# Seat time, scope, and pacing for simulations

> Use seat time as a scoping tool to decide how much one simulation should attempt.

- Canonical: https://docs.devlin.ai/design-guidance/seat-time-and-scope

Seat time is a scoping tool: deciding "this should take about eight minutes" forces the right conversations about how much one simulation should attempt. Decide it during design; when you have a target (a course slot, a compliance requirement, a shift-huddle window), build the simulation to honor it.

## What drives duration
- **Text simulations** run roughly 1-2 minutes per meaningful exchange: a focused scenario of 5-8 substantive learner turns lands around 8-15 minutes including reading and thinking time.
- **Voice simulations** run in real conversational time and always have an enforced session cap: the Max session length setting in the simulation's voice settings. When a session reaches the cap, it ends. The cap is the seat-time contract. Set it to the target, not above it.
- **Evaluation and feedback reading**, when evaluation is turned on for the simulation, adds 1-2 minutes after the conversation ends. Count it when filling a fixed slot.

## Matching scope to the target
- **Under ~5 minutes:** one narrow behavior, one situation, an early decisive moment. Good for drills and refreshers designed for repetition.
- **~5-15 minutes:** the sweet spot: one skill with genuine resistance, a complication, and a resolution the learner earns.
- **Over ~15 minutes in one sitting:** almost always a series pretending to be a simulation. Split by objective or by difficulty; learners retry a 10-minute simulation, they abandon a 30-minute one.
- Total seat time can be a SERIES budget: "45 minutes of practice" is better spent as four simulations of increasing difficulty than one marathon.

## Pacing levers inside a build
- End conditions determine real duration more than the optional max turns limit. A scenario that can end decisively when objectives are met (or clearly failed) respects the learner's time.
- Character resistance sets pace: more pushback = more turns per objective. Calibrate resistance down if the seat-time target is tight.
- Repetition is part of the time budget: a simulation designed for 3 attempts at 5 minutes teaches more than one attempt at 15.