hitstun-recover
Hitstun and recover
Resolve the column from the project before tuning:
| Column | After the hit |
|---|---|
| none | flinch presentation, no action lock |
| light-stun | timed action lock, then legal actions |
| knockdown | downed state plus authored wakeup options |
| juggle | airborne state with explicit landing and escape rules |
Hitstop and hitstun are different. Under action-feel, each affected actor pauses its move clock during hitstop while simulation input sampling continues. Count that actor's remaining stun or recovery on its advancing ticks, not render frames. Publish exceptions for asymmetric hitstop or other clock policies.
Timing contract
Record hitstop per participant, hitstun, blockstun, remaining active/recovery at impact, knockdown duration and wakeup invulnerability. Use half-open intervals [start, end) and a declared logical frequency. Do not assume wakeup invulnerability or blockstun shorter than hitstun for every design.
Frame advantage = victim first-action tick minus attacker first-action tick, measured on the same global simulation timeline. Positive means the attacker can act first. Only with equal freezes and matching interval conventions can this simplify to stun minus the attacker's remaining commitment at impact; total move recovery alone can miss remaining active frames.