The spawn window is now accurate. An internal design value tracks how many seconds a wave takes to spawn all its enemies — used only by the two internal design tools, not by the game itself. It was off by 1.4 to 3 times too high, out by 143 seconds on wave 49. Nothing about the game changes because of this.
- The old formula multiplied the wave’s entire enemy count by the drone spacing. But only drones spawn in a chain; swarms, orbs, spinners, fortresses and meteors are rolled from time windows instead. An oversized window made the tools’ “pressure per second” estimate too low — they reported dead time where there wasn’t any.
- It’s been replaced with two measured formulas, one for normal waves and one for guardian waves, both fit to match sampled data within about ±2 seconds. The old flat 45-second guardian-wave estimate hadn’t been accurate anyway since guardians started bringing an escort (see 0.113.0): actual measured time is 16 s on wave 10 and 64 s on wave 50.
- The design tool keeps the old estimate as a sanity check: if the window drifts far from the measured wave duration again, something changed in how waves are built without the formula being updated to match.
And a second bug the first one had been hiding. A separate internal difficulty-estimation tool carried its own model of player power and had never been recalibrated after 0.112.0 — missing the boon math, using an outdated weapon-level curve. As long as the spawn window was also wrong, the two errors roughly canceled out. With the window fixed, the remaining one stood out on its own, and the tool reported a “wall” in sector 3 for a Veteran run without hangar upgrades — where actual telemetry averages wave 9.9.
- The player-power model used by both design tools now lives in one place, so they can’t drift apart from each other again.
- The reported difficulty walls are plausible again and line up with the in-game difficulty descriptions: Veteran without hangar upgrades ends around sector 4 (“Without a half-upgraded hangar, it’s over by sector 3”), Legend without hangar upgrades around sector 1.
- Without telemetry data, the tool no longer assumes a boon bonus of zero — it uses the measured value on file (+7.4 % per boon) instead, while flagging that the value isn’t freshly calibrated.
- The damage-ramp formula from 0.114.0 now lives alongside the other wave-tuning numbers, so both tools read the same figure. Behavior is unchanged.