Abstract
Problem: What would Baldur's Gate 3 have looked like if Troika Games had made it instead of Larian Studios?
Approach: Tim Cain discovered Troika's original BG3 proposal document from July 2003, buried in his Temple of Elemental Evil archives, and walked through its design in detail.
Findings: Troika proposed a radically different BG3 — a first-person, real-time, action-oriented RPG with a fatigue-based magic system replacing D&D's spell memorization, only four attributes, eight classes with no Paladin, and competitive multiplayer. The proposal was submitted as Temple of Elemental Evil wrapped up but likely received no response from the publisher.
Key insight: Troika wanted to make "Adapted D&D" — D&D redesigned for real-time first-person play on PC — a philosophy born from learning that faithful tabletop-to-computer translation (as in Temple) wasn't the right approach.
1. Context and Discovery
Tim Cain found Troika's BG3 proposal document dated July 2003 buried in his Temple of Elemental Evil archive folder. This was written before Temple even shipped (fall 2003) and before Tim moved on to Vampire: The Masquerade – Bloodlines. Most of the team was busy finishing Temple or working on Vampire, so the proposal was put together quickly.
Troika submitted the proposal as an alternative project to pitch after it became clear the publisher didn't want Tim's preferred follow-up to Temple — a continuation through the classic D&D adventure modules Against the Giants, Vault of the Drow, and Queen of the Demonweb Pits.
2. Core Vision: First-Person Real-Time D&D
The game would have been PC-only, first-person, and real-time — a radical departure from the isometric Baldur's Gate formula. Melee combat would switch to third-person, similar to Jedi Outcast (2002), which used first-person for ranged combat and switched perspectives for lightsaber fights.
Tim called this design philosophy "Adapted D&D" — D&D rules redesigned specifically for real-time first-person computer play. He explicitly framed this as a lesson learned from Temple of Elemental Evil, where the team tried to faithfully recreate the pen-and-paper rules and found that approach limiting.
3. Overhauled Attribute System
The proposal stripped D&D's six attributes down to four physical attributes only:
- Strength
- Dexterity
- Constitution
- Charisma
Intelligence and Wisdom were removed entirely. Attributes were not rolled — they were determined by your class template and adjusted by race. As you leveled up, attributes increased automatically based on class, with no player choice involved.
4. The Fatigue System
The most radical mechanical change was replacing D&D's Vancian spell memorization (spells per day) with a fatigue-based resource system — effectively a mana pool called "Fatigue." Both combat maneuvers and spells drew from the same fatigue pool. This meant a multiclass fighter/mage would use the same resource for sword techniques and spellcasting, creating interesting trade-offs.
5. Races and Classes
All seven D&D races of the era were included: Dwarf, Elf, Half-Elf, Halfling, Human, Gnome, and Half-Orc.
Only eight classes were planned — notably with no Paladin and no Assassin:
- Bard — Songs and armored spellcasting
- Cleric — Divine spells and Turn Undead
- Druid — Spells, shape-changing, and animal companion
- Fighter — Highest base attack bonus, most feats
- Mage — Arcane spells and familiars
- Monk — Unarmed attacks, AC bonus without armor
- Ranger — Tracking and archery
- Rogue — Sneak attacks and evasion
The absence of Paladin and Assassin tied into the alignment system: all characters started neutral, and alignment shifted dynamically based on player actions along the lawful/chaotic and good/evil axes.
6. Skills and Feats Redesigned
Troika split character abilities into two categories:
- Skills — Player-activated abilities (active use). Improved with skill points gained each level. Every successful skill use granted a small amount of XP.
- Feats — Always-passive bonuses (e.g., Cleave letting you hit multiple adjacent enemies automatically). Gained one or more per level, with fighters getting the most.
7. Multiclassing and Prestige Classes
Full D&D-style multiclassing was planned — at each level-up, you could spend that level in any of the eight base classes. Prestige classes were also included, accessible after reaching sufficient levels in base classes.
8. XP and Progression
Experience came from three sources:
- Kills
- Completing quests
- Successful skill use (with combat skills weighted down since they're used more frequently)
9. Setting and Scope
The entire game would take place in and around the city of Baldur's Gate — Tim felt that as a sequel, it was important to stay in the same location, and that there were still plenty of stories to tell there.
10. Single-Character Focus with Followers
Unlike the party-based originals, this was a single-character game. You could recruit one or two followers, restricted by charisma, class, or level (the details weren't finalized). This fit the first-person action-oriented design.
11. Multiplayer: Co-op and Competitive
The proposal included multiplayer in two flavors:
- Co-op — Friends group together and quest cooperatively
- Competitive — Either arena PVP combat, or racing to complete the same quest first
12. Modding and Tools
Troika planned to ship BG3 with all of their development tools (built on the enhanced Arcanum engine), actively encouraging user-created content. Tim envisioned the game shipping with a large single-player campaign plus multiplayer arenas and race quests, with the community quickly producing more multiplayer content and eventually full single-player campaigns.
13. What Happened
Tim believes the proposal was submitted and either received a rejection or — more likely — no response at all. The game was never made. Tim noted the irony of discovering this document just as Larian Studios was about to release their own, very different, Baldur's Gate 3.
14. References
- Tim Cain. YouTube video. https://www.youtube.com/watch?v=dz6mHKFEAdU