The math adores Treant — the engine puts him at 50 or better against 39 of the 43 heroes it models — and the community still files him under C tier. Both sides are telling the truth. His garden of Spirit companions wins almost any game that runs long enough; the problem is that 'long enough' and a tournament round clock are natural enemies. Learning Treant is learning when to stop gardening and swing.
Your opening turns are already an engine — the job is to not break it. Fertilize hands you a Spirit every Upkeep before you touch a die, so tempo wasted on fancy fishing is doubly expensive. Splinter is the default: 50% on the first roll, 98% of turns with rerolls, and 5 to 7 damage as the branch count climbs. Against reroll-hungry opponents, pay Splinter's rider early — discard one Spirit to attach Barbed Vine, and their digging costs 1 damage per extra roll attempt, up to 2 a turn. When leaves and swirls flood the roll instead, pivot to Tend, a 46% line with best play: it deals nothing, but a drawn card, three grown Spirits, a Wellspring, and a fresh vine is a full turn of compounding. What you never do early is hunt swirls — the payoffs come later or not at all. One rules note: the start player skips their very first Income phase, so if you go first, sell a dead card early to stay on curve.
The middle game is one question on repeat: cash or grow? Let the clocks answer. Against slow heroes you have an enormous margin — 8.4 turns to kill Artificer against her 31 — so grow greedily and let Rooted turn their attacks into Spirits while blocking about 2.5 per hit. Against racers, flip it: cash Spirits the moment a kill or survival line appears, because a dead gardener grows nothing. Know your three companions cold. Seedlings (you can hold three) each spend for a single-die reroll — extra digs that never spend a full roll attempt. Saplings (hold two) spend for 1 health plus 1 CP, or with 1 CP added, a card draw. The Dryad (hold one) is the premium piece: +3 damage on any attack, or a shield that blocks an incoming negative status outright — his matchup files spend it on Stuns, Parlay, even an incoming Bleed. And keep Barbed Vine refreshed; it removes itself each time it bites, and an opponent who stops rerolling to dodge the tax has quietly handed you their consistency.
Endgames are about choreographing one cash-out the opponent cannot answer. Count their outs first — a Smoke Bomb, a Parlay, a Shadows token — and jab with Splinter until those are spent or baited, because a 10-damage Overgrowth thrown into a live answer wastes turns of growth. The double-cash Overgrowth is the standard finisher: 2 base plus 4 per Spirit removed reaches 10, a held Dryad pushes it to 13, past his printed 12 ceiling, and discarding a Wellspring makes the whole thing undefendable — no defense roll, only cards and status effects can touch it. If swirls arrive naturally with a big stack, Nature's Grasp is the luxury line: it grows two Spirits first, then deals 5 undefendable plus 1 per Spirit you own, landing 8-11 with a typical stack. Spend leftover Wellsprings in your Main Phase — roll a die, heal half rounded up, about 2 a pop. The ultimate is honest fine print: 1.3% even with full rerolls. Treat five swirls as a gift, never a destination.
His die is three branches, two leaves, one swirl — faces 1-3, 4-5, and 6 — and that lopsided composition writes the priorities. Branches are the default keep: Splinter runs 50% on the opening roll and 98% across three attempts, and each extra branch is another point of damage. The best pivot when the roll scatters is Vengeful Vines — any four-in-a-row lands 59% of turns with best play for 7 damage plus a free Barbed Vine, no Spirit discard required. Overgrowth's set of two branches and three leaves is a 49% line: take it when leaves show up early and you have Spirits worth cashing — its 2 base damage is nothing without the sacrifice. Tend at 46% is the value pivot when you want growth over damage. The chase lines get no respect from the odds: Call of the Wild's full straight arrives 25% of the time, Nature's Grasp needs four copies of a one-in-six face and hits 10% even fully dedicated, and the ultimate is 1.3%. Take swirl payoffs when three arrive by the second roll; never dig for them from zero.
Treant's economy is the quiet luxury of his kit: the engine runs on Spirits, not Combat Points — as his Tactician matchup puts it, his engine runs free on passives while the opponent's costs tokens and turns. That leaves his income unusually uncommitted, with three standing jobs. First, mobility tax: keep 2 CP banked against Knockdown heroes — his gunslinger matchup notes put standing up at 2 CP against losing your entire attack phase, and a hero whose plan is compounding turns cannot afford to skip one. Second, conversion: a Sapling spent with 1 CP added becomes a card draw, and a Sapling spent alone returns 1 CP with a heal, so the companion bank and the wallet trade in both directions. Third, cards and upgrades in the Main Phases — with no printed ability demanding CP, he can afford interaction where hungrier heroes cannot. Just remember upgrades can never be sold back: against the racers who already out-pace him, early CP sunk into permanence instead of survival is how favorable-looking games end early.
Honesty first: Treant's own action-card list is not in the verified data yet, so no named-card promises here — DiceSlayer will verify the deck on camera. The rules-level discipline still holds. You draw one card per turn, hold at most six at the Discard phase, and every card sells for exactly 1 CP — and Treant out-draws most income, since Tend draws on a 46% line and a Sapling plus 1 CP draws again, so sell ruthlessly or the hand cap does it for you without pay. Priority hold: Roll Phase cards, the dice-manipulation weapons — altering dice before activation is the only counterplay to any Ultimate in the game, and Rooted cannot roll against Ultimate damage, so that insurance matters more to him than his 2.5-prevention average suggests. On Hero Upgrades, the rules anchor is that they pay only the difference going from level II to III and can never be sold — fine in the long games he wants, a real tempo risk in the short games his worst matchups force.
Deny him time — that is the whole assignment, because his own matchup notes say it plainly: the engine wins long games. Deny Rooted its double payday: small defendable hits let him roll away 2.5 damage AND grow a Spirit per attack, so bring damage that is lumpy, undefendable, or upkeep-based — Poison, Burn, Bleed, and Nanites never face his prevention roll at all. Deny the cash-in: from two banked Spirits onward a 10-damage Overgrowth is live, so track the count and hold your negation — Smoke Bomb, Parlay, a Stun — for precisely that turn, never for Splinter chip. And deny Barbed Vine its tax by taking a good second-roll ability instead of paying health to fish for a perfect third.
Three windows. The opening turns, before Fertilize compounds: with no Spirits banked his ceiling is the printed 12 and his Splinter chip is all he has — the earliest turns are the cheapest you will get. The turn immediately after a Spirit dump: the stored damage is gone and even his free income needs several turns to rebuild a threat. And any turn your attack is undefendable: Rooted is his one and only defensive tool, so undefendable lines hit him at face value while his 6.3-a-turn engine has to grind through your full kit. One non-window to stop respecting: the five-swirl ultimate is a 1.3% event — plan around the Overgrowth spike, not the miracle.
Ban Treant based on your roster's speed, not his reputation. If you brought grindy setup heroes, he is quietly your worst nightmare — the engine has him 63.5 into Alchemist, 61.5 into Artificer, 61.2 into Paladin, and 60.3 into Necromancer (Huntress's raw 61.6 actually sits second, but her own matchup file tells you not to trust that number — Nyra isn't modeled) — and his C-tier reputation means almost nobody spends a ban on him, so the nightmare usually survives to the table. If you brought racers, keep your ban: Iceman (44.9), Ninja (47.4), Barbarian (48.2), and Pyromancer (49.2) already answer him on pace. And read the person, not just the hero — nobody brings the game's most complex hero to a tournament casually.
Treant is a specialist, not glue: the designated engine-slayer of a bring-3. Slot him next to a fast closer — the racer covers the four matchups where he drowns (Iceman, Ninja, Barbarian, Pyromancer), while he farms the slow setup heroes a racer hates grinding through. His community C tier is roster camouflage: opponents rarely burn a ban on a C hero, so your anti-engine tech tends to reach the table intact. The honest tax is practice — at complexity 6, the highest in the game, he is not a hero you pilot cold. Bring him drilled or bring someone else.
Here is the uncomfortable truth: 3/1/0 Swiss scoring is openly hostile to his best game. His own matchup files brag that the engine wins long games — and a long game that hits time is not a win, it is 1 point, the same as a stone-cold draw, on a day the podium needs 3s. Worse, his most lopsided matchups are his slowest — Paladin needs a computed 49 turns to kill him, but even the winning script there is Treant's own 13-turn clock, a long game of the slowest decisions in the format. So in Swiss, pilot against his nature. Cash Overgrowth at two Spirits instead of hoarding toward six. Spend the Wellspring for the undefendable line and take Vengeful Vines' 59% tempo swings. Treat the hoard-and-out-grow script as a luxury for untimed play, and accept a slightly worse winning position in exchange for a finished game. When you reach a dominant board early — the 8.4-versus-31 Artificer script — close with spikes, not sustain. Three points requires a knockout before the bell.