The recorded Tsukumo idea is to produce and copy trinkets repeatedly, then turn trinket count into damage, burn or faster ability cycles. Build the producer before chasing the payoff; the filmed run is one example, not an assured recipe.
See it in the game
Watch the creator play through this part of the game.
Watch this section in real gameplayOpen the original creator’s video player only when you want it.
Original video: Watch on YouTube ↗ · Footage remains with its creator.
See the actual run before copying a board
The Japanese September 16 recording begins by choosing Scavenger, looking for Tsukumo units, and repeatedly checking non-unique common trinket generation and bonus-copy effects. Later, the creator combines those rewards with effects that scale from trinket count and shorter cooldowns. Automatic subtitles mangle several proper names, so the guide avoids claiming an exact English name for every creature.
Four jobs the engine has to fill
| Role | What it should do | Question before buying |
|---|---|---|
| Producer | Create trinkets at battle start, cast, victory or another clearly labeled trigger. | Will this trigger happen while the unit is fielded? |
| Multiplier | Copy or produce additional eligible trinkets. | Does it exclude unique trinkets or certain rarities? |
| Payoff | Convert the expanding pool into combat value such as damage, burn or speed. | Does it scale per trinket, on gain, or only once? |
| Survivability | Keep the team alive while repeated triggers accumulate. | Can the fight last long enough for the engine? |
The filmed decision path
- Day one: identify a source of eligible common/uncommon trinkets before committing to a full Tsukumo board. Creator rerolls repeatedly while rarity is low; do not translate this into a fixed probability.
- Upgrade a producer: the footage shows the creator aiming to level a trinket-generating creature and watching copies increase, rather than assuming every extra copy activates automatically.
- Field the scaler: at roughly the 13-minute mark, the creator observes that an expected permanent-damage gain did not occur while a unit was off-board, then changes placement. Verify this with the printed ability in your client.
- Reduce cooldown: subsequent swaps move a cooldown support unit into the lineup and adjust its target. Check actual duration on the inspect sheet.
- Watch the counter: the run visibly reports trinket totals past 50 and later 100. That is what happened in this run, not a predicted output from following a guide.
When to abandon the chase
If you have spent multiple shop visits without a reliable producer, if your only payoff is a late rarity you cannot buy, or if your board dies before casts fire, stop forcing the same search. Move toward a damage-dealing unit already offered. The video itself has moments when a desired copy is missing and the creator considers replacing it.
How this differs from a tier list
This is a sequence of verifiable decisions, not a claim that one named creature is the strongest in every matchup. Rewatch the embedded segment around the midgame team rearrangement if your buffs appear not to stack.
