The Dota 2 Metagame

Some more item stuff coming later this week, but I want to tidy up some things first.  In the meanwhile, I thought I’d take some time to talk about the concept of a metagame and how it pertains to Dota, specifically how to strategize during the hero selection phase.

A lot of you might be groaning because the very mention of metagames means we’re veering dangerously close to League of Legends territory.  But I think the comparison is an important one to make.  I often hear about LoL players that are too intimidated to try Dota because they simply don’t understand the game.  By explaining the metagame, these players can have an easier time making the transition.  Not that this is a wasted exercise for the Dota crew either.  You can dramatically improve your public matchmaking win rate by just picking intelligently.  Taking a step back to ask why we pick certain heroes and lane them in certain ways also helps move us beyond just copying what works, so we can come up with new groupings that challenge the established conventions.

The most important starting point for understanding the Dota metagame, especially if you’re coming from a LoL background, is that most of the heroes in the game simply do not need items to function.  This is by no means a statement that CS is unimportant.  Instead, what you should take from it is that in LoL you farm items during the laning phase to gain an item advantage which leads to map control.  In Dota, you fight to establish map control during the laning phase in order to safely farm items to gain an item advantage so you can break into their base.  Your lane compositions dictate your plan for establishing that map control.

We can essentially divide Dota’s laning phase into 3 roles: carries, semi-carries, and supports.  Carries scale exponentially with gold and often have a weak or vulnerable laning phase.  Semi-carries scale primarily from leveling up their abilities.  Supports can contribute to fights without either an item or level advantage.  Like most things in Dota these categories are not rigidly defined.  Some heroes are essentially carry/semi-carry hybrids (Lycan, Templar Assassin).  Some semi-carries can function perfectly well as supports and vice versa (Nyx Assassin, Shadow Shaman).  Sometimes you’ll even see characters used as the premier carry for one team and the hardest support for another (Sven).

In Who is the Most Farm Dependent Hero?, I attempted to measure CS scaling.  While there are some issues primarily rising from the small sample size, shifts in how Normal/High/Very High play the game, and emerging trends during 6.74 (Bounty Hunter is likely too high, for example, and Tiny is likely too low) I’m generally pretty happy with how it turned out.


Most of 1-30 are carries.  They need a protected source of farm to become relevant to the game.  Some adjustments I would make

Lycan and Ursa are kinda unique in that they were successfully played as farming junglers, and because so much of their damage came from their abilities instead of their item builds they were practically semi-carries.  This data was from 6.74 and since then they have both been nerfed.

Doom is not a carry, but he’s definitely a semi-carry with a very vulnerable laning phase.

Bounty Hunter gets played as a semi-carry more often than not in high level play, but is usually treated as a carry in lower skill environments.

Recent patches have shifted Spirit Breaker to a more dedicated semi-carry role.

Alchemist is definitely a carry, but was weak in 6.74.

Tiny is now almost exclusively played as a carry.  His Aghanim’s upgrade was still new at this point of the game and has completely redefined how he is played.

Sven is still a weird case, but his post-6.74 buffs have made him much more viable as a carry.

Most of 30-60 are semi-carries.  They don’t necessarily need protected CS, but they do need a reliable source of levels to effectively influence the game.  This one can be tricky because the line between a dedicated semi-carry and a support can get awfully vague.  Also, some of 60-90 are semi-carries whose level of farm does not heavily influence the game like Dark Seer.

One last bit of information for any readers new to Dota.  The Safe lane is the bottom lane for Radiant and top lane for Dire.  It’s called safe because the creeps naturally clash much closer to your tower and is also important because it has the most setups for pulling neutral creeps into the lane so that your minions will tank them.  The Hard lanes are on the sides of the map and tend to leave you exposed and without any pull options.

So with those roles established, let’s see how lane setups evolve.


The most obvious setup from a new players perspective and also the most common setup in lower level Dota.  Ideally you would have 1 carry in the safe lane, one semi-carry in the hard lane, and a semi-carry or a hybrid like Templar Assassin in the center lane getting solo exp.  Let’s look at the dangers of using this setup.

1. Many 5-man compositions will have at least one weak sidelane in the 2-1-2.

2. Your Carry+Support will be facing a 2-man lane, forcing your Support to babysit full-time.  This leaves them unable to use neutral pulls to get extra farm and control the lane equilibrium.

3. You can’t use junglers effectively.

Defensive Tri-lane

This setup exploits the weaknesses of the 2-1-2 and can come in a couple variations.  Variations on this setup become relatively common in higher level matchmaking.

The basic idea is that you commit yourself to having a weak hardlane.  You stick a solo semi-carry or support in the lane to try to get as many levels as possible out of it while playing relatively safely.  If your opponents stick with a 2-1-2 they might have a difficult time forcing a sturdy hardlaner out of exp range, while on the other end of the map their 2 person hardlane is now having to deal with 3 heroes.  If they contest the lane, 3 should beat 2 consistently unless there’s a huge disadvantage in picks or skill.  If they give up the lane your supports now have the ability to pull neutrals with impunity and roam for mid pressure and rune control.

The other big variation of this is to use a carry/support/jungler.  The Jungler can provide tri-lane support when necessary and farm unimpeded when they aren’t needed.  Your support can then neutral pull and this effectively gives all five of your heroes some source of farm and levels.

Aggressive Tri-lane

Primarily this works as a counter to the Defensive Tri-lane, but can also counter 2-1-2 setups.  Your goal is to send a strong 3-man kill lane to your hard lane and make it impossible for their hard carry to farm.  Ideally you also send a good 1v1 carry like Lone Druid, Clinkz, or Weaver to your safe lane where they take advantage of a 1v1 against their hard lane solo, but this can be composition dependent.  Some variations can also include an aggressive jungler as your third.

This setup is relatively uncommon in random matchmaking because it requires the most coordination, but it’s definitely quite viable if executed correctly.

One big advantage to this strategy is that it can let you effectively farm a hard carry in each of your sidelanes.  This gives your team incredible late-game scaling if the opposing team can’t figure out a way to shut down either of your lanes.


Occasionally you will see a dual mid in organized play, but it’s extremely rare in public play.  This typically gets used to create a kill lane situation with an otherwise uncommon mid like Kunkka or Brewmaster.

2 hard lane, 1 safe lane, 1 jungler is a thing, but probably shouldn’t be.  It usually just means your team got greedy during picks, and it’s got huge vulnerabilities against every other setup.

Double jungle occasionally gets used in pubs, but it’s more of a gimmicky fun-time strat than anything else.


So how can we use this information during picks in general matchmaking?

  • For the most part you can assume your team is going to go 2-1-2 unless you see a jungler.  You can try to convince them to try other stuff out, but prepared to be met with confusion or silence.
  • Don’t pick a hard carry unless you’re sure you’ll have a source of farm.  If you think you have a farm source but doubt it’s security at least try to pick a strong laning carry like Viper, Drow, Luna, Naix, Lone Druid, etc.
  • Cultivate a list of heroes that work as flex picks.  Decent in a 2v2 hard lane, capable of handling a hard lane solo, can contribute to a defensive trilane while pulling.  Lots of heroes are viable at all of these (as least in your typical random matchmaking environment), and it’ll help protect you from having games ruined by bad picks.
  • If your team isn’t accomplishing anything in a 2 man hard lane, try to convince them to switch to a defensive tri-lane (or just tell your lane partner to play cautiously while getting safe levels and make the switch yourself).  It’s almost always better to win one side lane than it is to lose/tie both side lanes.
  • If you’re a jungler and you don’t put any pressure on lanes for the first 20 minutes, don’t be surprised if your team collapses.  This is quite frankly inexcusable at every level of play.
  • As a solo hard laner, don’t tunnel vision on getting farm.  Get whatever CS you can without risking death, and focus on getting the levels you need to accomplish things in other parts of the map.  Usually this means getting your ult and immediately moving to another lane to accomplish something with it.
  • Don’t get caught up in protecting towers, particularly as a hard laner.  If they really want your tier 1, there’s very little you can do solo to prevent it and just hanging under it is asking to get dived.  If you can teleport in to punish someone diving on a teammate, by all means, but your outer towers are ultimately expendable so don’t waste resources defending one against impossible odds.

On a related note, No Tidehunter ran a somewhat unconventional 5-carry strategy today that surprised a lot of people but makes perfect sense if you think about it in these turns.  Their lanes were Templar Assassin Mid, Lone Druid solo-Safe, and a Naix/Sven/Gyrocopter aggressive tri-lane.  This was against a Nyx Assasin Mid, Weaver Hard, and Luna/Shadow Demon safe lane with Chen in the jungle.

Their picks sound greedy, but TA vs Nyx is an even matchup, and TA syngerizes quite well with what they were trying to accomplish.  Lone Druid vs Weaver is a favorable matchup for the Druid, especially in the safe lane.  That leaves No Tidehunter’s unconventional aggressive tri-lane vs 3DMAX’s safe lane + jungler.

The first thing to consider is that Sven and Gyrocopter aren’t exactly the hardest of carries.  On my chart they came in at 40 and 41, so they don’t need a ton of farm to be effective.

More importantly, both Sven and Gyro have really useful abilities for roaming out of an aggressive tri-lane.  Sven has a very strong stun, and Gyro’s Rocket Barrage will tear up anyone who is isolated as the Chen quickly found out.  Homing Missile -> Storm Bolt -> Rocket Barrage is basically a guaranteed kill on anyone who gets caught out, and all you need for the combo is a level 2 on Gyro.

Finally, one last thing to consider that isn’t immediately obvious is that both Sven and Gyro have built-in cleaves.  As a support team, this allows them to quickly clear stacked jungle camps in the mid game in order to catch up on levels.  This really wasn’t necessary in today’s game, but it’s a nice option to have in your pocket just in case.

So while it looks like a really unconventional draft, it actually follows the basic principles of an aggressive tri-lane comp.  It’s a pretty good example of using flexible characters like Sven and Gyrocopter to create a lineup that the enemy team is incapable of countering.

7 Responses to The Dota 2 Metagame

  1. Monks says:

    Really interesting post, thanks for that. Do you have a link to the No Tidehunter game you mentioned? I definitely want to see it.

  2. TC says:

    What are your opinions on the DBR that’s coming out?

  3. aphocus says:

    What about the “Sucide-less” strat, seen this a couple of times in pro games, do a tri-line in safe, 1 mid, and 1 either jungling or roaming.

    if I can find a match id I’ll link.

    • phantasmal says:

      Good catch! I had made a note to include that one a while back, but totally forgot about it while writing the article. It’s unlikely to be the kind of strat you could ever convince a pub to run, but it’s definitely a competitive option so it should be mentioned for completeness sake.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: