Abstract
Problem: Why did Arcanum ship with both real-time and turn-based combat, and why were neither mode properly balanced?
Approach: Tim Cain recounts the development history — from the original turn-based-only vision through Sierra's insistence on multiplayer (and therefore real-time), to the decision to ship both modes.
Findings: The dual-mode approach was a mistake. Neither mode was adequately balanced, and many promising turn-based ideas were left unexplored due to scope and team size. Tim believes turn-based RPG combat is an under-explored design space with significant untapped potential.
Key insight: Trying to support both real-time and turn-based combat without fully committing to either results in neither being good — pick one and balance for it.
The Original Vision: Turn-Based Only
Going back to 1998, when Arcanum was first specced out (before talking to Sierra), the team planned a purely turn-based game. Coming off Fallout, they understood turn-based design well. Tim particularly valued turn-based for one reason: area-of-effect spells. In real-time isometric games like Baldur's Gate, trying to cast a fireball while characters moved around was maddening — targets would move out of the area, companions would move in, and the result was unpredictable chaos.
Sierra's Multiplayer Mandate
When Sierra got involved, they insisted on multiplayer. Their argument was that single-player RPGs without a multiplayer component wouldn't sell well. Multiplayer meant the game needed real-time combat, and Sierra's preference was to make the entire game real-time.
The Compromise That Satisfied No One
Tim admits what he should have done was one of two things:
- His preference: Drop multiplayer and real-time entirely, go pure turn-based
- The pragmatic choice: Drop turn-based entirely, go pure real-time and balance for it
Instead, they kept both modes. Tim thought he could make it work. He was wrong.
The team tried to balance both modes through a shared system — action points for turn-based were used to calculate animation and attack speeds for real-time. They assumed this would create parity. It did not. What shipped was unbalanced for both modes.
Fast Turn-Based Mode
A third mode, "fast turn-based," was added partly to convince Sierra that turn-based could feel snappy. It skipped animations entirely — clicks resolved instantly, damage numbers flew off, and turns processed rapidly. This was Tim's attempt to show turn-based didn't have to feel slow.
Why It Went Wrong
The root cause was scope. Arcanum had too many features for a dozen-person team to code, debug, and balance in a reasonable timeframe. Tim acknowledges that while players loved features like newspapers, some things should have been cut to allow proper balancing of core systems.
Unexplored Turn-Based Ideas
Tim had numerous turn-based mechanics he never got to implement. Several stood out:
Collapsed Enemy Turns
When enemies had consecutive turns in initiative order, collapse them so all enemies act simultaneously. Taking it further: give one initiative roll to all enemies. Combat would alternate — your party members go (in initiative order), then all enemies go, then back to your party. Tim experimented with this later and found the pros and cons balanced out nicely.
Separate Movement and Action Pools
Split action points into two separate pools: movement points and combat action points. Each round would have two phases:
- Movement phase: Everyone moves in initiative order
- Action phase: Everyone performs non-movement actions (attacks, spells, potions) in initiative order
Reversed Action Initiative
A variant where movement happens in normal initiative order, but actions resolve in reverse initiative order. Fast characters move first but attack last; slow characters move last but attack immediately. This created interesting tactical dynamics — a ranged character could step away from an approaching melee enemy, fire their ranged attack, and the melee attacker wouldn't be adjacent to strike back. It gave meaningful advantages and disadvantages to high initiative.
Double Initiative Rolls
Everyone rolls initiative for the movement phase, then rolls again separately for the action phase. Bonuses apply to both rolls but guarantee nothing. A slow, low-agility character might roll well and act first in either phase — unlikely, but possible. Tim loved this for the same reason he loved critical hits: it introduced meaningful randomness that could turn a losing battle around or humble an overconfident player.
Real-Time With Pause Doesn't Solve the Problem
Tim notes that when players pushed back against fully real-time RPGs, the industry responded with "real-time with pause." For him, this solved nothing. His core issue wasn't that combat moved too fast — it was that area-of-effect abilities are fundamentally incompatible with moving targets. Pausing doesn't fix that; turn-based does.
Turn-Based Combat Is Under-Explored
Tim's central thesis: turn-based RPG combat has barely been mined for interesting mechanics. The industry rushed to real-time, then to first-person perspectives, leaving turn-based behind. He believes there's enormous untapped design space — particularly in the indie scene, which he suspects may already be experimenting with novel turn-based systems.
He expressed genuine enthusiasm for seeing more innovation in this space, and invited viewers to share examples of turn-based RPGs with unusual combat mechanics.
References
- Tim Cain. YouTube video. https://www.youtube.com/watch?v=-QUDeipApgU