a roll of a d20 and compared it to a difficulty -- a skill system similar to that of earlier DC based RPGs and wargames.  It works, but 3rd edition had scaling difficulties and opposed rolls and gave skill points every level which made it so characters had to make sure that they continued to spend precious skill points in ways that continued to be useful.  Finding a difficult clue at 10th level had a higher DC than finding a difficult clue at 1st level.  This needed to be so because the characters would have skill bonuses, but the higher DC also required the higher skill bonuses -- a kind of skill system moebius strip a never ending cycle of increase.

4th edition partly solved for the ever spending of skill points by having a system of "trained" and "not trained," but then muddied the mix with scaling DCs as the characters went up in level that was tied to ever increasing stats and magical bonuses.  Ironically, the 4e system is actually flat.  The increases are an illusion.  Yes a 10th level character can beat up a 1st level character, but a 10th level character has as hard a time solving a 10th level problem as a 1st level character has of solving a 1st level problem.  Amusingly, sometimes these problems might be the exact same thing.

I've always thought that scaling difficulties, skill points to demonstrate improvement toward scaling difficulties, and a system of granularly listed skills to detract rather than add from play.  In the early days of Mutants and Masterminds, Steve Kenson spent many a day on the forums trying to convince people that Batman would just have Super-Dex and Super-Intelligence along with "training" in those skills that are "trained only."  This would mean that he was good at all the skills, and that he didn't need to spend individual skill points.  Players who favored granular itemization of skills won out, and the M&M point buy system has never really recovered its balance since.  Thankfully it has guidelines based on power level for power effects so the fact that skill points and powers don't really matchup mechanically point for point doesn't matter too much, but it still affects that game and makes that game feel less "superheroic" and more GURPS-ish.

I've always preferred games where characters can have skills that were independent of the level of the character.  In Call of Cthulhu there are no character levels, just skill improvements through the use of skills.  In D&D, there are character levels but I'd love to see them separated from skill use.  There are tons of people who would be level 0 who would have tremendous skill at some science or artistic skill.  There is no need to create nerfed NPC classes -- as 3rd edition did -- or to create bizarre DC scaling tied to level -- as 4th edition does -- all that matters is "trained" and "untrained" with a possible "how well trained" and "do you practice" added for good measure.

I'd love to see D&D Next adopt a skill system with fixed DCs, like Dragon*Age, that allowed characters to take whatever knowledge/artist/artisan/profession skills they wanted without limit, and in which training in a class gave some small bonus.  For example, a thief might just be a better climber than a non-thief and have a slowly scaling bonus (say +1 to +4 at level 20) against that fixed DC.  There is no need to have a large section of the rule book filled with skills and their descriptions and submechanics.  Look at Dragon*Age, there is no set of three subskills for any of the focuses -- unlike 4e's Acrobatics or 3e's Tumble.  There is no need.

Keep it simple.  Stop becoming more granular and complex.  Sure, some players want D&D to have Advanced Squad Leader and GURPS levels of detail and granularity, but most want to sit around a table and have a good time.  Let's find some compromise between the non-weapon proficiency system and the 3.x system.