At Gen Con in 1999 Wizards of the Coast revealed plans for a third edition of Dungeons & Dragons, scheduled for release in 2000. The upcoming release would preserve the “essential look and feel” of D&D while introducing intuitive changes like a unified d20 mechanic where high rolls meant success. Free t-shirts teased the upcoming changes.

During that 1999 announcement, I thought back to 1989 and how AD&D’s second edition rules addressed few of the game’s weaknesses. I remembered the disappointment that I had felt. I wrote, “Second edition had some virtues: It consolidated and organized the rules; it featured leaner, more readable writing; and it dumped some of the clumsier bits like giving distances in inches and dividing rounds into segments. But mostly the second edition just repackaged the hodgepodge of rules accumulated since the first edition. Almost all the game’s inconsistencies and clunky bits remained.”

Since 1974, the roleplaying games that followed D&D featured countless innovations that improved play, so in 1989, D&D’s rules looked as dated as a Model T. The 1999 announcement previewed something different: a third edition that redesigned the game from its foundation. That promise fired my excitement. I posted a wish list of the changes to D&D that I hoped to see. Now 25 years later, how many of my wishes came true? My original post appears in italics.

1. Unify the game mechanics.

Want to find out if you hit? Roll high on a d20. Want to jump a chasm? Roll low. Want to use a skill? Roll high. Want to use a thief skill? Roll low on a percentile. Listening at a door? Who can remember? Look it up. The third edition fails if it fails to make sense of this.

Third edition delivered the mechanic where players add an ability modifier and other adjustments to a d20 roll, and then compare the result against an AC or difficulty class to determine success. This mechanic remains at the heart of the game and ranks as the single biggest refinement over the game’s 50-year history.

Wish granted.

2. Smooth the power curve for magic user advancement.

Playing a 1st-level mage means spending your adventure waiting for your big moment when you can cast your single, feeble spell. The situation improves little until level 5 or so. Some say that this balances the class against the extreme power of high-level spellcasters. I say it’s bad design. Low-level casters should have more than one spell. If this unbalances the class, then reduce the power of high-level casters.

While third edition kept low-level magic users in the game by granting them bonus spells based on intelligence, wizards still surged in power at higher levels, eventually overshadowing other characters. Fourth edition D&D solved the problem, matching the power of different classes by giving every class the same allotment of daily and encounter projects. The uniformity left some gamers unhappy, and that included the designers. “I’d never wanted to use the exact same power structure for the wizard as every other class,” said lead designer Rob Heinsoo, “But we ran out of time, and had to use smaller variations to express class differences than I had originally expected.”

Fifth edition reverted to a magic system that resembled the Vancian magic of older editions, but the game made three changes that check the power of high-level wizards and other spellcasters.

• Many spells now require their caster to maintain concentration to keep their magic going. Critically, a spellcaster can only concentrate on one spell at a time. Now rather than layering haste, invisibility, fly, blur, resistance, and a few others, a caster must pick one.
• Spells cast by higher-level casters no longer increase in power simply based on the caster’s level. Now, for example, to increase the damage of a fireball, a wizard needs to spend a higher-level spell slot.
• Compared to earlier editions, spellcasters gain fewer spells per day, and much fewer high-level spells. In the third edition, a 20th-level wizard could cast four spells of every level including 8th and 9th level. In fifth edition, a 20th-level wizard gets 14 fewer spells per day, and only one 8th- and one 9th-level spell. Sure, the fifth-edition caster enjoys more flexibility, but far fewer slots.

The damaging cantrips introduced in fourth edition also let low-level wizards do something magical every turn, rather than spending most combats flinging darts like exactly zero magic users in our fantasy imagination.

Wish granted.

Related: How fifth edition keeps familiar spells and a Vancian feel without breaking D&D

3. Drop THAC0.

To the folks boasting that they have no problem with THAC0: Don’t spend too much time lauding your math skills. We all can do the addition and subtraction in our heads, but the system remains a clumsy and counter-intuitive vestige of D&D’s wargame roots.

In drafts of the original D&D game, a 1st-level fighting man could figure the lowest d20 roll needed to hit by subtracting the target’s AC from 20. New fighters needed a 20 “to hit armor class zero,” and with each level they gained, their THAC0 improved by 1. The to-hit tables that reached print lacked the same simplicity, but descending armor classes remained. So when the second edition dropped “attack matrix” tables in favor of THAC0 numbers, the game returned to its roots. Still, THAC0 used negative numbers for especially tough ACs, made mental math more difficult, and forced a counter-intuitive concept on players.

Second-edition designer Steve Winter said, “There were all kinds of changes that we would have made if we had been given a free hand to make them—an awful lot of what ultimately happened in the third edition. We heard so many times, ‘Why did you keep armor classes going down instead of going up?’ People somehow thought that that idea had never occurred to us. We had tons of ideas that we would have loved to do, but we still had a fairly narrow mandate that whatever was in print should still be largely compatible with the second edition.”

Third edition dropped THAC0 and descending armor classes in favor of a target to-hit number and to-hit bonuses players added to their roll. Suddenly, the game became more consistent and intuitive.

Wish granted.

Related: The Dungeons & Dragons Books that Secretly Previewed Each New Edition

4. Give complete rules for magic schools and specialization.

The second edition mentions magic users who specialize, but the core rules fail to give rules for magical specialists other than the example of illusionists.

Third edition included rules for wizards specializing in each school of magic. The 2014 version of fifth edition built on the idea by giving specialists abilities specific to their school.

Wish granted.

5. Give rules for different clerics for different deities.

The second edition mentions variant priests for other deities, but the core rules fail to give rules for priests other than clerics. Why shouldn’t a priest of a god of the hunt use a bow? Why shouldn’t the priest of a thunder god be able to cast a lightning bolt? Why should the rules force the DM to answer these questions?

In the third edition, clerics chose two domains such as the air and war domains suitable for a god of thunder. Domains brought extra powers and domain spells like chain lightning. By fifth edition, cleric domains turned into full subclasses with accompanying abilities. Alas, gods of the hunt still lack support.

Wish granted.

6. Make spell statistics consistent.

Playing spells in AD&D either demands that you memorize the details of each spell or that you waste game time thumbing through the PHB. Why can’t most ranged spells have consistent ranges? Why can’t most area-of-effect spells affect the same area? Similar spells should be grouped and should share similar numbers for range, duration, area of effect and so on.

The fourth edition of D&D granted this wish, but I also learned that part of D&D’s charm came from spells like polymorph, phantasmal force, and wish that no cautious game designer would pen. Unlike earlier editions, fourth edition provided spells with effects that confined their effects to grid and to a tightly defined set of conditions. Often spells no longer worked outside of combat. Even fireballs became square.

These changes simplified the dungeon master’s role by eliminating judgment calls over debates like whether a wizard could toss an earthworm above their foes, polymorph it into a whale, and squash everyone flat. The fourth-edition designers aimed to make the dungeon master’s role easy—something a computer could handle. But this approach discouraged the sort of ingenious or outrageous actions that create unforgettable moments. D&D plays best when it exploits a dungeon master’s ability to handle spells, actions, and surprises that no computer game allows.

Although fifth edition returned to spells that demand a DM’s judgment, the spells feature more careful and uniform descriptions than in the game’s early editions. Credit this improvement to a card game.

When the designers of Magic: The Gathering faced the problem of bringing order to countless cards, they used templated text: they described similar game rules with consistent wording imposed by fill-in-the-blank templates. Today, the patterns of templated text appear throughout modern D&D’s rules.

Early in the development of third-edition D&D, Wizards of the Coast purchased TSR. Skaff Elias had served as a designer on several early Magic sets and ranked as Senior Vice President of Research and Development. Skaff felt that the upcoming D&D edition could fix “sloppiness in the rules” by using templated text. Skaff and Wizard’s CEO Peter Adkison told the D&D design team to switch the spell descriptions to templated text, but the team kept resisting his directives.

Eventually, the D&D team readied the release of a playtest document that still lacked templated text. They claimed rewriting all the spell descriptions according to formula would prove impossible because hundreds of spells would need templating in 48 hours to meet their delivery deadline. Nonetheless, Adkison and Skaff took the challenge themselves, working through the night to rewrite the spells and meet the deadline. Even after that heroic effort, the rules document that reached playtesters lacked the templated descriptions from the CEO and the Design VP. The design team had simply ignored their bosses’ hard work.

The failure infuriated Adkison. He lifted Jonathan Tweet to the head of the third-edition team. Designer Monte Cook remembers Adkison’s new directive: “If Jonathan says something it’s as though I said it.” Unlike the TSR veterans on the rest of the team, Tweet had started his career by designing the indie roleplaying game Ars Magica and the experimental Over the Edge. As a member of the D&D team, he convinced the team to adopt some of the more daring changes in the new edition.

7. Make spell power consistent.

Spells of the same level suffer from wildly different power in game play. This means that each spell level includes a few powerful spells that see constant play and many weaker spells that are forgotten. Worse, spells like stoneflesh pose game balance problems.

As long as D&D adds spells, the game will fall short of making their power consistent. The 2024 version of the fifth-edition rules makes changes to nearly 200 spells from 2014. Some changes affect power levels by changing their damage or healing dice, by adding concentration, or even by substituting a different spell for an old name. The changes bring the power level of spells closer to other spells at the same level. Still, silvery barbs remains unchanged even though players widely regard it as broken. And the completely rewritten conjure minor elementals spell joins the list of broken spells often nerfed by house rules.

Related: How new changes created the 4 most annoying spells in Dungeons & Dragons and The 3 Most Annoying High-Level Spells in D&D

8. Shorten one-minute rounds.

Who besides Gary Gygax accepted minute-long rounds as plausible? I recently reread the second edition’s rationalization of why someone can accomplish so little in a minute, but found the explanation as silly as ever. While the length of time simulated by a round makes little difference on actual play, I still prefer some sense underlying the mechanics.

Original D&D specified that a 10-minute turn consisted of 10 combat rounds, making rounds one minute long. Nonetheless, the D&D Basic Set (1977) specified 10-second combat rounds. Editor J. Eric Holmes based the set’s shorter round on house rules like Warlock and The Perrin Conventions common among his circle of California gamers. The West Coast community tended to rely on combat experience earned in fights staged by the Society for Creative Anachronism. To them—and to most gamers—10 seconds seemed like a reasonable time for an exchange of blows. Meanwhile, Gygax stood loyal to D&D’s wargaming roots and favored the 1-minute rounds that let D&D fights mix with Chainmail miniature battles where armies of combatants needed minutes to mount attacks. In the AD&D Dungeon Master’s Guide, Gygax spends a quarter page defending 1-minute rounds against critics who argued that in a minute real archers could fire a dozen arrows and real fencers could score several hits. The second-edition Player’s Handbook dutifully spends four paragraphs justifying the 1-minute round by explaining how a character could easily need a full minute to drink a potion. So much for drinking a potion as a bonus action! When the third edition switched to 6-second rounds, few mourned the 1-minute round.

Wish granted.

Related: How D&D Got an Initiative System Rooted in California House Rules

9. Simplify experience.

Totaling experience shouldn’t require a spreadsheet. Why bother having exacting formulas for combat experience when non-combat experience remains largely up to the DM’s judgment? Newer games use extremely abstract experience systems because calculating experience adds no fun to a game.

Unlike the original D&D and AD&D rules, second edition avoids setting a standard experience point (XP) system for the game. Instead, the Dungeon Master’s Guide reads like a draft presenting two methods: Half the text suggests that DMs improvise awards based on how well players foster fun, show ingenuity, and achieve story goals. Half suggests DMs award XP for beating monsters. The book cautions against awarding XP for treasure. “Overuse of this option can increase the tendency to give out too much treasure.” In the end, the text offers little guidance. “The DM must learn everything he knows about experience points from running game sessions.” Obviously, when I griped about point totals that required a spreadsheet as if that were the only option, I forgot the current edition’s indecisive XP rules.

In practice, the best experience system matches the style of play that gamers want.

For games centered on looting dungeons, D&D’s original XP system worked brilliantly by rewarding players for claiming treasure and escaping the dungeon with as much as they could carry.

For games focused on stories that resemble an epic fantasy tale, awarding experience for overcoming obstacles works better. This builds on the essentials of storytelling where characters face obstacles to reach their goals. Most second-edition players chose to award XP for defeating monsters. This method based XP awards on the game’s most frequent and quantifiable obstacles, namely monsters, so the method avoided relying on arbitrary decisions. However, it fails to account for games without much combat or where smart players reach goals without a fight.

Fifth edition offers an XP system based on beating monsters while suggesting awards for overcoming non-combat challenges.

As a flexible alternative to grant XP for overcoming obstacles, DMs award XP to parties when the reach their goals—the Dungeon Master’s Guide calls these goals “milestones.” Many DMs skip the point totals by simply granting levels for meeting goals.

For gamers who stick to points, the totals still add up to thousands, high enough numbers to merit a spreadsheet. Those high numbers remain as a vestige of the original D&D game where one gold piece brought one experience point.

Related: XP Versus Milestone Advancement—At Least We Can All Agree That Awarding XP Just for Combat Is Terrible

10. Publish fewer rule books.

All the optional rule books make the second edition too sprawling, too heavy, too expensive, and too intimidating. Plus, if you use all the options, then you hardly play the same game.

Ever since TSR Hobbies learned that D&D fans would buy books of campaign lore and character options, the D&D team faced a temptation to publish too many D&D books. As sales sagged in the second-edition era, TSR tried to bolster revenue with more and more books. When the third edition’s sales boomed, Wizards of the Coast tried to capitalize by publishing more books.

But even the most eager gamers can only spend or play so much, so inevitably those additional books meet indifferent players and languish on store shelves. Rows of D&D books in game stores tend to intimidate potential new players who wonder how much they need to buy to start playing and how much they need to master to join existing players. The editions of popular roleplaying games typically end with game store shelves glutted with product. By fourth edition, the D&D team knew this pattern, but they still published hardcovers almost monthly.

Fifth edition stands as the first to stick to a slower release schedule that may avoid the risk of overwhelming fans with products. Until 2018, Wizards of the Coast only published three books a year. Still, the number of D&D hardcovers released annually has climbed to four, then five. Even as a D&D enthusiast, I no longer find time to play or even read every release. Count the wish as granted, but perhaps only through 2021 when five books a year started reaching stores.

Related: How the end of lonely fun leads to today’s trickle of D&D books

11. Make healing spells consistent with the hit point concept.

I can accept the rational behind the gain in hit points a character receives with each level: A 10th-level fighter cannot actually survive 10 times more injury than he could at 1st level. Instead, his combat skills enable him minimize the damage from a stab that once would have run him through, so instead of a fatal wound, the 10th-level fighter suffers a minor cut.

Fine. But why does a cure light wounds spell have a negligible effect on the 10th-level fighter when the same spell can bring a 1st-level fighter from death’s door to perfect health?

I admit this issue is more of a peeve than a serious problem with the system, but a bit of consistency would make the rationale for hit point advancement much easier to accept.

Fourth edition granted this wish. In the fourth edition, rested characters gained a limited number of healing surges, and then healing magic let characters trade surges for healing. Because the number of points healed by a surge grew with a character’s hit point total, magic like healing potions cured wounds in the same proportion at every level. Fourth edition’s treatment of hit points and healing ranks as one of the edition’s best innovations, but it needed the edition’s hit point math to work, so fifth edition returned a scheme where a 1st-level cure light spell offers minimal healing to a high-level character.

Related: D&D’s Designers Can’t Decide Whether Characters Must Rest for Hit Points and Healing, but You Can Choose