Our own data, from ranked matches above 7500 MMR.

Quebra de Mana
Queima o mana de um adversário a cada ataque e causa dano proporcional ao mana queimado.

Translocar
Teletransporte a curta distância, permitindo ao Anti-Mage entrar e sair do combate.

Contrafeitiço
Concede resistência mágica passivamente. O Contrafeitiço pode ser ativado para criar uma bolha antimagias ao redor do Anti-Mage que bloqueia e reflete habilidades alvejadas.

Opressor
Os seus ataques retardam inimigos proporcionalmente ao mana faltante. Retardo mínimo com 60% de mana e máximo com 0% de mana. Não surte efeito caso o inimigo tenha mais de 60% de mana.

Vácuo de Mana
Causa dano em uma área baseado nos pontos de mana faltantes na unidade-alvo, que é miniatordoada.
Anti-Mage is a hero of the clock. Not of kills, not of damage, not of teamfights: of the clock. He exists to reach a point in the game before the enemy reaches theirs, and everything the hero does from the first wave to Manta serves that race. The problem is that most players understand half of it. They understand they need to farm fast, and stop there. Farming fast is the means. Turning farm into map is the end, and that is where the hero loses.
An Anti-Mage with Battle Fury at minute 12 and Manta at minute 20 is on schedule. If at minute 25 all three enemy tier 1 towers are still standing, he is late, and no GPM number fixes that. Every minute the AM farms without threatening a tower is a minute the enemy team farms too, with one more hero in every fight that happens.
The question that separates the Archon AM from the Divine AM is not how much gold per minute. It is: where are the enemy waves dying? If they are dying at your tower, the enemy controls the map and you farm what is left. If they are dying at their tower, you choose where to farm, and that is worth more than any item.
Anti-Mage's Blink is the most misused ability on the hero, and the mistake comes in two opposite forms:
The use that defines rank is the third one: Blink to cut the wave behind the enemy line when their team is grouped to push. It is the AM that makes the enemy choose between their tower and yours, and while they choose, the AM is already gone.
| Mark | Where the AM has to be |
|---|---|
| Battle Fury | 11 to 13 min, with at least two waves cut in the enemy jungle |
| Manta Style | 19 to 21 min, first tier 1 tower already down |
| Third big item | 26 to 28 min, game with at least two tier 2 towers under pressure |
If a row of the table slips by three minutes, the item slips by three minutes and the game slips by ten. The AM does not recover time: he only gains what the enemy loses. That is why the hero fits teams that know how to play with four on the map, and falls apart with teams that need five in every fight.
Mana Void is not an ability to finish a carry. It is an ability to punish whoever spent everything. The right target is almost always the big-mana hero who just used the whole kit: the Storm who jumped in, the Crystal Maiden who cast the ultimate, the Invoker at the end of his rotation. The Void on that target is area damage in the middle of the enemy team, at the moment they have no answer left.
Mana Void on a carry with 200 mana at the start of the fight is the most common wasted ability in student replays.
The AM loses before the game starts in two situations. Against a fast push lineup (Chen, Enchantress, Lycan, Beastmaster): the enemy ends the map at minute 18 and no clock solves that. And against a low-mana lineup: an Anti-Mage against Ursa, Bristleback, Legion Commander and Wraith King has nothing to burn, and Mana Break becomes a small bonus instead of the reason for the pick.
In both, the hero can win, but not for the reason the player picked him. And that is usually enough to lose.
*If I keep farming for three more minutes, what will the enemy have taken down in that time?*
If the answer is nothing, keep going. If it is a tower, the AM is being paid in map for the item he is stacking. And map is the currency in which Anti-Mage is richer than any other hero. Spending it badly is the only way to lose with him.
The whole catalog (real match analysis, by hero, position and patch) opens with the subscription.
See plansMost played18 matches, 50% win rate
1
2
3
4
5
6
7
8
18
19
20
21
22
24
25Highest win rate3 matches, 100% win rate
1
2
3
4
5
6
7
8
9
11
12
13
14
15
17Starting items
96%
74%
73%
68%
54%
52%
17%
4%Core, in the order it usually comes out
min 1493%
min 1791%
min 2087%
min 2559%
min 2746%
min 2751%Situational
min 699%
min 1293%
min 3054%
min 3626%
min 3527%
min 3727%
min 3723%
min 2619%
min 2010%
min 2711%
min 459%The numbers come from the top of the ladder. What they do not say is whether the build holds up in the games played here. Your answer is what adjusts that.
Advantage in points over Anti-Mage's usual win rate: negative means trouble. World data comes first (STRATZ and OpenDota); our high-MMR matches take over once the sample is big enough.
OpenDota, public ranked matches from the last 7 days
Win rate
50.6%
505,351 matches
Pick rate
12.4%
of matches include the hero
Pro scene
0 picks and 1 bans

