What would the perfect fantasy role-playing game look like? An awful lot of game designers seem to think it should feel somewhat like Dungeons & Dragons but use different mechanics. I suggest we should use as much of the tried-and-true mechanics of D&D as possible and only fix what’s actually broken.

Old-school D&D (TSR days) has good emphasis on exploration but:

• Lacks character advancement options for non-spellcasters.
• Lacks maneuvers to do in combat beyond “attack with my weapon”.
• Is too miserly with healing and death and–at low levels–hit points and spells.

Modern D&D has clean and consistent mechanics but has:

• Too many character-building options.
• Too much emphasis on combat and lack of support for exploration.
• Is too generous with healing, death saves, and spells.

Start Simple

Our perfect system should start with a clean, uncluttered rule set. Unfortunately, there isn’t a version of D&D like that. Fortunately, we can pluck clean mechanics out of 3e and use them to replace the clunky ones in AD&D. 3e did a great job of simplifying the mechanics of D&D… and then drown them in pedantic rules.

Part of keeping the system simple is avoiding a complex “action economy” that results in awkward phrasings like “I use my action to X, then I use my bonus action to Y.” These drag out combat, because they encourage players to treat each turn as a puzzle to find the optimal use of their actions. And thruout the system, using natural language and plain meanings of terms is important to acceptance and easy learning. This probably means limiting “actions” to one per round (except for a special feat or haste effect), altho a character should still be able to both move and attack, as in OD&D.

If you’re struggling to find a word to describe some mechanism, it’s probably not a mechanism that we naturally think of as a thing. Armor class is a natural mechanism, because we inherently know that some armor is more effective than others, and it originally went down in D&D because the military considers “class 1” to be the best, not the worst.

Rules should be presented simply, in a clean layout. Long-winded explanations are even less helpful than too-short ones. The cleanest layouts I’ve seen are in Shadowdark and Old-School Essentials, so those will be our guides. (Well, I’d kill for my game to look as gorgeous as Pathfinder, Song of Swords, or Zweihander, but I’m wary about emulating games that need 1100 words to explain stealth or 400 words to explain cooking.)

The system should strive to avoid bookkeeping as much as possible. There should be no tracking of arrows, torches, rations, etc., altho a simple check would be okay. And damage to armor, hit location, ongoing wound bleeding and infection, etc. likewise need to be left out or reduced to simple checks. Even calculating and tracking experience points should be reduced to achievement awards. (Finish an adventure and get XP based on how tough it was and how much you succeeded.)

Abilities

Abilities should be more-or-less the same as those in D&D. There’s nothing particularly broken about strength, dexterity, constitution, intelligence, wisdom, and charisma. Traditionally, charisma got treated with disdain, because it wasn’t as fully utilized by the rules as the others, but there are simple ways to fix that just by making it more important (reaction rolls, resistance to mind control spells, etc.). However, perhaps wisdom and charisma could be combined into a single sense ability that measures the character’s “horse sense and people sense”.

It might not be necessary to have a 3 to 18 index number for abilities. If we use standard modifiers, then it’s the modifiers that matter. Your strength of 2 gives you +2 to damage, etc. However, if we want to use an opponent’s ability as a difficulty target for opposed checks or your strength score as the number of items you can carry (which I do), we’ll need to keep the index number. Otherwise, we’ll be using the 2 as an index on a table for how many items you can carry, so we haven’t gained anything.

Then we eliminate the dice-rolling to create a character. You want to play a wizard? You’re a wizard! Assign a standard array of numbers to your various abilities, with perhaps one point to put anywhere. Some will use it to top off their brilliant mind; others will use it to improve some other ability enough to gain a bonus there.

Abilities should have broad application. We should make sure that it’s reasonable to play an intelligent fighter. You wouldn’t want intelligence as your primary ability, but there should be some advantage to being smart. (But that doesn’t need to be true of everything, like a wizard and strength.) However, rather than merely adding your intelligence modifier to your attack roll or something, the advantage should be that you can learn more feats/exploits or learn ones that are only open to characters with higher intelligence.

Add Character Options

All of OD&D and AD&D were lacking options for non-spellcasters. These were added to modern D&D, but modern D&D has gone bananas with them, creating literally hundreds of sub-classes, features, and feats across various editions and endless supplements.

Class Features & Feats

Class features and feats will need to be rethought from the beginning and kept very limited, as I did with my fix for 2e. We have the luxury of not needing to sell scads of supplemental books to players, so we don’t need to entice them with endless options.

Something that modern players like that’s underdone even in 5.5e and nonexistent in earlier editions is character backgrounds. This should be a feature of the classes themselves. Ideally, it would be integrated with a campaign setting, but a good campaign setting could overwhelm the rule system.

I now think that the perfect rule system does need to have an integrated setting, but that setting should be fairly generic or it could doom the system. Dolmenwood and Land of Eem are brilliant and gorgeous, but I can’t stand the fairytale and cartoonish aspects.

Skills

Character skills are also very useful, but probably are a bit overdone in modern systems. Characters should get one or two skills from their background and a few skills from their class. For example, if you’re a fighter-type with a rural background, you could get a nature-based skill, but if you have an urban background, you get a more city-appropriate skill.

As they rise in level, characters should get better at all their skills. However, unlike 5e’s proficiency bonus, weapon skills and spellcasting skills should be kept separate from ordinary skills for clarity and to protect their niches. (In 5e, a wizard is just as good with a dagger as a fighter or thief of the same level, because they’re all proficient with the dagger. It’s feats that make the difference.)

You should be able to specialize, but there shouldn’t be fiddly allocations to various skills as you rise but rather certain feats/features that let you give an extra bonus to one skill over another. So, if you want to specialize in something, you take the feat that gives you a bonus to that one thing (and probably lets you do some trick).

Classes

Most D&D-like systems reject race-as-class, so we’ll borrow from AD&D 1e and 2e for our base character ancestries and classes. But to eliminate multi-classing, there should simply be additional classes that in effect are hybrids of other classes.

I don’t like the idea of subclasses; it’s simpler to just let players have a couple of meaningful choices that differentiate one type of monk from another. This avoids the problem of players complaining that the subclass they like the flavor of is weaker than the others.

1. Warrior (melee or archery fighter with some nature or urban skill)
2. Knight (heavy fighter; gets a squire and horse)
3. Paladin (knight/cleric)
4. Ranger (light fighter with nature skills)
5. Bard (fighter/wizard)
6. Wizard (memorizes spells)
7. Sorcerer (casts at will but fewer spells)
8. Cleric (memorizes spells; worships gods/saints)
9. Druid (casts at will; worships nature or sun & moon)
10. Monk (cleric/fighter with hand-to-hand skills)
11. Witch (druid/wizard)
12. Rogue (burglar; chooses more fighting or more sneaking skills)
13. Mountebank (rogue/wizard)
14. Swashbuckler (rogue/fighter with urban skills)

It may be useful to put the classes into categories so that, for example, certain magic items can be reserved for martial-types. Rather than have categories like 2e with classes in them, I would merely label each class with its category or categories. So swashbuckler is labeled “martial & rogue”, for example and can therefore use magic items, feats, and skills restricted to martial classes and rogue classes.

• Martial
• Arcane
• Divine
• Adroit

Each class should be presented on one page with a piece of character art. That limits the amount of folderol that can be added to a class. Class features/feats can be defined elsewhere, since many will be shared between classes, just as spells are defined in their own section.

Advancement

My system of advancement uses small-scale experience points and renown points that accumulate as you spend treasure. This is both a simple and flexible, so advancement and rising social rank are simple and easy to track.

It probably makes the most sense to say that characters should advance in level after every adventure; this means adventures need to be fairly discreet, with a clear objective that, when fulfilled, signals the end of the adventure. This is a traditional concept in old-school adventuring, but modern games often pile one challenge on another in a jumble of “adventuring” that often leaves players calculating that they advanced 20 levels in only a few weeks on the in-game calendar.

Each class should get to choose options (either spells or feats) as they advance.

Straighten out Spellcasting

An obvious thing to do (but which few D&D-like systems do) is to fix the fact that D&D spell levels don’t match the caster’s class level. It might seem simple to say a 4th-level wizard can learn 4th-level spells, but that means players would have to juggle several lists of spells even at mid-level, and the tables for learning and memorizing spells would be huge–more than 15 columns wide to go to 15th level.

Instead, I think the thing to do is to consolidate spell levels and call them something else:

• Very simple spells are cantrips (arcane) or blessings (divine) along the lines of Prestidigitation and Thaumaturgy (but including Read Magic and Detect Magic). Use of these is free and unlimited.
• 1st- and 2nd-level spells are now enchantments and graces or 1st order, available from 1st level.
• 3rd- and 4th-level spells are now dweomers and intercessions or 2nd order, available at 5th level.
• 5th- and 6th-level spells are now glamours and miracles or 3rd order, available at 9th level.
• 7th-level spells are now rituals or 4th order that take longer to cast and work best with multiple casters, coming available at 13th level.
• 8th- and 9th-level spells are now legendary magic only for NPCs, monsters, and items.

At each tier, the old higher-level spells would cost more to cast or would be toned down to be equal in power to the lower-level spells. Since 2nd order spells aren’t available until 5th level, we can keep players from feeling they’re not getting anywhere from 1st to 4th level by having spells get more powerful based on the character’s level.

For example, instead of 2nd-level Invisibility, 3rd-level Invisibility 10′ Radius, and 4th-level Improved Invisibility, there would be one enchantment called Invisibility. Perhaps at 1st-level the caster can become invisible while staying still and concentrating; at 2nd-level he or she can move but must still concentrate; at 3rd level, the spell ends when he or she attacks or becomes unconscious; at 4th level, it also affects one other creature that remains in contact with the caster; at 5th level, it affects all creatures in a 10-foot radius. These should probably have escalating costs but don’t require the caster to learn a new spell.

Arcane Spellcasting

For arcane spellcasters, I like the idea of wizards being able to cast any spell they know but having to make a spellcheck each time. This check would be pretty easy right from the beginning, but you could never cast the same spell very many times a day. (The limit would probably be something like a modern D&D proficiency bonus.) If you fail the spellcheck, you suffer a minor wild magic effect and can’t cast that spell again until you’ve had a good night’s rest.

Material components have never been popular. Except for really expensive components, these don’t help class balance much and aren’t really fun to role-play finding or shopping for. I suggested elsewhere that spellcasting should require a material component that’s easy enough to get but costs good money; I call it “felstone”.

Finding felstone powder as treasure would be a nice alternative to gold and magic items. So, yes, you get access to Sleep right out the gate as an enchantment, but can you afford to use 20 gp worth of felstone powder to cast it when you could cast Magic Missile twice? But later on, spells may cost you more than money: you may have to expend hit points to cast them.

Arcane casters would acquire spells by learning them from spellbooks they recover as treasure.

Divine & Nature Spellcasting

Clerics and druids and such can pray for any spell of their class’s list that they know of (that is, that is in the rule books) but must make a check to see if their god grants it. Failure means having to choose a different spell to learn, but you can try again after gaining a level.

For diving and nature casters, I like the idea that they can cast spells basically just like clerics and druids in old-school D&D: choose your spells for the day and cast them without making a check.

Make Combat Dangerous (a bit)

Initiative

In a system couched in the simplicity of OD&D but willing to use the embellishments of modern D&D, initiative can be simple: each side rolls 1d12, and the winning side chooses who among them goes first. Then it alternates. But, there can be a feats that give a character +1 on their group’s initiative or that allow the character to sometimes act first in a round. Re-roll initiative only after an area-effect spell or similar disruption.

Combat should use the d20 vs armor class method of D&D, but there should be some simple tactical combat and varying effects for weapons or at least some tactical exploits. It’s important that the combat system be as quick and simple as possible to keep the game moving. The biggest complaint with modern systems is how slow combat is–which is weird, because modern D&D is basically a combat simulator, so just, like, don’t be like that, D&D.

Fun Combat

Combat should be fun and interesting, and hopefully cinematic and perhaps a bit tactical. It should therefore provide things the heroes can do and objects to interact with during combat (such as a table of useful things commonly found in a dungeon chamber or building room). Using these things should either be in addition to the character’s usual attack or more effective than the character’s usual attack–or else players will forego using them in favor of their usual attack. This could be part of a character’s feats, or it might be a common combat tactic (available to anyone or at least anyone of a martial class) that allows characters to give themselves advantage or their opponent disadvantage or to foil an attack or even to disable an opponent entirely.

Plenty of spells disable opponents, why not let a fighter drop a chandelier on a couple of opponents like you see in movies? Of course, the GM has to provide a chandelier for that purpose. The perfect game system would somehow ensure that that sort of thing is sometimes available (or allow the player to generate it). Perhaps the character needs to spend a point of luck and then roll to see if they can perform a “daring feat”, then the GM is required to describe some interactive item that allows an effect appropriate to the hero’s level.

Power Balance

A big problem in modern D&D is that heroes are given so many and such a variety of powers that they can often one-shot the villain the moment they meet. Video games avoid this by ensuring the heroes never have such power. The perfect TTRPG would do the same: don’t have spells that can take out any enemy if that enemy misses a saving throw; and don’t allow class features or magic items to stack damage to absurd levels.

In modern D&D, this is the name of the game, because WotC made their fortune with Magic: the Gathering, where deck-building for sick combos is what it’s all about; but MtG is carefully balanced to avoid broken combos; D&D isn’t.

Healing

I’ve already laid out ideas for injury, falling damage, and death. The system should be more lethal than 5e’s overly generous rules but less brutal than old-school D&D. Compare old-school with modern:

• Old School: Heal at 1 hp per day. Dead at 0 hp. A cleric’s healing requires touch and can happen over and over in battle, but that’s rare because of the scarcity of magical healing.
• Modern: Healing thruout the day with short rests. Fully healed with overnight long rest. Unconscious at 0 hp but make three death saves before three failures and you’ll live. Plus, a cleric can heal you from afar, and it’s common, so you can get knocked down and pop back up over and over in battle.

Doesn’t it seem like there ought to be a happy medium in there?

• New: Heal 2 hp per level with overnight rest. Below 10 hp, you have disadvantage on attacks and strength and dexterity checks. Unconscious at 0 hp but make a death save, and you’ll live. A cleric’s healing requires touch and is somewhat common but will usually leave you with less than 10 hp and thus at disadvantage.

Creatures

Creatures should have a specified level that makes them, in the recommended numbers, a good match for a party of characters of the same level.

The monster section should start with:

• Treasure type tables, with different treasures for each type
• Reaction type tables, with different reactions for each type

Then each creature should get a half page–one column–that contains:

• A picture of the creature
• The stat block (with reaction type and lair/habitat type)
• A description that mentions special abilities
• A treasure type with an entry for what parts are worth harvesting
• A table listing (for boss-type monsters) why the creature might be the villain of an adventure or (for animals and such) what it’s doing when encountered

At the back, there should be a special abilities section that explains what the various resistances, vulnerabilities, magic abilities, attacks, and defenses do, rather than shoehorning that into each monster description and repeating it for each monster that has, for example, a grab-and-crush attack or enthralling ability.

Monsters should come with an entry for the habitat where they live and the type of lair they dwell in–a cave complex? a ruin? a treehouse? There should be a 2-page spread of burrows, warrens, caves, caverns, tunnels, mines, encampments, and such. There should also be a table for each creature type that provides an idea of what that creature is doing when randomly encountered and how they might react to adventurers.

Integrated Systems

It’s important that our perfect system feature a working gold economy, monster harvesting system, and crafting system. These need to work as incentives to follow the rules and not as fiat rules, so that they’re gamified to be fun to work with. That allows simple rules to generate complex results that are fun, surprising, and interesting. Players love having an affect on the game world.

Ordinary purchases, lifestyle costs, leveling up costs, and donations to the faith should be used create a working gold economy that keeps heroes from accumulating gold with no end in sight. These can feed a character’s renown, unlocking NPC contacts of higher status. (Or purchases could just add to experience points directly.)

There should also be a simple travel system with a calendar and weather and advice on mapping your realm and adventures.

Example

You go fight goblins to rescue a merchant trader and recover his wagon of trade goods. He thanks you by giving you the trade goods, which you sell for a nice bag of gold, which is important because goblins don’t have any other treasure.

Next, you seek out and battle a troll to his doom and lug his carcass back to the city to claim the bounty and then sell the corpse to an apothecary. The apothecary pays you partly in gold and partly in potions of healing, because troll blood is a primary ingredient in potions of healing. You promise to be on the lookout for a hellhound, whose heart can be brewed into a potion of fire breath or a potion of fire resistance.

So you take your loot and buy a good horse and some fine clothes. Your purse is lighter, but now you look successful, so your renown goes up, and the local lord hears about you and requests your aid with a problem he’s been having.

You also donate a chunk of your treasure to the temple and pray to your god for a fine enchanted sword. This doesn’t improve your renown, but lo and behold! That night, you have a vision of just such a sword in the hands a warrior of the Foul Wastes. Now wither, adventurer? Seek out hellhounds, the lord’s problem, or the enchanted sword?