Poison is the comp that wins runs instead of saving them. It holds the highest win rate of any core in this window with 50 or more ranked games, which means it does not finish fourth and call it a day — it converts. The cost is that it is the comp most punished by rolling at the wrong time.
The core
| Unit | Why it is in the core |
|---|---|
| Bulbasaur | The Poison front body, and the unit that carries the item nobody else wants. |
| Nidoran | Second Poison body. Cheap, and the pair is what reaches the threshold early. |
| Diglett | Ground 1 for the damage line and the reason the comp is not just two walls. |
| Paras | The Bug piece. Two cost, and the unit that keeps the board's damage-over-time identity alive in a long fight. |
The trait spread is wide — Poison 3, Grass 2, Bug 2, Ground 1 — and that is the design: none of the secondary thresholds cost a slot, so the board gets four bonuses for four units. Compare it to Fire 3, which gets three thresholds but spends its cheap units getting there.
What the current file says
The Poison synergy in the current data file reads 50 / 120 / 170 damage per second at 2 / 4 / 6 units, applied for four seconds by an attack. Paras' Growth adds 10% Attack per use, with 4% lifesteal on its first three uses, so a long fight is where that unit pays for its slot.
That is also why the comp is played the way it is: cheap bodies, bought and starred rather than levelled into. Bulbasaur, Nidoran and Diglett are one cost and Paras is two, so the gold goes into stars, not into a legendary.
meta.json publishes a stage snapshot for this core: guide.stages carries stage 3 and stage 4, each with the units on the board and their star level, the level distribution when that stage ended, and the HP and gold bands. The block above prints each of those numbers next to the count of games it was computed from, and the cell block prints how many games record a board at all. Both counts are below the core's game count, because a game has to reach the stage — or report its final board — to be counted there.
Items
Bulbasaur is the item holder, and its most common pick is not a damage item — it is Flame Orb, on 52% of boards: HP and Defense, plus a per-second drain and a heal block on everything within one hex. That is the whole reason the unit is in the comp: the Poison threshold does damage over time, and Flame Orb is a second instance of the same idea. A Bulbasaur on a generic damage item is a worse unit than one on Flame Orb, and the feed's shares say so — Leftovers (42%) and Sitrus Berry (29%) are the next two picks, both survival items.
Diglett takes the crit and attack-speed line: Scope Lens on 85% of boards, Choice Band on 51% and Choice Scarf on 42%. It is the only real damage source on the board, and at one cost itemising it is not a waste.
Where players stand it
The cells in the summary at the top of this page are the aggregated placement from the games that record a final board, and the shape is the point: four of the five most-recorded cells are the front row — Paras, Weedle, Bulbasaur and Nidoran in cells 3 to 6 — and Diglett, the damage line, stands in cell 21 behind them. The cheap bodies absorb the opening; the carry stands where it is not the first thing hit. Poison wants the fight to last long enough for the damage over time to matter, and the placement is doing that job.
When not to play it
- You rolled at stage 3. This comp's entire advantage is that its units are cheap; spending gold to find them defeats the point.
- You already have a two-star expensive carry. Poison does not have a slot for it. Playing Poison around a Mewtwo is a board with two plans and neither of them finished.
- You want a high floor instead of a high win rate. Fire 3 reaches the top four more often than any other core in the window; Poison is the one that converts a top four into firsts. Pick by which of those two numbers you actually need.
Questions players ask
Poison 3 or Poison 4? Poison 3 with a real damage line. The fourth Poison body is usually a worse unit than the off-trait unit it replaces, and the jump from four to six units (120 to 170 damage per second) costs two slots.
Is Paras still worth a slot after the nerf? At two stars it is a cheap Bug body that the comp does not pay much for. The feed's data for this comp has it in 92% of boards — the second-highest share of the four, behind Bulbasaur — so players are not treating it as a slot to cut.
Why does this comp have fewer games than Fire 3? Because it is harder to play, not because it is worse. The average place and the win rate are the columns that compare comps; the game count only tells you how many people tried.