Modern Dungeons & Dragons includes both skills and character classes, but in the early days of the roleplaying hobby, gamers often saw skills and classes as incompatible. Some gamers touted skills as the innovation that freed roleplaying games from character classes. Three years after D&D reached hobby shops, new games like Traveller and RuneQuest eliminated classes in favor of skill systems. Advertisements for RuneQuest in The Dragon trumpeted, “No Artificial Character Classes!!” Such games eliminated the unrealistic class restrictions that prevented, say, a fighter from learning to climb walls or from mastering a spell. “Mages can wear armor and use blades.” The ad credits RuneQuest to designer “Steve Perrin and friends.” Remember that name, because Perrin returns to this tale later.

D&D co-creator Gary Gygax favored classes because they resonated with the fantasy archetypes everyone knew. He warned, “If characters are not kept distinct, they will soon merge into one super-character.” He had a point. Skill-based games gave every character the ability to improve the same common adventuring skills, leading to a certain sameness among adventurers.

Classes let characters make distinct contributions to a group’s success. In a 1984 interview in DRACHE magazine, Gygax said, “The D&D game is based on the theory that there is so much to know and to do that nobody can do everything on his own. The team aspect is important. Each player has to use his strengths at the right place. Otherwise, the group won´t be able to survive.”

As long as Gygax controlled D&D’s development, he kept skills out of the game. His Unearthed Arcana (1985) added weapon proficiencies as a sort of weapon skill, but their narrow scope kept the sharp lines between classes.

Still, TSR designer Dave “Zeb” Cook saw a need for character development beyond class. “One of the things dreadfully lacking from Advanced Dungeons & Dragons was any sense that your character had a real life beyond class skills.” When Cook wrote Oriental Adventures (1985), he brought a taste of skills to D&D in non-weapon proficiencies—skills without the name. These new proficiencies never overlapped with class abilities. Characters gained skills such as calligraphy, etiquette, animal handling, and bowyer. Non-weapon proficiencies “gave players a way to create a more culturally-informed background for their character.”

Checks finally reached AD&D in the Dungeoneer’s Survival Guide (1986). Although this book’s title suggests a focus on player strategy, this guide brought key rules innovations to AD&D. Here, the non-weapon proficiencies from Oriental Adventures became options in the primary game. When players used non-weapon proficiencies, they made proficiency checks to determine the outcome. These checks filled the place of ability checks. The new system of featured all the ingredients of a modern skill system, although class features still covered most of the actions characters attempted during an adventure, so thieves still rolled on their private tables to climb walls and move silently.

In a convention appearance, Dave “Zeb” Cook and fellow designer Steve Winter talked about how these first-edition books led to a second edition. “Oriental Adventures was the big tipping point because Zeb Cook put a lot of really cool stuff in OA,” Winter said. “We felt like, wow it would be great if this was actually part of the core game, but it’s not.”

“Because of the way we had to treat those books, you couldn’t actually consider them canon when you were writing product or doing modules,” Cook explained. “You always had to assume that players only had the Dungeon Master’s Guide and the Player’s Handbook.”

Even after Gygax left TSR in 1985, designers like Cook and Winter lacked the clout to make sweeping changes to the company’s flagship game. TSR management insisted that second edition AD&D remain broadly compatible with the original. The Player’s Handbook (1989) included non-weapon proficiencies as an optional rule. Ability checks entered the core game, but languished in the glossary. Nonetheless, these additions inched AD&D closer to matching the ability checks and skills in other role-playing games.

But TSR sold two D&D games, an advanced version that got more scrutiny from management, and a basic version that offered more freedom to designers. By 1988, RuneQuest designer and freelancer Steve Perrin was gaining assignments writing D&D supplements. His GAZ5 The Elves of Alfheim (1988) for the D&D campaign setting of the Known World introduced skills by name to the game. “Due to their background, elves have a variety of skills that are neither shown in the rule books, nor related directly to combat, thieving, or magic. These are optional additions to your D&D campaign.” RuneQuest’s designer put more cracks in the wall between skills and D&D’s classes.

A year later, GAZ11 The Republic of Darokin (1989) by Scott Haring expanded this skill system beyond elves.

“Each skill is based on one of the character’s Abilities (Strength, Intelligence, Wisdom, Dexterity, Constitution, and Charisma). When a circumstance arises in which the DM feels the use of a character’s skill is needed, he asks the player to roll a d20 against his current score with the Ability. If the result of the d20 roll is less than or equal to the Ability, the skill use succeeds. A roll of 20 always fails, no matter how high the chance for success.”

The gazetteer listed skills from advocacy and animal training to woodworking, but the options still kept away from the class specialties of combat, thieving, and magic.

In 1991, the Dungeon & Dragons Rules Cyclopedia gathered all the rules from the basic line into a single hardcover that included the skill system. Meanwhile, AD&D would spend another decade forcing players to say “non-weapon proficiency” in place of “skill.”

For D&D’s third edition in 2000, the designers finally gained permission to correct old drawbacks. “We knew we wanted to make a more robust set of skills,” designer Monte Cook said in an interview. “You had thieves‘ skills, which were different and they worked completely differently, because they were percentage based. So we wanted to marry all of that together.” Like RuneQuest and virtually every other contemporary roleplaying game, the new edition would adopt a single, core mechanic to resolve actions. Players made checks by rolling a d20, adding modifiers, and comparing the result against a difficulty class number. Skills now offered bonuses to these checks.

The older D&D skill system and AD&D proficiency checks had created in impression that the third-edition designers worked to avoid. In both systems, skills seemed like a requirement to attempt many tasks, so characters needed gemcutting skill to even attempt a radiant cut. That adds up. On the other hand, surely anyone could attempt bargaining and gambling, yet D&D’s original skill checks only applied to characters with a skill.

D&D’s new d20 core mechanic meant that skills expanded to include actions characters actually did in the game. For instance, rogues got skills rather than a private table listing their chance of hiding and picking pockets. “D&D was still a class based game, but the idea that you were not a thief, so you can’t climb and you can never climb, didn’t really hold a lot of water.” The system allowed any character to attempt to hide and climb. Unskilled characters just suffered worse odds of success. Good luck with the gemcutting.

By fourth edition the games designers worked hard to reach Gary Gygax’s ideal of teamwork—but only during combat. On the battlefield, each character class served a distinct role like striker and defender. For tasks outside combat, the designers contrived a skill challenge system aimed at ensuring that every character gained an equal chance to contribute.

During fifth edition’s design, the D&D designers planned to sideline skills in favor of simple ability checks. “We’re making skills completely optional,” lead designer Mike Mearls wrote. “They are a rules module that combines the 3E and 4E systems that DMs can integrate into their game if they so desire.”

But playtesters liked the depth that skills gave characters. Also finessing the game’s math so it played equally well with or without skill bonuses doubtless proved troublesome. So skills stayed part of the D&D core. The designers still chose to rename skill checks as ability checks. This further avoids from the implication that characters need a skill to attempt certain tasks. Without formal skill challenges, fifth edition allows characters with particular skills to shine more as individuals who bring special talents to contribute to the team.

And in the end, no one had to say or type “non-weapon proficiency” ever again (unless they tell this story).