Dragon Knight
Our own data, from ranked matches above 7500 MMR.

Soprar Fogo
Libera um sopro de fogo na frente do Dragon Knight que queima e reduz o dano de unidades inimigas atingidas.

Cauda de Dragão
O Dragon Knight golpeia uma unidade inimiga com o seu escudo, atordoando e causando dano a unidades ao seu redor.

Ira da Serpe
O sangue vital do Dragão fortalece as habilidades e itens do Dragon Knight, concedendo-lhes um raio maior às áreas de efeito e fazendo os seus ataques causarem dano mágico adicional a unidades inimigas.

Bola de fogo
Ateia fogo em uma área, causando dano ao longo do tempo a inimigos que estiverem em cima. O alcance é reduzido quando em forma corpo a corpo.

Sangue de Dragão
O sangue vital do Dragão aumenta a regeneração de vida e fortalece a armadura do Dragon Knight. A regeneração e a armadura são aumentadas em 50% durante a Forma de Dragão Ancestral .

Forma de Dragão Ancestral
O Dragon Knight assume a forma do Dragão Ancestral, aumentando o alcance das suas habilidades, e recebendo velocidade de movimento adicional e um ataque de longo alcance com novas propriedades. O dragão evolui a cada nível e os efeitos são cumulativos. Nível 1 — Verde : concede dano corrosivo ao longo do tempo que também afeta estruturas. Nível 2 — Vermelho : concede dano em área aos ataques do dragão. Nível 3 — Azul : concede um ataque congelante que retarda as velocidades de movimento e ataque, ignorando imunidades a efeitos.
Skill order
1
2
3
4
5
6
7
8
9
10- 11
12
13
14
15- 16
18- 20
- 25
Talents
- +50% Wyrm's Wrath Bonuses81% · WR 64%25+12 Dragon Blood HP Regen/Armor19% · WR 66%
- +220 Breathe Fire Damage95% · WR 60%20+150 Elder Dragon Form Attack Range5% · WR 54%
- 15+0.4s Dragon Tail Stun100% · WR 56%
- -2s Breathe Fire Cooldown84% · WR 51%10+15 Damage16% · WR 50%
Items
Starting kits
2
2
29655%
2
3
23151%
2
3
21051%
Starting items
97%50%
87%51%
47%51%
33%52%
32%52%
31%51%
27%49%
24%49%
24%48%
Core, in the order it usually comes out
min 1395%52%+1.3 pp
min 2379%55%+2.4 pp
min 2951%64%+3.1 pp
min 3533%
min 3048%65%+3.4 pp
min 3533%
Situational
min 799%51%+0.6 pp
min 2183%54%+2.5 pp
min 2752%63%+3.6 pp
min 2824%
min 3029%
min 3428%67%+3.3 pp
min 3029%
min 3227%63%+3.6 pp
min 3421%65%+3.8 pp
min 2019%
min 1514%56%-1.9 pp
min 3018%
min 3619%
min 2514%59%+0.1 pp
min 3813%
Did this build work for you?
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.
Trend in the 7500+ ladder
4,332 ranked matches in the last 8 weeks
- Win rate
- 50.3%
- Pick rate
- 26.2%
Rating: 70% win rate (adjusted for sample size) and 30% pick rate, compared with heroes in the same position. S is the top 10%, A the next 20%, B 30%, C 25%, D the rest.
Off-meta that works: Dragon Knight with Manta Style (+11.7 pts), Dragon Knight with Orchid Malevolence (+7.2 pts), Dragon Knight with Radiance (+5.2 pts)
Dragon Knight worldwide
OpenDota, public ranked matches from the last 7 days
Win rate
52.8%
497,280 matches
Pick rate
14.1%
of matches include the hero
Pro scene
66 picks and 51 bans
53.0% win rate when picked
Win rate by rank bracket
- Herald and Guardian53.2% · 29,377 matches
- Crusader and Archon52.9% · 151,926 matches
- Legend and Ancient52.7% · 217,793 matches
- Divine and Immortal52.3% · 66,313 matches
Win rate by match duration
- up to 25 min48.2% · 4,778 matches
- 25 to 35 min48.9% · 12,255 matches
- 35 to 45 min49.9% · 7,089 matches
- over 45 min48.4% · 3,650 matches
Advantage in points over Dragon Knight'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.
Worst counters
Enigma39.8% win rate in 327 games · our high-MMR matches-10.7
Visage42.1% win rate in 114 games · our high-MMR matches-8.3
Slark43.0% win rate in 507 games · our high-MMR matches-7.4
Io43.3% win rate in 180 games · our high-MMR matches-7.1
Brewmaster44.0% win rate in 193 games · our high-MMR matches-6.4
Nyx Assassin44.9% win rate in 443 games · our high-MMR matches-5.5
Doom45.5% win rate in 336 games · our high-MMR matches-4.9
Techies45.8% win rate in 565 games · our high-MMR matches-4.6
Best synergies
Enigma62.4% win rate in 189 games · our high-MMR matches+12.0
Visage62.4% win rate in 947 games · STRATZin our high-MMR matches: 59.2% in 49 games+9.7
Ursa58.3% win rate in 120 games · our high-MMR matches+7.9
Elder Titan60.2% win rate in 601 games · STRATZin our high-MMR matches: 51.2% in 43 games+7.6
Bounty Hunter57.0% win rate in 363 games · our high-MMR matches+6.6
Slark56.7% win rate in 335 games · our high-MMR matches+6.3
Primal Beast58.4% win rate in 1,357 games · STRATZin our high-MMR matches: 43.4% in 53 games+5.8
Timbersaw55.8% win rate in 190 games · our high-MMR matches+5.4
Dragon Knight: too strong to play scared, too slow to hunt heroes
Across the 15 Dragon Knight lessons in the catalog, the coach repeats the same idea to players from very different brackets: "you're really strong and you don't understand how strong you are". The mistake shows up in two opposite ways. In lane, the player trades with fear. After level 6, with the dragon and the lead in hand, he goes hunting heroes and stops making gold. The root is the same: DK's strength comes from trading in lane and from damage on buildings, not from the hunt.
In lane, trade like a Morphling
The regeneration from the innate ability defines DK's laning stage. The coach compares him to Morphling: trade every wave, go back to full HP and trade again. Walk up to the hero, Dragon Tail, two or three hits and back off.
In the review of an Immortal DK against Bounty Hunter, the player was in a 2v1 against a hero with no boots and spent the lane trying to match the creep count, with 800 gold sitting unspent and two Bracers in a lane that dealt no damage. The fix: boots and Gloves of Haste instead of the second Bracer, and hitting the Bounty every time he walked up to a creep. In an unranked replay against Pudge, the diagnosis was the same: "you're playing like the Pudge is level 25".
The trade filter is the enemy's spells. Against a Timbersaw, who deals 300 pure damage every time, you don't trade. Against a Bounty Hunter, who hits you with a 170 damage Shuriken, you always trade. And two rules:
- Dragon Tail early when the lane has a hero who can die. In the review of a Legend 5 DK, the student only took the stun at level 8: "that's a crime in any country".
- Got a kill, buy regen. The same student died to a Phantom Assassin with 500 HP and 500 gold in his pocket: "proof you didn't buy regen".
The dragon is for hitting buildings
With Elder Dragon Form and no risk of dying, a good DK ults right away and goes to hit the tower. The poison leaves damage over time on the structure: one hit, step out, come back when the poison ends. If the tower aggros you, click one of your own creeps and the aggro switches without you leaving its range.
Taking the mid T1 doesn't end the job. The waves start meeting closer to their base, and you gain time for one more camp before you need to show up on the wave. In the lesson on a Divine 4 mid DK against Zeus, the student took the tower and sat at 6k gold when he should have had around 8,500, because he didn't use the space he had opened.
The items that aren't up for debate
In the patch 7.41f lessons, the item the coach demands most is Aghanim's Shard. In the Legend 5 and in an Ancient 1 mid DK against Shadow Fiend, the mistake was the same: BKB without Shard. "You're signing a contract: I won't farm creeps on DK anymore." The order the coach gives is Blink, Shard, BKB, and after that you adapt. BKB stays out of the first four items only "one in every 100 games"; in an Archon 4, the student swapped it for Monkey King Bar against three magic damage cores. The value of a fixed pattern is freeing your head for the map.
When the carry turns into a support
A DK with no farm starts chasing the "shiny dot" on the minimap. In the same Legend 5, the student had 97 last hits and spent his time hunting heroes in the middle of the map: "you promoted yourself to pos 4". In one fight he used Blink and stun on the Pudge, with 2,100 HP, while a 900 HP target stood right next to him.
The cycle for a Blink and stun hero: jump on the target that dies, kill it, step out into the fog, wait for Blink to come back and repeat. To find that target alone, the side lane pays off more than mid, where fights tend to be 5v5.
| Bracket | The mistake corrected in the replay |
|---|---|
| unranked | afraid to trade with a level 1 Pudge |
| Archon | leaving the wave to hunt kills; MKB instead of BKB |
| Legend | BKB without Shard; chasing heroes with 97 last hits |
| Ancient | mid DK without Shard, focusing the tank with a save behind him |
| Divine | taking the mid T1 and not using the space it opened |
| Immortal | with a lead, clicking forward instead of managing it |
With Aegis, the delay has to be the enemy's
In an Immortal replay, the one commenting was the carry from the other team. The student had Aegis, ult and items, and still walked around as a group chasing him while he cut waves on the other side of the map. "You're buying my delay." With Aegis, the rule is to push two waves and split the map: "you don't need five heroes behind you to hit my T2".
In the DK against Bounty Hunter, the student was hitting a tower with backdoor protection on while the Nyx Assassin cleared his wave, and he was also sharing camps with his own Necrophos. The coach's reminder: "you're going to finish BKB and their throne is going to explode". Until then, "carry mindset is making as much gold as possible".
Gold or buildings
Before every DK play, ask: does this end in gold or in a building? Hitting the tower with the dragon ends in a tower. Blink on an isolated support in the side lane ends in a kill and farm. Following the Pudge around the map, hitting a T2 with no wave and taking a 5v5 fight without BKB all end in wasted time. DK is too strong to play scared and too slow to play as a hunter.
Locked on exactly what you wanted?
The whole catalog (real match analysis, by hero, position and patch) opens with the subscription.
See plansBuild of the players who play Dragon Knight the most
7500+ ranked matches our radar captured in the last 60 days: accounts with at least 10 matches on the hero, and the build from their matches only. This is not the pro league: that one is in the hero's pros.
5 specialists, 282 matches, 55% win rate on the hero
- Specialist 167 matches, 57%
- Specialist 264 matches, 53%
- d1man41g54 matches, 57%
- Specialist 451 matches, 49%
- Specialist 546 matches, 59%
Items
Starting kits
2
2
2662%
2
3
2662%
2
3

1759%
Starting items
98%54%
73%58%
54%56%
44%56%
37%60%
31%58%
22%56%
21%44%
Core, in the order it usually comes out
min 1297%56%
min 2384%59%
min 2951%72%
min 3048%72%
min 3430%71%
Situational
min 799%55%
min 2188%57%
min 2653%70%
min 2423%59%
min 2622%62%
min 3514%66%
min 3013%58%
min 3611%69%
Skill order
Most played121 matches, 53% win rate
1
2
3
4
5
6
7
8
9
10- 11
12
13
14
15- 16
18- 20
- 25
Highest win rate10 matches, 80% win rate
1
2
3
4
5
6
7
8
9
10- 11
12
13
14
15- 16
18- 20
- 25
Who plays Dragon Knight in pro Dota
Pro leagues since patch 7.33. Recent matches also include 7500+ ranked games our radar watches live.
20 matches shown, 45% win rate
10 min 20 min NW 4.6k 13.2k LH 65 178 DN 9 Lane +1.1k 
















2 min
3 min
6 min
10 min
14 min
16 min
18 min
19 min
19 min
10 min 20 min NW 3.7k 9.6k LH 49 131 DN 3 Lane +0.7k 












2


3 min
7 min
8 min
9 min
12 min
15 min
22 min
25 min
10 min 20 min NW 4.7k 9.1k LH 65 132 DN 5 Lane +0.7k 












2


3 min
4 min
6 min
10 min
15 min
23 min
32 min
32 min
10 min 20 min NW 3.5k 8.4k LH 62 158 DN 9 Lane -0.2k 










3
2


3 min
4 min
6 min
7 min
19 min
27 min
10 min 20 min NW 5.2k 12.6k LH 89 255 DN 7 Lane +1.1k 









2
2


1 min
3 min
12 min
14 min
16 min
22 min
27 min
31 min
33 min
10 min 20 min NW 3.5k 8.8k LH 58 134 DN 11 Lane -0.6k 












2



3 min
3 min
8 min
14 min
17 min
25 min
34 min
10 min 20 min NW 3.6k 9.5k LH 61 143 DN 6 Lane +0.1k 









2

2


2 min
5 min
6 min
7 min
12 min
15 min
16 min
22 min
28 min
31 min
38 min
10 min 20 min NW 5.0k 10.1k LH 58 140 DN 8 Lane +1.3k 












2


3 min
5 min
10 min
15 min
22 min
29 min
29 min
33 min
10 min 20 min NW 3.5k 8.6k LH 55 148 DN 7 Lane -1.0k 

















4 min
6 min
10 min
13 min
16 min
25 min
10 min 20 min NW 4.5k 12.8k LH 69 208 DN 16 Lane +1.8k 











2



0 min
3 min
4 min
13 min
17 min
21 min
24 min
29 min
32 min
34 min
38 min
10 min 20 min NW 2.6k 7.9k LH 35 116 DN 2 Lane -2.1k 










3




3 min
4 min
5 min
8 min
15 min
16 min
23 min
32 min
32 min
37 min
44 min
48 min
48 min
48 min
48 min
10 min 20 min NW 4.2k 11.0k LH 67 151 DN 11 Lane +0.6k 












3


4 min
5 min
6 min
11 min
15 min
21 min
27 min
32 min
32 min
10 min 20 min NW 3.9k 9.1k LH 62 151 DN 9 Lane -0.6k 





![Sand King (No[o]ne-) Sand King](/dota/herois/icone/sand_king.webp)



2
3


2 min
4 min
6 min
12 min
15 min
22 min
24 min
28 min
37 min
10 min 20 min NW 4.7k 9.6k LH 69 159 DN 9 Lane +1.0k 
![Dragon Knight (No[o]ne-) Dragon Knight](/dota/herois/icone/dragon_knight.webp)








2

3

1 min
4 min
9 min
15 min
15 min
21 min
27 min
38 min
38 min
10 min 20 min NW 3.9k 12.1k LH 59 185 DN 5 Lane -0.9k 










3




1 min
4 min
9 min
15 min
17 min
21 min
23 min
25 min
28 min
31 min
35 min
40 min
40 min
47 min
51 min
56 min
10 min 20 min NW 4.8k 10.9k LH 78 213 DN 15 Lane +1.2k 









4



3 min
4 min
13 min
16 min
17 min
19 min
24 min
29 min
33 min
43 min
10 min 20 min NW 4.1k 10.0k LH 70 210 DN 10 Lane +0.9k 










2




3 min
4 min
15 min
19 min
20 min
26 min
34 min
37 min
40 min
40 min
10 min 20 min NW 4.6k 10.2k LH 70 140 DN 6 Lane +1.8k 
















4 min
6 min
10 min
15 min
22 min
28 min
10 min 20 min NW 4.7k 10.0k LH 83 171 DN 8 Lane +0.5k 
















0 min
2 min
4 min
16 min
22 min
25 min
29 min
34 min
10 min 20 min NW 3.6k 9.6k LH 54 133 DN 5 Lane +0.4k 










2





3 min
3 min
4 min
7 min
13 min
16 min
22 min
26 min
Pro league and 7500+ ranked (seen live by our radar), newest first. In the tree, the lit branch is the talent the player took. In ranked, last hits and denies come from the replay and show up a few hours after the match. Position is estimated from gold per minute within the team.
Patch
Position
Tournaments
Minimum matches
With the item at the end of the match
| # | Player | W x L | Last | |||
|---|---|---|---|---|---|---|
| 1 | skiter | 80 | 53 x 27 | 66.3% | 60.4% | 11/14/2026 |
| 2 | YroboroS | 19 | 16 x 3 | 84.2% | 60.2% | 11/25/2024 |
| 3 | Malr1ne | 54 | 37 x 17 | 68.5% | 60.1% | 11/13/2026 |
| 4 | No[o]ne- | 31 | 23 x 8 | 74.2% | 59.9% | 11/20/2026 |
| 5 | Хёма Чигири 10 место | 101 | 64 x 37 | 63.4% | 59.2% | 10/09/2025 |
| 6 | jikroy | 33 | 23 x 10 | 69.7% | 58.3% | 04/03/2026 |
| 7 | GODCOMPLEX | 109 | 67 x 42 | 61.5% | 58.1% | 08/08/2025 |
| 8 | gotthejuice | 39 | 26 x 13 | 66.7% | 57.6% | 02/14/2026 |
| 9 | gpk~ | 55 | 35 x 20 | 63.6% | 57.4% | 08/21/2026 |
| 10 | NothingToSay | 31 | 21 x 10 | 67.7% | 57.1% | 02/01/2026 |
| 11 | Nisha | 26 | 18 x 8 | 69.2% | 56.9% | 01/19/2026 |
| 12 | jmixheal | 42 | 27 x 15 | 64.3% | 56.7% | 06/05/2024 |
| 13 | Pure | 32 | 21 x 11 | 65.6% | 56.3% | 05/30/2026 |
| 14 | Self hate | 32 | 21 x 11 | 65.6% | 56.3% | 02/26/2026 |
| 15 | Young PH | 25 | 17 x 8 | 68.0% | 56.2% | 12/17/2025 |
| 16 | DarkMago♥ | 38 | 24 x 14 | 63.2% | 55.8% | 11/17/2025 |
| 17 | v1olent– | 15 | 11 x 4 | 73.3% | 55.5% | 01/17/2026 |
| 18 | bzm | 53 | 32 x 21 | 60.4% | 55.4% | 10/25/2026 |
| 19 | naive- | 21 | 14 x 7 | 66.7% | 55.0% | 11/11/2026 |
| 20 | shigetsu | 23 | 15 x 8 | 65.2% | 54.8% | 04/05/2026 |
| 21 | SLATEM$ | 23 | 15 x 8 | 65.2% | 54.8% | 10/20/2026 |
| 22 | casino | 23 | 15 x 8 | 65.2% | 54.8% | 09/24/2024 |
| 23 | m1CKe | 34 | 21 x 13 | 61.8% | 54.8% | 10/02/2026 |
| 24 | bb3px | 27 | 17 x 10 | 63.0% | 54.5% | 11/20/2026 |
| 25 | kek0vka | 27 | 17 x 10 | 63.0% | 54.5% | 07/04/2024 |
Adjusted pulls the win rate of players with few matches toward the hero's average in these filters (48.8%): 14 wins in 16 does not beat 92 in 134. Position is estimated from gold per minute within the team.