Weapons with a mind of their own have long been a staple of sword-and-sorcery tales ("Hellooooo, Elric"). But almost no one was using them in 3rd edition D&D.

Diagnosing the Problem

Two years of D&D play convinced us that intelligent weapons were vanishingly rare, despite the rule in the 3rd edition Dungeon Master's Guide that said 15% of magic weapons were intelligent. Nobody was actually using that rule. Even we weren't using that rule in our games, and we wrote the book. The RPGA wasn't really using intelligent weapons at all. This section of the DMG was basically gathering dust.

Another problem: The system was based entirely on the roll of the dice. You rolled to set the brains of your magic sword, you rolled to see which powers you got, and you rolled for alignment. Worse, the power tables had an escalating feature: If you rolled high, you got to skip ahead to a table with more powerful abilities.

Don't get me wrong -- I love escalating die tables. But in this case, they weren't appropriate because the cost for the item had already been set. You could potentially get a sword with teleport, heal and disintegrate for only 10,000 gp. No DM is going to let such a weapon into the game, obviously. But the DM has no guidance for less obscene cases. You have a talking sword with three primary abilities, worth 25,000 gp according to the table. What powers are appropriate? What's too little? What's too much? The DM has nothing to go on beyond educated guesswork.

Knowing What You Want

After a lot of discussions with DMs, we knew we wanted to keep intelligent items relatively rare. They're effectively secondary characters that go everywhere the PCs go, but you don't want the talking sword taking the spotlight away from a living, breathing player at your gaming table. If every PC in the party has one or more talking magic items, that's a recipe for cacophony that'll bring the game grinding to a halt.

We also knew that players and DMs generally like designing their own magic items, and they appreciate the structure of the pricing formulas we've established for other items. We wanted to extend the "build it yourself, and pay for it yourself" philosophy.

Finally, we knew we wanted intelligent items to behave more like the creatures that they were.

Reassembling the Pieces

First, we kissed that "15% of weapons are intelligent" rule goodbye. The new rule says what we want it to: "In general, less than 1% of magic items have intelligence. Use them sparingly in your campaign." We also broadened the focus from magic weapons to magic items in general. Stormbringer-style magic swords probably will still be the most common intelligent items, but even spellcasters who rarely draw weapons can possess intelligent items.

Then we tore apart the tables and established an "a la carte" pricing structure. Now you pay for the brains of your intelligent item, and you pay for each power it has. That makes it easy for the DM to design an intelligent item that fits into the economy and power level of the ongoing campaign.

Because we wanted intelligent items to feel a little more like creatures, we gave them a power that normal magic items don't have: the power to take actions on their own. It's a subtle power, and I think people might miss its importance on a first reading. Actions are the fundamental currency in the D&D game -- the one thing no character can ever get enough of. Intelligent items effectively activate themselves when appropriate. A paladin can run around the battlefield smiting her foes while her magic sword heals her, checks on her allies with a status spell, and casts dimensional anchor so the demon can't flee back to the Abyss. The sword can do all of that without the paladin taking any actions to activate those powers. The whole point is that the sword is intelligent, after all, so it uses its powers whenever circumstances are right.

What Didn't We Do?

Along the way, we tried a number of new rules that didn't work out. We tried replacing the Item Ego system with a formula based on the market price of the intelligent item. In theory, it ought to work, because all of the aspects of an item that add to Ego also add to cost. But in practice, the only formulas that gave us similar results to the Item Ego table looked like calculus car-wrecks. They were all so complicated that consulting the table was faster than plugging numbers into the formula.

We also tried integrating the intelligent item's purpose with a dedicated power and a dedicated curse. The idea was that a specific purpose (for instance, slay servants of Erythnul) should have a specific power attached to it (maybe some kind of continuous area-effect regeneration, calm emotions, and some other healing powers) and a specific curse that the item would use if its owner misbehaved (like wracking pains whenever someone within a mile suffers a violent death). It's a promising notion, but we soon realized that a good list of linked purposes, powers, and curses would take up too much space -- especially for a bit of the rules we wanted to remain relatively rare.

Return to Main Page