Mirana Guide 7.41f: Build, Skill Order and Matchups

What Changed for Mirana Since 7.39

Mirana’s latest change came in 7.41f on September 15, 2026: Celestial Quiver base charge restore time went from 6 to 7 seconds, and Sacred Arrow’s maximum stun went from 2.6/3.4/4.2/5 to 2.3/3.2/4.1/5 seconds. The full text is in Valve’s official 7.41f patch notes. The table below lists every Mirana change in Valve’s patch-note feed from 7.39 to today. The feed shows no Mirana changes in 7.39b, 7.39c, 7.39e, 7.40b, 7.40c, 7.41b or 7.41e.

The pattern is not the usual nerf-after-nerf story. After the 7.41 rework introduced Quiver, three of her four follow-up patches were buffs (7.41a, 7.41c, 7.41d) and only the last one trimmed her. A hero who is buffed three times and then nerfed once, and still sits at 51.5%, tells you the nerf did not remove the core of what makes her work. The rest of this guide shows where that core is.

Is Mirana Good Right Now?

By public win rate, yes: Dotabuff’s last-seven-days All Pick data on October 25, 2026 puts Mirana at 51.54%, 2.09 points above the median hero, and she is above the median in all eight rank brackets, though by only 0.14 points at Immortal. The table below uses Dotabuff’s heroes table with the rank filter applied to each row, so the median hero is recalculated for each bracket.

Source: Dotabuff heroes table, All Pick, last seven days, read on October 25, 2026. Community data, not Valve data.

OpenDota’s heroStats feed splits picks by bracket as well, and it shows the same shape. Her win rate there is 50.27% in Herald (5,385 picks), 52.28% in Guardian (20,966), 52.05% in Crusader (46,294), 51.82% in Archon (83,831), 51.84% in Legend (109,667), 51.57% in Ancient (70,393) and 50.51% in Divine (53,703), against bracket medians of 49.14% to 49.83%. The gap is positive in all seven brackets OpenDota reports, between +0.68 in Divine and +3.11 in Guardian. In Turbo she wins 52.94% over 182,939 picks. OpenDota’s Immortal bracket had no data in the feed we pulled, so the Dotabuff table above is the only source for that rank.

Why the Edge Disappears at Immortal

Look at the last two columns of the Dotabuff table together. Her pick rate more than doubles from Herald (6.51%) to Immortal (14.01%), and her win-rate gap shrinks from +2.99 at Guardian to +1.68 at Divine to +0.14 at Immortal. At the top of the ladder she goes from a clearly winning hero to an average one while being drafted more often than at any other rank.

The data cannot tell us why, so what follows is our interpretation. Mirana’s strengths are the kind that low and mid brackets struggle to answer: a long-range stun that punishes standing still, a gap-closer and escape on one button, and Moonlight Shadow in games where nobody buys detection. Immortal players buy the detection, break the arrow line and draft the heroes that handle her (see the matchup section). Pick rate rising while win rate falls is also what you see when more players reach for a comfortable hero than the matchup supports.

Practical read: below Divine, Mirana is a hero you can pick on reputation. At Immortal, pick her when the enemy draft has no illusion cores or tanks who ignore a single stun, not by default.

Her Kit With Current Numbers

All values below come from Valve’s official hero data feed for patch 7.41f. Mirana is an agility hero with 20 base Strength (+2.2 per level), 26 base Agility (+3.1) and 22 base Intelligence (+1.9). She attacks at range 630 with a 1.7 base attack time, a projectile speed of 900 and 285 movement speed. At level 1 that gives 560 health and 339 mana by the standard formulas.

Aghanim’s Scepter makes Sacred Arrow release a Starstorm on enemies along the arrow’s path (search radius 500), with an additional 80% Starstorm on the unit the arrow hits. Aghanim’s Shard adds one Leap charge and grants three temporary Quiver charges while the Leap buff is active. Valve’s current hero data lists no facets for her, and Dotabuff’s facet panel returned “No data available” when we pulled it.

Sacred Arrow: What the 7.41f Stun Nerf Costs

Sacred Arrow’s text says the stun grows by 0.5 seconds for every 150 units of travel, up to a level-based cap, and that bonus damage grows with distance up to a maximum of 180. The arrow flies at 900 units per second. From those numbers we can work out exactly where the 7.41f nerf bites. This is our arithmetic on Valve’s ability text, so check it in a demo lobby before you rely on a decimal.

The nerf only matters if you hit from far away. A cap of C seconds is reached at 300 times C units. The table shows the distance where each level’s cap is reached before and after 7.41f.

At a distance under 690 units, an arrow does exactly what it did before the patch. A level 1 arrow at 600 units stuns for about 2.0 seconds both before and after. A level 4 arrow is untouched at every distance. The people hurt are those who max-range a level 1 to 3 arrow: they lose 0.3, 0.2 or 0.1 seconds, which is 12%, 6% and 2% of the old maximum. Most of that loss lands in the laning stage, where you have one point in the arrow and play at long range anyway.

The second half of the story is the arrow’s travel time. Here is the relationship between distance, stun and flight time at level 4.

A full-range hit is a 1.67 second flight after a 0.5 second cast point, so it is a prediction shot, not a reaction shot. That is why root effects matter so much for her.

Rod of Atos plus Sacred Arrow is guaranteed by the numbers. Rod of Atos roots for 2 seconds at a cast range of 1100. If you fire the arrow the instant the root lands, it leaves your bow after the 0.5 second cast point and needs 1.22 seconds to cover 1100 units, so it arrives at 1.72 seconds, with 0.28 seconds of root to spare. Against a rooted target at 1100 units, that is a 3.7 second stun and 132 bonus damage. The window closes at 1350 units, which is beyond Atos’s cast range, so anything you can root is inside it. This assumes you are already facing the target and that they do not pop Black King Bar, a dispel or a Linken’s Sphere on the Atos.

Also remember what the ability says: the arrow stops at the first enemy unit it strikes and instantly kills the first non-ancient creep it hits. Firing through a creep wave wastes it. And it can hit invisible units, which makes it one of the few answers to an enemy who dodges your vision.

Celestial Quiver and Leap: The Charge Economy

Since 7.41 Mirana is a charge hero. Celestial Quiver adds bonus magic damage to your next attack while you have a charge, and it replaced the old Selemene’s Faithful innate. The 7.41f nerf slows the recharge, so it is worth knowing what a charge is worth.

Health and mana in this table come from the base values and per-level gains with no items or talents. Max charges follow Valve’s note of 2 plus 1 every 6 levels, so we read it as an extra charge at levels 6, 12, 18, 24 and 30.

What 7 seconds per charge costs. At 6 seconds a charge she regained 10 charges a minute, at 7 seconds she regains 8.6, a 14% drop in sustained recharge. In a lane, where you trade one charged attack at a time, that is a slower harass loop. In a teamfight it matters far less, because you arrive with a full set of charges whichever number is on the patch notes.

The Shard burst. Valve’s notes say that casting Leap with the Shard grants 3 temporary charges for the duration of the buff, that these ignore the max charge count and that they stack from each Leap cast. With the Shard she has three Leap charges, so one fight can start with several casts and a pile of empowered attacks. At level 20 with the +50 talent, each charge is worth 115 magic damage, so three temporary charges from a single Leap are 345 damage on top of the charges you already had. This is arithmetic on the patch text, not a measured damage log.

Leap itself covers 650 units (800 with the level 10 talent) at a speed of 1300, so the jump takes half a second. Charges come back every 45/35/25/15 seconds depending on Leap’s level. At level 1 that means two jumps and then a long wait, with only a 6% speed bonus. At level 4 a charge returns every 15 seconds and the buff is 24% movement speed and 100 attack speed for 5 seconds, which is a real damage steroid with the attack speed talent (another +100 at level 15).

Starstorm and Moonlight Shadow

Starstorm is her wave-clear and her anti-illusion tool. It hits every enemy in a 675 radius for 75/150/225/300, then a second meteor hits the closest enemy to Mirana for 80% of that. Against a single target at level 4 that is 300 plus 240, or 540 damage for 110 mana on a 12 second cooldown. With the level 20 talent (+200) the same cast is 500 plus 400, or 900. It is not blocked by Linken’s Sphere and it cannot hit invisible units, so a Riki or an invisible Clinkz that you have not revealed stands in the middle of the storm unharmed.

The cooldown is 12 seconds, or 10 with the level 10 talent, so in a lane you can cast it on nearly every creep wave.

Moonlight Shadow is the reason people draft her. It turns the whole allied team invisible for 18 seconds, with 9/12/15% movement speed and 9/12/15% extra outgoing damage for Mirana herself. If a hero is revealed, invisibility returns after the fade delay (2.5/2/1.5 seconds) as long as the 18 seconds have not run out, and affected units do not auto-attack. The cooldown is 120/110/100 seconds, or 70 with the level 25 talent.

Two practical uses that win games at every rank: starting a Roshan or a high ground push in the dark so the enemy has to buy detection mid-fight, and the late-game save when the enemy team has to commit sentries to find you. If the other team has an answer to invisibility in the draft, her ultimate loses most of its value, which is one reason she looks weaker at the top of the ladder.

Talents

Dotabuff’s talent panel shows pick rate and win rate for each option when you hover over it. The win rates are inflated: the games that reach level 25 are usually games her team was already winning. Read the pick rate for what players do and the win-rate gap between the two choices as a loose hint, not as proof.

Source: Dotabuff talent usage, read on October 25, 2026. Community data.

The public consensus is clear at levels 20 and 25. Nine in ten players take +200 Starstorm damage and the Multishot arrows. The level 25 Multishot talent adds two extra arrows in a cone 25 degrees to each side, and they cannot hit the same target as the main arrow. Level 10 and 15 are the real choices. The cooldown on Starstorm wins the pick rate at 10, and the +20% evasion from Moonlight Shadow at 15 works at all times during the buff. We would take the Leap talents (+150 distance, +100 attack speed) in a carry-style game where you chase, and the cooldown and evasion in a support-style game where you cast. Note that the +50 Quiver talent is picked by under one player in ten, even though it is her only Quiver upgrade, and that is a gap worth testing in your own games.

Skill Order

Dotabuff’s most popular ability build, used in 12.38% of her games this week, starts with Sacred Arrow at level 1 and Leap at level 2, then puts the next three points in Starstorm and takes Moonlight Shadow at 6 before returning to Starstorm. Sacred Arrow gets its remaining points after that. We are not listing every level because the build panel’s level numbers did not line up cleanly with the talent levels when we read it.

The logic behind that order holds up on the numbers above:

• Sacred Arrow at 1. It instantly kills the first non-ancient creep it hits and it is your only lane stun. Costs 90 mana at every level, so extra points buy damage and stun length, not efficiency.
• Leap at 2. One point gives you a 650 unit escape on a 45 second charge. The speed bonus (6%) is small, and the point is the jump.
• Starstorm at 3, 4, 5. Level 3 already hits for 225 plus a second meteor of 180, and it is her wave-clear. Mana climbs from 80 to 110, so watch the pool.
• Moonlight Shadow whenever a rank unlocks (levels 6, 12 and 18). The cooldown improves from 120 to 100 seconds across ranks, and the fade delay goes from 2.5 to 1.5 seconds.

Item Build: What the Public Buys

This is Dotabuff’s most-used items list for Mirana over the last seven days, with Valve’s item numbers for what each one does. Win rates here are not a ranking of item strength: expensive items are bought in games that lasted long enough to afford them, and a winning team has the gold. The 38.41% on wards is the opposite effect, a support selection bias.

Source: Dotabuff most-used items, last seven days, read on October 25, 2026. Community data. We left out a few entries (Smoke of Deceit and two others that are not standard shop items).

What this adds up to is a support-oriented build. Magic Wand and Arcane Boots are the early staples, and the two Aghanim’s upgrades are the next spend. We rate the Shard as the better value of the two: one more Leap charge and a big Quiver spike for 1400 gold. Scepter at 4200 is the long-game purchase. It matters that Scepter is the single most-bought item at 224,343 matches, more than the Magic Wand: the arrow with Starstorms along its path is a real power spike, so it is worth planning your game around reaching it.

Our reading of how to adapt it by role, and this is advice and not data:

• Position 4 (soft support): Magic Wand, Arcane Boots, Shard first, then Rod of Atos or Force Staff for the setup, Glimmer Cape or Eul’s for saves, Scepter after that.
• Position 3 (offlane, 55.85% of her games): same early game, but add Eul’s and Atos for pickoffs. At 381 GPM in the offlane on Dotabuff she is not a farm-heavy hero, so cheap impact items matter more than big late-game purchases.
• Mid or safe lane (11.82% and 23.81% of her games): prioritize the Shard, since Quiver damage scales with your level and a fast Leap-and-attack rotation is the damage source.

Lanes and Positions

OpenDota tags Mirana as Carry, Support, Escape, Nuker and Disabler. Dotabuff’s lane panel shows where she is actually played.

Source: Dotabuff lane presence, last seven days. A caution: weighted by presence, the four lane win rates average about 49.0%, while her overall rate on the same site is 51.5%. The panel’s numbers do not reconcile, so treat the ordering as the signal and the exact decimals as approximate.

The offlane is where she lives, with more than half her games and the best win rate of the four lanes. The mid lane shows the highest KDA (3.50) and by far the highest GPM (475), which fits a hero with Starstorm wave-clear and Quiver scaling in a solo lane. It also has a lower win rate than the offlane. We would not read much into that gap with the reconciliation problem above, but it is a reason not to treat mid Mirana as a hidden strength.

Game Plan: Early, Mid and Late

Early game (levels 1 to 6). The job is to get Quiver charges into the lane trade and keep mana for the arrow. Auto-attack when a charge is up, back off when it is spent, and use the 7 second recharge as your rhythm. Starstorm kills creep waves and punishes a hero who stands next to them, but the second meteor targets the closest enemy to you, so it is stronger when you are the one standing in front. Do not spend your Sacred Arrow on a creep wave: it dies on the first creep it meets.

Mid game (levels 6 to 20). This is when Moonlight Shadow gets its value. The right use is initiating: wait until the enemy has farmed spread out, turn the team invisible and walk into a tower or a Roshan pit. Look for the arrow on a distance greater than 700 units from a safe position (at level 4 that is a 2.3 second stun or more). If the enemy has no detection, push everything. If they do, keep the ultimate for a defensive use.

Late game (levels 20 and up). Level 20 gives her the +200 Starstorm talent (500 plus 400 on a lone target) and level 25 the Multishot arrows. Your role in a teamfight is to open with the arrow on whoever the enemy needs most, chain it into Starstorm and use the Shard Leap burst on the second target. In our view her damage scales by hero level (Quiver) and by item timings (Scepter) but has a lower ceiling than a hard carry, so she wins by getting a decision and closing before the enemy catches up.

Matchups

The table below is built from Dotabuff’s Mirana counters page. We show the heroes with at least 25,000 matches in the sample, so every row is a large one. The win-rate column is Mirana’s win rate against that hero. Dotabuff’s own “advantage” ordering is a different statistic, so we use the raw win rate to avoid mixing the two.

Source: Dotabuff Mirana counters page, read on October 25, 2026, default date window, heroes with at least 25,000 matches. Community data.

The worst-matchup list has a clear theme, which is our reading and not a Dotabuff statement. Phantom Lancer (illusions) and Wraith King (a tank who reincarnates) are heroes that do not care much about a single stun, and that is the one reading we are comfortable offering. We would not guess at the reasons for the rest. Meepo is also at 45.74%, but on only 17,048 matches, so we leave him out of the table. On the strong side, Shadow Fiend at 59.23% and Nature’s Prophet at 58.79% are the standouts.

One more number worth knowing: her win rate against Kez is 57.20% on 39,481 matches, so if that pick shows up in your lobby the history is on your side.

How to Play Against Mirana

If you are on the other side of the arrow, the numbers above are a checklist.

• Make the arrow short. A stun from under 300 units lasts about a second. Closing distance on a Mirana that wants to snipe you is a legitimate counter, and a hero who stays at the edge of the map and gets hit at 1500 units takes a five second stun and 180 bonus damage.
• Break the line. The arrow is 115 wide and travels at 900, so at long range you have well over a second to step out of it. At full range that is 1.67 seconds of flight after the 0.5 second cast.
• Buy detection before the ultimate. Moonlight Shadow lasts 18 seconds and her whole team is invisible. A Sentry or a Dust turns a good initiation into a fight on your terms, and revealed heroes re-fade only after the fade delay.
• Do not stand next to a creep wave. Starstorm’s second meteor hits the closest enemy to her. A hero that stays in the wave is the one that gets hit twice.
• Know what does not work. Starstorm is not blocked by Linken’s Sphere. A Black King Bar answers the stun (Valve’s item text: negative effects from enemy spells have no effect), and the matchup table shows which heroes win the fight without one. Our hero win rates page has the full ranking if you want to draft against her.

What the Pro Data Shows

OpenDota’s per-hero match feed lists the last 100 league games where Mirana was played. The sample runs from October 12 to October 24, 2026. It is a mixed-tier sample: 43 of the games are DreamLeague Season 30 Qualifiers, 34 are EPL Masters 2026, 8 are NEXUS SERIES II and the rest are smaller events including the PARI UNIVERSE by FISSURE stage. These are not top-tier finals, so treat what follows as a read on the competitive scene below the biggest stage.

Source: OpenDota heroes/9/matches, last 100 league games, October 12 to 24, 2026. Community data, small samples per row.

Mirana went 49 wins and 51 losses over the sample, with an average game of 38.9 minutes and an average line of 4.4 kills, 5.8 deaths and 14.4 assists. Our read is that the high assist count points to a setup role rather than a kill-farming one. The length splits are noisy with 11 to 44 games per row, and we would not draw a conclusion from the 50-plus minute rate. The player with the most games is listed by OpenDota as Se of Inner Circle x Insanity, with 18 games and 11 wins, followed by kreker of Team Lynx with 10 games and 6 wins.

The pro and public pictures do not match. Public players see her as a clear winner at 51.5%, and the pro sample sits at an even 49%. We cannot separate the causes from this sample: the mixed tier of the games and the small number of games are both in play. For what it is worth, it points the same way as the Immortal row earlier in the article.

The Bottom Line

Mirana in 7.41f is a good public hero who is easy to pick and harder to pick well. Nothing in the September 15 patch changed that: a slower Quiver and a shorter long-range stun are real, but they hit the margins of her kit. The part of her game that wins ranked games is still intact: Moonlight Shadow initiations, the Shard Leap burst, Starstorm in lane and a Rod of Atos plus Sacred Arrow combo that cannot miss inside 1100 units.

If you play her, we would do four things. Spend your arrow at range and never through creeps. Buy the Shard early. Plan the game around a Moonlight Shadow engagement. And at Immortal, check the enemy draft for Phantom Lancer, Wraith King and detection before you pick.

If you want help turning hero knowledge into a higher bracket, our MMR boosting guide explains how rank services work and our breakdown of whether Dota 2 boosting is safe covers the risks. You can also compare services in our best Dota 2 boosting services review. For more 7.41f hero breakdowns, see our Hoodwink guide, Slardar guide and the full hero guide index.