Movement & Scale#

How the sim's distances and movement speed map to FFXIV. The goal is that a fight ports over 1:1 from the game: 1 sim unit = 1 FFXIV yalm, and travel times match.

Constants#

All defined in src/engine/sim.ts. Movement is server-authoritative — only these engine constants govern it; the client just sends intents and renders snapshots.

A standard single-circle arena is radius ≈ 20 (FFXIV's common arena size). Author distances, radii, spawns, and waymarks directly in yalms.

Why 6.0 (and not 5.69)#

Derived from real FFXIV measurements in a radius-20 arena (cardinal waymarks at the edge, C→A diameter = 39.81 yalms):

• C→A normal run was measured at 7 s → 39.81 / 7 ≈ 5.69 yalms/s.
• FFXIV's canonical run speed is 6.0 yalms/s (39.81 / 6 = 6.64 s).

The 0.36 s gap is FFXIV's standstill acceleration ramp — the player spends the first fraction of a second speeding up. This sim has no acceleration model: players hit top speed instantly. So we match FFXIV's top speed (6.0), which makes steady-state movement identical. The trade-off is that a full edge-to-edge crossing in the sim is ~0.3 s faster than the in-game stopwatch (the sim doesn't lose the ramp time). Matching the measured 7 s instead (5.69) would have made all steady-state movement ~5% too slow, which is worse.

Sprint is 1.3×, matching FFXIV's in-combat Sprint (+30%).

Notes#

• botIntent.ts imports MOVE_SPEED, so bots and players always move at the same speed — no separate bot constant to keep in sync.
• Some older raids (e.g. sample-raid.yaml at radius 38) predate this scale and were not rescaled. New raids should be authored in yalms.
• Changing MOVE_SPEED can break tests with fixed-duration travel assertions (e.g. the "walks off arena" fall test, whose tick window assumed the old speed). Run bun test and adjust durations, not behavior, if any surface.