There are 3 major types of damage that heroes can deal in DotA 2:
- Physical damage: Hero right-click attacks deal physical damage, as well as some spells (e.g. Bristleback’s Quill Spray, Tidehunter’s Anchor Smash, Witch Doctor’s Death Wards)
- Magical damage: Most spells in DotA 2 deal magical damage, and so do some items, like the Dagon
- Pure damage: A small subset of spells and passive abilities. Pure damage isn’t reduced by armor or magic resistance.
It’s clear a mixture of damage types is important when playing and drafting in DotA 2. Focus too heavily on physical damage, and your opponents can respond by purchasing items like Solar Crests and Shivas Guard, crippling your one damage source. An over-reliance on magical damage and opponents can counter with a Pipe of Insight, Glimmer Capes, and Black King Bars [BKBs].
While players should always be aware of this, it can still be beneficial to tilt damage towards either physical or magical. This can be in response to early picks from enemies that are weaker against magical damage (e.g. Timbersaw), or weaker against physical damage (e.g. Anti-mage). Pure damage is traditionally used as a counter against Huskar and Anti-Mage, who both have strong magic resistance abilities.
In addition to leveraging a damage type for countering, teams can stack a damage for synergistic reasons. Heroes like Ancient Apparition, who uses his Ice Vortex to reduce enemy’s magical resistance in a small area of effect [AoE], empower heavy magic damage heroes like Zeus or Lina, who can now suddenly deal 30% more damage.
Similarly, the Dazzle, Slardar, Templar Assassin trio is often seen drafted together. All 3 heroes are capable of greatly reducing enemy heroes armor [often well into negative values] allowing Templar Assassin to sometimes even one-shot cores. High physical damage line-ups also benefit from being able to kill Roshan very quickly, allowing them to secure aegis’ more safely and earlier on in a match.
We thought it would be interesting to see what damage types the teams at TI7 were prioritizing in their drafts, as well as what damage compositions led to victories and defeats.
The left-most winning data-point comes from TNC beating Secret in the group stages. TNC had less then 10% of their damage from physical sources! Their radiance Naga Siren and Pugna cores dealt mostly magical damage, while their Timbersaw dealt magical and pure damage.
TNC made a string of great rotations in the early mid-game that set them up with a 5k gold lead at only 15 minutes. If it wasn’t for such a dominating early game, Secret might have been able to stack pipes and BKB’s, and could have resulted in a different outcome.
We have a bit of a chicken-and-the-egg situation on our hands. Our graphs show winning teams more likely to have high physical damage. That’s a fact. However, this does not show whether high physical damage sources lead teams to victory; or winning teams tend to rack up more physical damage than magical damage.
Spell-based heroes often purchase Veil of Discord to boost their magic damage output in fights, yet their other items usually provide maneuverability [Force Staff], control [Eul’s Scepter of Divinity, Rod of Atos], mana regeneration [Bloodstone, Eul’s], or a combination of the above. After Veil, any extra gold they gain from being ahead doesn’t go into items that significantly boost their magical damage.
For physical damage cores we’re much more likely to see stat-boosting items picked up, which both increase their farming efficiency and their physical damage output in fights. Thus leading heavily farmed physical damage cores to greatly exceed their average hero damage when being ahead.
How can we tell if the visible shift from win to loss in Fig. 1 is a product of victory, or a cause?
We attempt to deal with this uncertainty by grouping the TI7 heroes into groups based on their highest damage type output. Heroes who played fewer than 10 games, or averaged under 10,000 hero damage per game, were excluded as being insignificant. We saw that on average a magical damage hero would rack up ~15% more hero damage when they won, as opposed to when they lost. Contrast this with heavy physical damage heroes, who dealt on average 50% more hero damage during their victories.
This indicates the shift in the graph towards physical damage for wins, stems from winning contributing to much higher physical damage outputs. Therefore, picking heavy physical damage doesn’t necessarily increase the likelihood of victory.
DotA 2 games can be won by both heavily magical damage and physical damage team compositions. The important aspect to utilizing damage types to win, is in drafting either a balance of types, exploiting opponent’s potential damage weaknesses, or exploiting your own teams possible synergies.
