The community filed Moon Elf in the bottom tier — a 'basic, low-power starter hero.' The engine read the same board and scored her above 50 into 37 of the 43 modeled opponents. Somebody is wrong, and the gap is the whole lesson: her arrows are small, but her real weapon is a tax code — a fizzled attack here, a stolen reroll there, and +2 stapled to everything you land. Play her as a damage hero and the D tier is right. Play her as a tax collector and the math flips.
Your opening turns have one job: get Targeted on the board and keep arrows flying while you do it. Longbow is the default — 50% on the very first roll, 98% across three attempts — and 4-7 early damage is honest pressure. Demising Shot (three arrows plus two moons) is the Targeted play, but it only lands 31% of dedicated turns, so don't force it: take it when two moons show naturally by the second roll, because Targeted is persistent — mark them on turn two and every attack for the rest of the game hits +2 harder. What you never do early is fish moons for their own sake: Eclipse hunts land 10.4% of the time and the ultimate 1.3%, so a greedy moon turn is usually a zero. One rules detail worth a Combat Point: the start player skips their very first Income phase, so going first means starting a point behind — sell a dead card early to catch up.
The middle game is one question asked every turn: which tax hurts THIS opponent most? Entangle punishes reroll-hungry heroes — anyone hunting four copies of a one-in-six face is living on their third roll, and you just deleted it. Blind is a held threat: one time in three their next attack fizzles outright, so spend it into their payoff window, not their chip. Targeted never comes off the priority list — re-apply it the moment it's cleansed, because +2 per hit at a 98% attack rate is nearly 2 expected damage per turn, the difference between 5.8 and 7.8. Then read the clock. Into engines and turtles you are the aggressor: the model has you killing Artificer in ~9.1 turns to Artificer's ~31 and Paladin in 14.8 to his ~48.8 on paper, so keep shooting and never give them setup turns. Into racers, invert: Covered Shot fires 49% of turns for 7 damage plus an Evasive token, so bank dodges while you chip and make every one of their hits swingy.
Late game, the pile wins it. Eclipse — 7 damage plus all three of her status effects in one swing — is the turn that breaks a defense built to answer one problem at a time, and one Cleanse token can't strip three effects. Lunar Eclipse! is the true closer: 12 unpreventable damage plus the full status pile plus an Evasive for you, and once it activates nothing in the game responds — no defense roll, no token, no card. At 1.3% you don't fish it; you buy it. The moon is her 6 face, and roll-phase dice manipulation converts Combat Points into moons — three natural moons plus purchased sixes is how good pilots fire a 1.3% line on purpose. Two cautions while you set up: never launch the Eclipse turn into waiting attack negation — a Shadows token, a Parlay, a held Smoke Bomb — feed those a plain Longbow first and land the pile after. And know your own insurance: Evasive is a one-in-three save per token, and you can hold up to three — roughly a 70% dodge (if stacking on one hit is legal — DiceSlayer verifies on camera) — but it cannot be spent against an Ultimate — the only Ultimate counterplay in the rules is altering dice before it activates, so a held roll-phase card is your real late-game shield.
Her die is three arrow faces, two feet, and one moon — arrows on values 1-3, feet on 4-5, the moon on 6 — and that composition writes the priority list. Any two arrows on the opening roll is a keep: finish Longbow, the 98% line. Entangling Shot — any four straight values — lands 59% of turns with best play for 7 damage plus Entangle, and it's your best pivot when the opening roll scatters; keep exactly one of each middle value while you chase it. Covered Shot (two arrows, three feet) comes in at 49% and pays 7 plus an Evasive — take it when feet flood the roll and you want insurance. Demising Shot is a 31% line — commit only when two moons arrive by roll two, for the permanent Targeted mark. Exploding Arrow (one arrow, three moons) fires 24% of the time and rolls five fresh dice for about 7 expected damage, drains roughly 0.8 of their Combat Points, and Blinds — a fine cash-in for a moon-heavy roll you didn't ask for. Blinding Shot's large straight is a 25% windfall, never a plan, and Eclipse hunts connect 10.4% — three moons by roll two or walk away. Last rule: when YOU are Entangled, you're down a reroll, so take the good roll-two ability instead of gambling a broken hunt.
Her board spends nothing: no ability costs a Combat Point, and there's no stand-up tax hiding in her kit — so her entire CP curve funnels into cards, and the job of her cards is moons. The moon is her 6 face, which makes roll-phase dice manipulation a direct currency exchange: points in, moons out, and suddenly the 10.4% Eclipse and the 1.3% Lunar Eclipse! are purchases instead of prayers. Exploding Arrow plays economy defense at the same time, draining about 0.8 CP from the opponent per cast — against heroes whose comeback lines run on points, that's a second tax. Keep a small reserve for the defensive-roll window: it is the last legal moment to spend status effects and cards before damage resolves, and for a hero whose defense is conditional, that window is worth more than any main-phase luxury. Every card you'll never play sells for exactly 1 CP — that's another die set to a moon later.
You draw one card a turn, hold at most six, and every sale is exactly 1 CP, so hand discipline is simple: keep what makes moons, sell what doesn't. Priority one is roll-phase dice manipulation — it converts near-miss moon rolls into Eclipse turns, and it is also the only counterplay to an enemy Ultimate anywhere in the rules (alter their dice BEFORE activation, because afterward nothing responds). The hero whose Evasive tokens and Missed Me are both worthless against Ultimates needs that insurance more than anyone at the table, so the last die-altering card in your hand is never a value play — it's armor. Priority two is anything that interrupts or prevents damage, because a big prevention effect is the guaranteed defense her board never printed. On her own upgrade paths, honesty over guesswork: Moon Elf's card list isn't in the verified data yet, so no upgrade-priority claims here until DiceSlayer verifies them on camera — the one rules anchor is that upgrades can never be sold back, so against the racers who beat her on pace, be suspicious of early points spent on permanent value instead of survival.
Deny her the mark: cleanse or purge Targeted on sight and re-check it every turn — it is a persistent +2 on everything she lands, the single number that turns her from a 5.8 hero into a 7.8 hero. Deny her defense the chance to matter by bringing guaranteed damage: poison ticks, upkeep chip, and undefendable attack lines never meet the Missed Me roll. And deny the pile: save your attack negation — a Shadows token, a Parlay, a Smoke Bomb — for the turn she's holding moons, not for a 4-damage Longbow.
She is weakest before the mark lands: every turn without Targeted on you, she's a 4-to-7-damage archer with a coin-flip defense. Her dedicated moon hunts are your free turns — an Eclipse fish connects 10.4% of the time, so nine of those turns in ten produce nothing while you swing freely. She's thinnest right after her Evasive bank empties: each token is a one-in-three save, and once they're spent her only defense halves damage on about 54% of rolls. And when your Ultimate is up, she has no answer at all — Evasive can't be spent against it and Missed Me never rolls.
Nobody burns a ban on a 'D-tier starter,' and that reputation is exactly why she travels well — she arrives unbanned while the engine quietly scores her above 50 into 37 of 43 opponents. The real logic: if your roster leans slow — engine builders and late-game scalers — she is a legitimate ban (63.7 into Alchemist, 61.8 into Artificer, 61.1 into Paladin). If you brought a racer, keep the ban: Iceman (44.4), Ninja (47.1), Barbarian (47.9), and Pyromancer (49.0) already beat her on pace, and a ban spent on a hero you outrun is a ban wasted.
She's an anti-engine specialist wearing a starter hero's reputation. In bring-3-ban-1 she slots next to a fast closer: the racer covers the four pace matchups she loses (Iceman, Ninja, Barbarian, Pyromancer), while she farms the slow heroes a racer hates sitting across from — Alchemist, Artificer, Paladin, Huntress are all 61-plus. The tier list is her camouflage: opponents who prep by community placements don't ban her and don't fear her, and then their slow hero walks into a modeled 60-40 with a permanent +2 stapled on by turn two.
In 3/1/0 Swiss, a control plan that times out scores 1 point on a day the podium needs 3, and her natural game trends long — the modeled clocks read 14.8 turns against Paladin and ~13 against Gunslinger even in matchups she's winning. Targeted is the accelerator: applied early, it lifts her rate from 5.8 toward 7.8 and cuts roughly a quarter off her own clock, so make the mark as early as the dice allow and keep arrows flying. Save the full status-lock script for opponents who threaten to out-race you; against fellow grinders, shoot first and tax second.