
Outworld Destroyer
Mid lane is a war over mana: if the enemy can't cast, he can't kill you. Steal a chunk of the enemy mid's max mana with every Astral Imprisonment, then turn the gap into pure damage with Sanity's Eclipse. And unlike most pub heroes — the better your bracket, the better he gets.
This hero wins the lane before it starts: Astral Imprisonment steals the enemy mid's max mana — no mana, no spells, no way to trade back.
Patch verdict: he scales with your bracket — Herald 48.2%, Divine 54.9%. Buffed twice this year (7.41d attack damage, 7.41e barrier); 7.41f left him alone.
Draft discipline: if the enemy locked in zoo or push — Chen, Beastmaster, Broodmother — don't pick OD. They don't care about mana, and they end the game before you scale.
Is he still good this patch?
Yes — and he got quietly stronger all year. 7.41d gave him free attack damage for last-hitting, 7.41e turned his new E (Objurgation) into an instant-cast barrier with a shorter cooldown, and 7.41f didn't touch him at all. On the current patch he sits at 52.9% win rate with a 9.4% pick rate. Through 7.41a–7.41d he was stuck around 48% on a 4.1% pick rate; the step came at 7.41e and the pick rate has more than doubled since.
But the number that actually defines this hero is the shape of his curve (current patch, all brackets):
The higher the bracket, the better he is — the exact opposite of the classic pub stomper. Two reasons: his kit rewards timing (Astral has a 0.3s cast point, and a banish can win or throw a fight), and his damage is math — the mana gap you spend all game engineering cashes out in one button. Low brackets pick him as a gimmick. High brackets build around him.
What's the skill build?
One decision, two builds: fight early (max Astral) or farm for the late game (max Arcane Orb). The majority (55%) fights — Astral is damage, a banish and the mana steal that feeds your ult in one button — and it performs slightly better. Pro player Speeed's farming version takes two early points of Astral, then maxes the orb to clear waves faster. Either way: Objurgation's first point comes at level 4 (it's your only gank survival), ult at 6, 12, 18.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
18
20
25What's the item build?
Mid-lane intelligence opening into attack speed. The order below follows the median purchase times across current-patch games — read it as a schedule, not a law.
Default build:
Tango
Faerie Fire
Branches ×2
Circlet→
Null Talisman
Boots→
Power Treads (INT)→
Witch Blade→You initiate
Blink Dagger→they dive you
Force Staff→
Dragon Lance→
Hurricane Pike→Late
BKB
Aghanim's→Close it
Scythe of Vyse
Parasma
Three things about this list. Power Treads is non-negotiable — 96.3% of OD games buy it. Keep it on INT when you cast (your whole kit reads your mana pool) and swap to AGI when they focus you; tread swapping is free. Witch Blade is the power spike — attack speed plus a bonus nuke, and it makes your orb harass miserable to trade into (median timing 13.7 minutes). Hand of Midas is a low-priority greed buy: the parsed games lean hard against it (buyers won markedly less often), and it delays the Witch Blade that makes you dangerous. In a slow, free game it's defensible — just know what you're trading.
Aghanim's Scepter is not a damage item on OD — it's a second life. Drop below 20% HP and it auto-casts an undispellable barrier plus a strong dispel (80s internal cooldown). If the enemy has a pick-off hero, buy it earlier than feels right. The classic closer is Scythe of Vyse: hex covers your ult's wind-up, and it's the single strongest late-game purchase on this hero (top relative win rate when bought, median timing 36.1 minutes). The pure-greed closer is Moon Shard — more attack speed means more orb hits, and it shows up in most late-game build lists.
About Meteor Hammer — the one item the community argues about, and the argument recently flipped. English-language players were still lab-testing it in September (Topson included); the newest Chinese guides now say 'just buy damage items instead'. The numbers sit in between: 11.5% buy rate, small positive win delta, median timing 9.8 minutes — a real but niche choice. Buy it when your Astral aim is the reliable part of your game and the enemy can't interrupt the channel. And remember: Topson building it doesn't mean you should.
How do you play the lane?
Mid, obviously — 94% of his games are mid, and the mid cohort wins the most (53.7%). The lane has two jobs: win the mana war, and last-hit under orb pressure.
The harass trick: manually casting Arcane Orb doesn't draw creep aggro. Stand at the edge of their wave and orb the hero — the creeps ignore you. Leave auto-cast off until the lane is already won; an orb on cooldown at the wrong second is 20% of your mana gone.
The mana war: Astral Imprisonment the enemy mid every time it's up. A level-2 Astral takes 19% of his max mana for 40 seconds — two casts in, he can't afford his own combo, and that's when the lane is over: he can't trade, and he can't leave either.
Pro player Speeed's rule of thumb: right-click him, and the moment he's about to step out of your 450 attack range, Astral him — the banish reaches 650, so the gap is free damage plus free mana. And the dirty detail: Astral secures the ranged creep — banish the enemy as it dies, or banish the creep itself. Either way the last-hit is yours, and he's down gold and mana.
What do you do in teamfights?
One combo, in order: Objurgation shield on yourself → Astral Imprisonment their highest-mana hero → Sanity's Eclipse. The banish isn't just setup — the steal applies the moment you cast, so the gap your ult reads is already widened. And the ult resolves on imprisoned units: the blast lands on them while they're still inside the prison.
Everyone gets the ult math wrong, so here it is once: Sanity's Eclipse deals 200/300/400 base plus 0.4 × (your max mana − their max mana). Into a Lion or a Zeus it's a nuke; into a strength brawler with 600 mana it's an execution; into another INT hero — Pugna, Invoker, the enemy OD — the gap is zero and your big red button makes a sad noise. Check mana pools before you press R.
Your position is artillery, not initiation: 450 range behind your frontline, orb whoever steps in, and keep the barrier (150–300 flat + 12% of your max mana) for the diver who reaches you. If they drafted Rubick, shield before you Astral anything — a stolen Imprisonment turns your own mana pool into his weapon. And respect BKB: it switches you off. The orb's bonus damage and the ult both fail into debuff immunity — when their carry pops BKB, your damage collapses until it drops. Hold the blast for after.
Who's the worst to face?
The pattern across 75,000+ recorded games: OD loses to summons and push — damage that doesn't come out of a mana pool, so draining it changes nothing (Chen, Beastmaster, Broodmother, Enigma) — and to heroes that punish him for casting: Pugna's Nether Ward taxes every Arcane Orb, and Clockwerk's cogs leave him nowhere to run. He farms heroes who live on big spell buttons, because every Astral makes their kit more expensive than they can afford.
Hardest matchups
Chen37.1%
Beastmaster39.5%
Broodmother41.2%
Pugna42%
Clockwerk42.5%
Enigma42.6%Easiest matchups
Lion55.7%
Tidehunter55.3%
Dragon Knight53.9%
Anti-Mage53.4%
Sand King53.4%
Bristleback53%Chen posts the scariest number — 37.1%, albeit on the smallest sample (345 games). The zoo doesn't care about mana, and it takes your tower while you're still stacking yours. On the other side, Lion at 55.7% (788 games) is his most reliable prey: a mana-hungry support with no escape. Just remember his pool is big — orbs kill him, the blast doesn't.
37.1% — the scariest number on the table, though on the smallest sample (345 games). Summons don't care about mana steal, and the push ends the game before you scale. Dodge, or end it early.
55.7% against Lion (788 games) — his most reliable prey. Lion lives on a mana-hungry combo and has no escape: steal a fifth of his pool and he's a squishy hero who can't fight back. One caution — his pool is big, so your ult does less to him than to a strength hero. Kill him with orbs, not with the blast.
53.4% against Anti-Mage — the counterintuitive one. Mana Break eats your current mana, but your ult reads max mana, and AM's pool is tiny. Banish, blast, collect.
Things you might not know





Any off-meta ideas?
The honest answer: there isn't one. 94% of OD games are mid. Carry is 1.6% of games at 48.5%, offlane 1.9% at 47.8%, support 1.4% at 49.1% — every non-mid role is a coin flip or worse. His entire kit is built for a solo lane with fast levels: the mana war needs a 1v1, the orb needs levels, the ult needs items.
Want proof from the very top? At BLAST Slam VIII (October 2026), Team Spirit actually ran carry OD — and afterwards their carry Yatoro said, roughly: 'I hope we stop picking OD at position 1.' Even a TI-winning core couldn't make it work.
Don't step in these traps
You already know the plan — win the mana war, cash out the gap. The only question left is when the fights start. Queue with DOTA timer running: Bounty Rune, Wisdom Rune, Lotus, Tormentor, Roshan — it pings on time, with voice reminders. All you do is play.
