What Is a Good KDA in Dota 2? By Role and Rank

What KDA Means in Dota 2

KDA stands for Kills, Deaths, and Assists. Every Dota 2 match produces three raw numbers per player: how many enemy heroes you killed, how many times your own hero died, and how many kills you assisted by being nearby or dealing damage. These three numbers appear on the post-game scoreboard.

The KDA ratio condenses those three numbers into a single figure:

KDA = (Kills + Assists) / max(Deaths, 1)

The “max 1” denominator prevents division by zero when a player finishes a game without dying — which does happen, especially on support heroes who play very conservatively or in stomp games that end quickly.

A KDA ratio of 3.0 means that for every death, you contributed to 3 enemy kills or assists. A ratio of 1.0 is break-even. Anything below 1.0 suggests your deaths outpaced your contributions. The higher the ratio, the more you contributed relative to the resources you gave away when dying.

One important note: Dota 2 does not display this ratio anywhere in the client. The game shows raw K/D/A numbers on the scoreboard. The ratio itself is a community-derived calculation, popularized by third-party tracking tools like Dotabuff and Stratz. When Dota 2 players discuss a “good KDA,” they almost always mean this computed ratio — not the raw kill, death, and assist counts on their own.

How Dota 2 KDA Differs From Other Games

In games like League of Legends, the KDA ratio is displayed in the client and tracked prominently. In Overwatch 2, kill-to-death ratio is a headline stat. Dota 2 deliberately takes a different approach: the game distributes assists broadly and does not reward the killing blow with additional gold beyond the base kill bounty. Multiple heroes can receive assist credit for the same kill, which inflates aggregate KDA numbers relative to games where only one player receives credit.

This means a Dota 2 KDA ratio is structurally higher than a League of Legends KDA for the same quality of play. A Dota 2 support accumulating 25 assists in a match produces a ratio that looks dramatic on paper — but reflects a system designed to include supports in the numbers, not superhuman individual skill.

Context is therefore essential when interpreting KDA in Dota 2. The number only means something in relation to the role played, the rank of the lobby, and the heroes involved.

Why KDA Is Not the Whole Story

KDA is a useful sanity check but a poor summary statistic for Dota 2 performance in isolation. Here is what it consistently fails to capture.

Farm Efficiency Predicts Outcomes Better Than Kill Counts

A carry who finishes 4/1/2 but hits a 14-minute Battle Fury or 15-minute Manta Style is often more valuable than a carry who finishes 12/4/10 with 6,000 net worth at the 20-minute mark. Dota 2 has a gold economy that translates directly into item power: last hits per minute, GPM, and first item timing predict game outcomes more reliably than kill counts at every bracket below Divine. The player with fewer kills but better farm timing regularly outperforms the player who fought more.

Deaths Have Asymmetric Consequences

In Dota 2, dying at minute 5 and dying at minute 40 are not equivalent events. A support dying at minute 3 defending a tower costs the team roughly 100 gold and 40 seconds of respawn time. A carry dying at minute 40 with a full inventory costs the team thousands of gold in lost buyback capacity, several minutes of respawn time, and potentially the Aegis. KDA treats every death as equal. They are not. A player who dies once at minute 38 with a Rapier has effectively lost the game for their team in ways that a 2/1/3 mid at minute 10 never would.

Some Heroes Are Designed to Die Productively

Heroes like Axe, Bristleback, Tidehunter, and Clockwerk are built to initiate into enemy teams and absorb damage to set up kills for cores. A 4/6/18 KDA on Axe in a game where his team won every teamfight because of his initiations represents excellent play. An Axe player who never dies in such a game is probably too passive to serve his role. The deaths were productive investments, not mistakes.

Support Contributions Are Mostly Invisible to KDA

Warding, dewarding, smoke initiations, courier management, regen sharing, pulling, stacking, saving cores with Force Staff or Glimmer Cape — none of these contribute to KDA in any way. A Pos 5 who places 18 observer wards across 40 minutes and uses Glimmer Cape to save the carry three times might finish 0/4/16 — a ratio of 4.0 — while the reason those 16 assists happened is entirely because of their vision and save utility, not despite their death count.

What Is a Good KDA by Role?

The benchmarks below are community estimates derived from public Dotabuff and Stratz match data across a broad sample of ranked games. They describe what the top 30 to 40 percent of players achieve in each role at their bracket over a season of games — not absolute requirements for “good” play. Individual games will vary significantly based on hero selection, opponent quality, and game duration.

Position 1 — Hard Carry

The carry role centers on farming first and fighting second, at least until the first major item spike. Carries who fight excessively before items are complete often find themselves burst down quickly and feeding enemy cores. The ideal carry game involves efficient farming, minimal unnecessary deaths, and meaningful teamfight presence from the mid-game onwards.

• Below average: KDA under 2.0 (dying too often or contributing too little)
• Average: KDA 2.0 to 3.5
• Good: KDA 3.5 to 5.0
• Excellent: KDA 5.0+

A practical example of a solid carry performance: 10/4/6 produces a ratio of (10+6)/4 = 4.0. That reflects a player who contributed meaningfully to fights and managed their deaths at an acceptable level given the farm total. At Herald and Guardian, community data suggests average carry KDA sits around 2.0 to 2.5 because positional discipline is still developing. At Immortal, top carry players regularly hold seasonal averages between 4.0 and 6.0.

Position 2 — Mid Lane

Mid players are expected to fight more than carries and rotate early to support side lanes. Rune control, pressure on towers, and early rotations to assist the offlane or safe lane are all part of the mid role’s job. This pushes their raw kill and assist counts up but also raises death counts because they are in the action consistently.

• Below average: KDA under 2.5
• Average: KDA 2.5 to 4.0
• Good: KDA 4.0 to 6.0
• Excellent: KDA 6.0+

A mid player finishing 14/5/12 ends with (14+12)/5 = 5.2 — solidly in the good range. A 7/7/8 scoreline at (7+8)/7 = 2.14 suggests they are dying at an unsustainable rate relative to their contribution. Mid players benefit particularly from heroes with strong escape mechanisms — Storm Spirit, Windranger, Queen of Pain — that allow players to take fights without committing to unfavorable trades. Learning when to take favorable fights versus when to retreat is the primary lever for improving KDA in this position.

Position 3 — Offlane

Offlane is where KDA expectations drop most significantly. The role exists to pressure the enemy safe lane, absorb resources from the enemy team, create space for the carry to farm, and initiate in teamfights. Dying as an offlaner is frequently correct — if an Axe is being chain-initiated onto while his carry farms three camps uncontested, the offlaner is doing exactly the right thing.

• Below average: KDA under 1.5
• Average: KDA 1.5 to 3.0
• Good: KDA 3.0 to 4.5
• Excellent: KDA 4.5+

Context is especially important for this position. An offlaner finishing 2/6/14 with a massive building-damage number and a 72% kill participation rate has almost certainly outperformed an offlaner who finished 5/2/4 by playing too safely and creating no space. Do not judge an offlaner by KDA without also checking how much space their presence created and how many teamfights they initiated successfully.

Position 4 — Soft Support (Roaming Support)

The soft support role centers on roaming to create early kills, enabling core heroes through kill setups, and maintaining vision across the map. Kill participation is frequently a more useful metric for this position than raw KDA, because a pos 4 who is present for every fight but is not the last hitter will accumulate assists rather than kills.

• Below average: KDA under 2.5
• Average: KDA 2.5 to 4.0
• Good: KDA 4.0 to 6.0
• Excellent: KDA 6.0+

A pos 4 finishing 2/3/22 ends with (2+22)/3 = 8.0, which usually reflects a stomp game where the team snowballed and assists were easy to accumulate. A more representative strong performance in a close game might be 4/4/18 at (4+18)/4 = 5.5 — reflecting a player who was in every relevant fight and managed their deaths acceptably despite aggressive roaming.

Position 5 — Hard Support

Hard supports should theoretically post the highest KDA ratios on paper because they assist on nearly every kill while also prioritizing survival to maintain long-term utility. A support who dies repeatedly in the laning phase loses the ability to ward, save, and peel for their cores — which are the actual contributions that win games.

• Below average: KDA under 2.0
• Average: KDA 2.0 to 4.0
• Good: KDA 4.0 to 7.0
• Excellent: KDA 7.0+

A strong hard support performance might look like 1/2/24 — ratio of (1+24)/2 = 12.5. That reflects a player who accumulated assists throughout a long game, died only twice, and was present at nearly every kill on the team. That player almost certainly drove significant impact through vision control, saves with utility items, and enabling core heroes to take safe fights. The raw number is impressive but the story behind it is the support playing within their role.

These are community estimates based on aggregated Dotabuff/Stratz match data and coaching observations. Individual games will vary widely based on hero, opponents, and game duration.

What Is a Good KDA by Rank?

KDA norms shift considerably as MMR increases. In lower brackets, chaotic games with constant fighting produce high kill counts but also many deaths. In higher brackets, games become more deliberate, deaths are more consequential, and players understand when not to take fights. The result is that the “average” KDA number climbs alongside rank — but so does the standard for what “good” looks like.

Herald and Guardian (0 to 2,000 MMR)

Games at this bracket are characteristically chaotic. Players fight frequently regardless of vision or advantage, warding is inconsistent, and fights often extend far beyond the point where one team clearly has the advantage. Death counts are high across the board. Kill counts are also inflated because fights are long and messy with many players able to secure assists on the same enemy.

Community estimate of average KDA across all roles: approximately 2.0 to 3.0.

A player holding a 4.0 KDA consistently in Herald or Guardian is performing genuinely above the average of their lobby. They are dying less than teammates and contributing to more fights — a reliable signal that their positioning and game sense exceed the bracket average even if mechanical execution is still developing.

Crusader and Archon (2,000 to 3,500 MMR)

This is often described as the “average” Dota 2 player range in global terms. Players at this bracket have more map awareness than Herald, understand roles better, and die slightly less per game. Warding is more consistent, rotations happen with some coordination, and game sense starts to separate players more clearly from one another.

Community estimate of average KDA across all roles: approximately 2.5 to 3.5.

A 4.0+ KDA at Crusader or Archon is a strong indicator that you are above average for your bracket and playing in a way that reflects better decision-making than most opponents. Players who sustain this over 50+ games are typically on an upward MMR trajectory. If you are maintaining 4.0+ KDA but not climbing, the issue may be hero pool consistency, communication, or loss-streak patterns rather than individual performance.

Legend and Ancient (3,500 to 5,000 MMR)

At this range, mechanical understanding is solid and game sense is developed enough for players to make coherent decisions under pressure. Rotations happen efficiently, drafts are more intentional, and deaths feel more punishing because teams are better at exploiting them. Kill participation becomes a stronger differentiator as players understand when fights are worth taking.

Community estimate of average KDA across all roles: approximately 3.0 to 4.5.

Players climbing through Legend and Ancient tend to show more consistent KDA because they have fewer impulsive death decisions and more intentional fight engagement. Outlier games in either direction become less common with experience at this bracket. A Legend or Ancient player aiming to climb should target sustained 5.0+ KDA over a hero pool of two or three consistently played heroes.

Divine (5,000 to 6,000 MMR)

Divine is where players have strong intuition about the entire map state — what the enemy is doing, where they are likely to rotate, and when fights are fundamentally unwinnable. Farm patterns are efficient, item timings are consistent, and deaths are increasingly rare because players know which situations to avoid.

Community estimate of average KDA across all roles: approximately 3.5 to 5.5.

A Divine player whose seasonal KDA falls below 2.5 is almost certainly either playing a position that does not suit their strengths or taking too many unfavorable fights out of habit. Above 5.0 is solid for this bracket. Above 7.0 indicates play that is genuinely above the Divine average and likely trending toward Immortal if win rates are consistent.

Immortal (6,000+ MMR)

At Immortal, every death is a decision that carries weight. Players at this level have internalised enough game knowledge that unnecessary deaths are uncommon events rather than regular occurrences. Games are coordinated at a higher level, fight selection is sharp, and players understand the exact moment when an engagement becomes unfavorable.

Community estimate of average KDA across all roles: approximately 4.5 to 7.0+.

Top 1000 Immortal players frequently hold seasonal averages above 6.0. The very top of the leaderboard — the players competing for the highest-ranked spots — commonly post 8.0 or higher across hundreds of games. At this level, the KDA number reflects not just survival skill but the ability to be present and impactful in every fight while avoiding the unnecessary risks that cost lower-bracket players games.

All figures are community estimates derived from public Dotabuff/Stratz data and coaching observations, not figures officially published by Valve.

Kill Participation — The Better Metric

While KDA ratio gets the most attention in discussions about individual performance, kill participation rate is often a more reliable indicator of whether you are genuinely impacting your team’s success — particularly for support players.

Kill participation rate = (Kills + Assists) / Total Team Kills x 100%

If your team secures 30 kills in a game and you finish with 8 kills and 15 assists, your kill participation is (8+15)/30 = 76.7%. That is a strong number for any role. Kill participation removes the denominator problem of KDA entirely — it does not punish you for dying. It rewards presence in fights.

Kill participation is valuable specifically because it connects your individual performance to team outcomes. A carry who farms safely but is absent for every teamfight might have a reasonable KDA ratio but a 35% kill participation — indicating they are not contributing to the fights that determine the game’s direction. A support with a modest KDA ratio but 85% kill participation is almost certainly playing excellent Dota.

Dotabuff tracks kill participation as a percentage on hero profile pages and player statistics screens. Most high-impact players hold kill participation above 65% for their role across a season of games.

Kill participation benchmarks are community estimates. Hard supports in stomp games may exceed 90%; in close or defensive games they may sit lower without this reflecting poor play.

Heroes Where KDA Is Particularly Misleading

Some hero archetypes produce systematically skewed KDA numbers that do not reflect performance quality in a meaningful way. Understanding these categories helps you interpret your own statistics and those of your teammates more accurately.

Bait Initiators

Heroes like Axe, Tidehunter, Clockwerk, and Centaur are designed to dive into enemy teams and initiate fights that their cores then close out. A successful Axe Blink plus Berserker’s Call initiates a fight resulting in three enemy deaths and one team member dying — a spectacular trade. But if Axe also gets burst down during the execution window, his KDA suffers even though he made exactly the right decision.

Evaluating a Tidehunter or Axe by KDA alone is similar to evaluating a goalkeeper by goals scored. The metric simply does not capture the contribution the role is designed to make. For these heroes, look at teamfight win rate, building damage, and kill participation instead.

Global Ability Carries

Heroes like Zeus, Sniper, and Silencer accumulate kills through global or long-range ultimates that allow them to secure kills from across the map without meaningful positioning risk. A Zeus with 17 kills in a match often secured several of them by firing Thundergod’s Wrath at enemies already at low health from teammates’ damage. The KDA ratio looks exceptional but the player did not necessarily outplay opponents to earn each kill.

Comparing a Zeus game to a Juggernaut game on the basis of KDA is meaningless. Zeus’s kit inflates raw kill numbers structurally. The more useful question for Zeus is whether their net worth is high enough to cast spells at critical moments and whether their skill timing won fights that would otherwise have been lost.

Zone-Out Supports

Heroes like Ancient Apparition, Keeper of the Light, and Crystal Maiden played at position 5 frequently contribute through long-range harassment, debuffs, and global pressure rather than being present for every kill. Their kill participation is lower not because they play poorly but because their contribution is positional and preventive.

A Crystal Maiden who zones the enemy carry out of the lane for six minutes — preventing hundreds of gold in last hits — has made a massive impact that shows nowhere in any KDA-adjacent metric.

Tower-Focused Offlaners

A Pugna, Nature’s Prophet, or Lycan played in the offlane can spend significant time rat-pushing side lanes and destroying towers while their team fights elsewhere. A game where they destroy three Tier 1 towers in 18 minutes and force constant enemy response might end with a 2/3/6 scoreline — low kill numbers because they were not in the teamfights. But their building damage contribution may have been the decisive factor in the game.

How to Improve Your KDA in Dota 2

If your KDA is consistently below the averages for your role and rank bracket, it usually points to one of a small number of addressable problems.

Reduce Preventable Deaths First

The majority of deaths in matches below Ancient are not the result of skillful enemy counterplay — they are positional mistakes. Moving into fog without vision, chasing a fleeing enemy into their base without teammates, diving towers with insufficient HP or support, and taking fights after a spell rotation that leaves you without important abilities. These are decision errors, not mechanical failures.

The fastest way to improve KDA is to reduce these deaths. Use the minimap actively every few seconds. Before moving into any position where you can be ganked, ask: where are the relevant enemy heroes? If the answer is “I do not know,” treat that position as dangerous until you do know. Building a habit of checking vision before moving is the single highest-leverage adjustment most lower-bracket players can make.

Increase Kill Participation

If you die rarely but your assists are low, you are probably farming safely while your team fights without you. This is common among carries who over-stay in jungle camps during midgame teamfights. Evaluate whether each camp you take is worth more than the fight your team is currently taking without you. Often it is not. When a fight initiates near you and your team has the numbers advantage, joining it converts one camp into multiple kills and assists that are each worth more than the camp gold.

Match Hero Choice to Your Skill Set

Some heroes are mechanically demanding in ways that punish players while they learn them. Playing Invoker, Meepo, or Arc Warden while still developing basic positional instincts means you are spending cognitive budget on mechanics instead of decisions — and dying as a result. A simpler hero that lets you focus on positioning, fight timing, and map reading often produces better KDA numbers and faster rank improvement than mechanically impressive heroes played sub-optimally.

Pick two or three heroes per role that you can play with less mental overhead. Track your KDA on those heroes over 30 games and compare it to heroes where you are still learning the kit.

Stop Chasing Kills

One of the most common patterns in brackets below Legend is chasing a fleeing low-health enemy hero across the map at the expense of taking a nearby tower, forcing a Rosh attempt, or simply getting back to farming. The kill is worth 150 to 300 gold. The tower is worth 150 to 250 gold plus structural advantage plus time saved for your team. Choosing the tower almost always produces better game outcomes — and often the chase leads to a death that drops your KDA rather than the kill that would have improved it.

When you feel the impulse to chase, ask whether your position after catching the enemy is safe. If reaching them puts you in fog, near their base, or away from your team without buyback, the correct answer is almost always to stop and take the nearby objective instead.

Work With a Coach

Replay self-review catches some errors but misses others entirely — particularly the errors of omission, where you should have moved somewhere but did not, or should have taken a fight but misjudged it as unfavorable. A coach watching live or reviewing a replay identifies these moments in real time and explains the correct decision in the moment, which is far more effective than reading about it abstractly.

Working with an Immortal-level Dota 2 coach through TeamSmurf gives you a structured breakdown of exactly where your deaths are coming from and what your positioning looks like relative to what it should be. Players who combine coaching with deliberate practice on a small hero pool typically see KDA improvement within 10 to 15 sessions of concentrated play.

If you are interested in understanding how your MMR correlates with your actual game performance, that guide covers how the rating system weighs wins, losses, and individual performance into your MMR trajectory.

Seasonal KDA vs Per-Game KDA

A single game where you finish 20/1/15 produces a KDA of 35.0 — but it says almost nothing about your actual skill level. Single games are dominated by noise: matchup luck, enemy mistakes, stomp conditions, or a particularly favorable draft. What matters is your seasonal average — your KDA across 50, 100, or 200 games on a consistent hero pool. That is the signal; individual games are the noise.

Dotabuff tracks seasonal and lifetime KDA per hero, per role, and per position. When you open a hero profile and see a “KDA Ratio” figure, it is the average over the selected time period. That number — especially over 50+ games — is the one worth monitoring and improving deliberately.

If you are averaging 4.5 KDA on a hero over 80 games, that is a meaningful indicator of consistent performance at your bracket. If you had one game at 22.0 and the rest averaged 1.9, your displayed average of 2.4 reflects variance, not improvement.

To build a meaningful seasonal KDA benchmark, play the same one or two heroes in the same position consistently for three to four weeks. Track the trend. If the seasonal KDA is rising, your approach is working. If it is flat or declining despite more games played, the approach needs deliberate adjustment — a new positioning habit, a different laning strategy, or a hero swap.

Players who combine a small hero pool with consistent role discipline almost always show cleaner seasonal KDA numbers than players who play five different roles across ten different heroes and have no controlled comparison. The data from your own Dotabuff profile is one of the most useful coaching tools available to you for free — and most players never look at it seriously.

If you have reached a point where your KDA is solid but your MMR is not moving, the problem may be win-rate variance rather than individual performance. A stretch of games where you play well but lose due to team-related factors can stall MMR for weeks. Some players in this situation choose to use a professional boost from TeamSmurf to bridge the gap between current performance and their actual bracket, then focus on maintaining the new rank through the skills they have already developed. Whether that makes sense depends on your specific situation, goals, and how much of the gap is team-related variance versus genuine skill gap.

The short answer to whether boosting is a safe approach in 2026 is that it depends entirely on the service. Is Dota 2 boosting safe in 2026 — that guide covers the actual risk profile for different service types and what to look for.