Flex heroes, or who can play on every role

Carry. Mid. Offlane. Support.

Long ago four roles lived together in harmony. Then, everything changed when the role stealer attacked.

Only the Flex Hero, master of all four roles, could stop them, but when the team needed him most, he vanished…

Dota is a weird game. Everything can work, everything can find its place. And even the most “boring” (for a common man) hero can have not one or even two, but many different playstyles.

If you’ve been following my posts (or seen them before), you may already know that I’m a big fan of obscure playstyles. And, well, it’s not just fun, but is also pretty handy: you may get very deep into playing just one hero, and find an answer for every role!

What playing in party with me feels like

Today I’d like to discuss one rather interesting topic: why heroes can be played on different roles and which of them do it most often?

There won’t be much numbers and stats. Instead, I’ll share with you some of my knowledge I’ve got over the years of playing the game, researching meta trends and working with players, and note the most flexible heroes, based on positions stats for 6k+ MMR games in patch 7.27d.

Russian version

Video version (english subs)

Core patterns

There are a couple of main reasons for a hero to be played on different roles:

a. The hero is too strong (in the current meta) and you don’t want to give it away to the enemy, so you’d get it for yourself instead and shift the hero to any free role.

It doesn’t happen as often in pubs, but is pretty common for competitive scene. The most important here is the reason why a hero can be considered “broken”: it may be a part of a very strong meta combo, or maybe it has an ability that is way too strong.

It may not be tied to the meta directly, too, if opponent loves to play a hero and is pretty successful with it. For example, if enemy loves to play Mars and has a high winrate, why not take it for youself?

It may work a bit differently in pubs, however. When a hero gets picked in the first draft stage too often and gets its role shifted (usually to a support role), then people start doing it in pubs as well.

b. The hero has a lot of playstyles, based on different skills and aspects of the hero

For example, Enchantress as a Core hero is usually built around Impetus, and Support is built around enchanted creeps and healing. Core Dark Willow is based around Shadow Realm damage, attack speed and elusiveness, coupled with the way to farm efficiently. Support Willow is based around utilizing her AoE control capabilities.

Arc Warden is a rather interesting example. This hero can be played as a classic hybrid core (Hand of Midas, Mjolnir and focused around Spark Wraith), as well as an aggressive physical damage core (based around Magnetic Field and Flux, built around Power Treads, Mask of Madness, Drum of Endurance and Desolator/Diffusal), an aggressive magic damage ganker (based around Flux range, Blink Dagger and Dagon), an utility Arc (focused on Magnetic Field and Meteor Hammer, usually played as an offlane core) or even a support (based around Magnetic Field, Drum of Endurance and Solar Crest).

tfw you are your own support

Obviously, only one of these playstyles is the most accepted, but if you’ll look close enough and research matches of different Arc spammers at high MMR (and pros), you may notice, that all of these playstyles have a right to live — under specific conditions or in the right hands. Aside from that, some of these playstyles may be different to a point when they can be considered a different hero.

The most interesting and probably most common reason to switch, however, is adjacent roles.

Adjacent roles

Roles in Dota are not fixed. In fact, borders between them have been blurred to a point when roles can be considered a spectrum, rather than a set of five distinct positions.

Here’s when you can get to the first conclusion: heroes are often switching between two closest roles.

It’s most noticable with 1st and 2nd position, 3rd and 4th and between support roles.

The best examples of it may be Miracle- and MATUMBAMAN, who are famous for switching lanes and even heroes, or RAMZES666 and No[o]ne-. Another good example is Zai, who’s switching his roles pretty often: at different times and on different teams he could be an offlaner, while usually being a position 4, or a position 4, while usually playing as an offlaner.

Usually this kind of switch happens based on specific in-game situations and drafts, but more often the reason is a search for a better early game — in both competitive and pub games. So it’s not unusual to see “midlane” hero as a “carry” (and vice versa) if the hero’s early game may decide the direction for the whole match.

Supports are another story. If a couple of years ago there used to be a strict line between the 4th and 5th positions, it’s not so strict now, with free wards and cheaper support items. Heroes, that usually are played as position 4, mostly are potential midlaners (or cores in general), while position 5 heroes are usually played exclusively as supports.

And it’s still kind of hard to understand who is position 4 and who is 5 in a particular game, since the division is not as obvious as “offlane support is position 4” or “more wards means position 5”.

Another example of lane and role swaps is more likely to happen in pubs — lane swap between side lanes. It usually happens first when looking for a better laning matchup, but can become more common later. For example, it’s how Wraith King, Ursa and Chaos Knight started to be played as offlaners, and Bounty Hunter or Bristleback are pretty popular position one heroes at higher ranks and in competitive games. It’s most common for heroes who are strong during laning stage and in midgame, and who can pressure enemy safelane core.

This kind of metamorphosis can result in some other funny situations, when heroes (usually the ones who are strong in the current meta) switch through different roles to position 4 or even 5. At different times and in different patches this kind of thing happened to Wraith King (1 → 3 → 4), Sven (1 → 3 → 4), Chaos Knight (1 → 3 → 4), Monkey King (1/2 → 3 → 4), Invoker (2 → 3 → 4), Nyx Assassin (2 → 3 → 4/5). Sometimes it may to in the opposite way: it happened to Io (4 → 3 → 2/1) and Snapfire (4 → 5 → 4 → 3 → 2), but it’s much more rare and is pretty sensitive to any meta shifts.

Unique role in this regard is the middle lane core. Heroes here can be drastically different and can switch from any position. The main things here are the ability to win your lane (so you need to choose your hero for a specific matchup) and a hero’s ability to use the earned advantage (it can be a strong lategame potential, flash farming potential or gank potential).

And if the situation with safelane cores is already quite obvious and it’s mostly about matchup for them, it’s a bit more than that for others. Classic offlaners are mostly able to not lose their lane, and are also quite efficient at jungle farming. They can also make life much harder for their opponents, while feeling great throughout the game themselves. Good examples of this are Kunkka, Magnus, Axe (who is particularly interesting because he is able to skip waves between towers easily, which makes the lane much harder for enemy).

And since most of the support heroes are pretty strong laners by themselves, for them the main reason to switch their role is some kind of change that allowes them to scale better into the lategame.

Good examples of this are Dark Willow (who was pretty popular mid hero after being released, and became popular again after she got an Aghanim’s upgrade), Earth Spirit (who was pretty popular choice for midlane ever since his release because of his gank potential), Snapfire (after the hero’s release key factors were great base damage, efficient flash farming and lategame potential with the Lil Shredder level 20 talent, and came an Aghanim’s scepter), Dazzle (after his Poison Touch rework, addition of Bad Juju and Guardian Greaves change), Winter Wyvern (at first — because of her aghs, then because of added possibility to attack heroes under the Winter’s Curse effect).

In general, “carry potential” is not as important for a mid hero: most of the relevant mid heroes are usually played as something between core and support.

Many of transitions like this are not as typical for pubs. People here are more likely to stick to competitive meta and pick “for sure”, so their choices are more conservative. Often this kind of trends — be it playing on a new position or a new build for Anti-Mage — are easy to notice when watching high rank matches (7–8k+ MMR), but not as obvious when you’re looking at the “bigger picture” or even 6k+ MMR trends.

My most favorite example of this is 2B and Support Invoker. 2B is a 7k+ MMR player and streamer, who’s famous for constantly playing traditionally core heroes as supports and is quite successful at it. Often he’s being unfairly criticized for it by his team, and when he started playin Support Invoker, their reaction wasn’t any different from it when he played support Troll or Anti-Mage. And yet other players started noticing this kind of playstyle.

mah boy 2B — https://www.twitch.tv/2b_doto

Support Invoker’s popularity wasn’t higher than 10% compared to the hero’s other roles at 6k+ MMR, but everything changed when YapzOr played it in an official match. After that part of Invoker’s games as a support skyrocketed to 76% and even now it still remains stable at ~12%. And it was pretty similar story for Ember Spirit and Sniper as supports.

Flex heroes

But what heroes are the most commonly accepted “flex picks” at the moment? Who can be picked for different roles and switch between them?

To find out I used positions overview for patch 7.27d from my ranked reports based on matches of 6k+ MMR players.

My main criteria for selecting heroes was having at least 10% of games on a non-primary role. It doesn’t mean that hero-role pairs with less than 10% are not significant — even 2–3% of games are already significant enough to pay close attention to a hero, but we are looking for the most common ones. In some specific cases I allowed myself to include heroes who are below this line, based on my experience from previous patches or observations from competitive matches/scrims.

Heroes are divided in 5 tiers:

  • Tier 5: two roles, one playstyle
  • Tier 4: two roles and different playstyles
  • Tier 3: two roles with closer distribution of matches or three roles (that may not be represented in stats as much)
  • Tier 2: more noticable distribution between three roles or four roles
  • Tier 1: five roles

It’s also worth to point out: I didn’t pay attention to winrates as much. Aside from a few exceptions, win rates of different roles for one hero are pretty much the same, and win rates in general in this case is more about understanding the hero and its playstyle on a role better. Besides, if our goal is to find the most accepted flex heroes, win rates have the least importance.

The last thing I’d like to point out: support role stats is not separated because of technical reasons. So it may be both position 4 or position 5. I marked support positions mostly based on my observations, but support roles are so close to each other, that you can consider them one role.

Tier 5

Heroes, who stand out the most in terms of two position divisions, without changes in playstyles, are:

Here we can already notice the same transitions I talked about before. The most common pairs of roles are safelane and midlane cores, as well as offlane and support. The only two interesting cases are Lone Druid and Dragon Knight.

Lone Druid is traditionally coming to offlane, and yet his playstyle on all three core roles is pretty much the same, so you can say that LD is surprisingly stable.

DK is, however, the closest to cross the line of two different playstyles.

Tier 4

And here, where playstyles for two different roles diverge enough, you can see a pattern: the division often happens between a core and a support roles (the only exception is Death Prophet, but historically she became a common offlane core through being played as a position 4 first). The difference can be easily explained: usually these heroes are required to do different things on these positions, and are more likely to change both their item build and playstyle based on a different amount of resources they have.

The best examples of it are Lina, Mirana and Abaddon. In case of the first two, one build is based around cheap and efficient items (Tranquil Boots, Mekansm, Eul’s Scepter), spell damage (aghs, lens), auras and surviveability, while the other is based around attack damage (Monkey King Bar, Daedalus, Manta Style).

Abaddon’s builds are pretty similar, but different enough: support build is cheaper in general, and the game is mostly built around surviving, healing allies and buffing the team; offlane build has a bit more focus on attack damage, team auras and is based around the idea of being a tanky frontliner.

This tier is also the place where you start noticing significantly more different roles in a pair: midlane core and supports. It’s hardly a coincidence: many common mid picks have been supports back in the day (or became supports today), and usually we are talking specifically about position 4. The reason for it may be better lategame potential and greediness on one hand, and strong spells, coupled with meta, on the other. It also makes it clear: most midlane cores can become supports (and vice versa) if you’re creative enough.

Tier 3

In this tier we start to see many heroes, who are rarely played on some positions in pubs, but can be noticed on the highest level of ranked matches, in competitive games and in teams’ training matches.

One particularly interesting example of it is Mars. This hero is not as common for midlane anymore, and the support Mars trend didn’t get enough momentum yet to become noticable in stats. However, in the last 6 month more and more players started to pay attention to this hero as a support, and now at least four different teams are training strategies that have Mars as a position 5 support. Maybe we are going to see it in a competitive game soon, who knows?

Arc Warden is in a similar situation. Self has a very clear distinction between safelane and midlane, and it’s also one of the heroes who has different winrate for different roles (48% for sidelanes and 54% for midlane), but some teams and pub players are seeing it as an offlane or even support hero. There are a couple of reasons for it: first, he has a strong gank potential, second, he has a strong combo with Drum of Endurance and Solar Crest, and third, the Magnetic Field. But even though players were looking at the hero on these roles for a while now (especially in SEA and China), it still remains a niche pick that is hard to get going.

Enchantress was considered equally good as an offlane core and as a support for a long time, but not since the hero got changed last time. Today she’s almost exclusively played as a support — even though position 4 and position 5 playstyles may be drastically different for her — and popularity of other roles decreased significantly.

For other heroes roles division isn’t that big of a secret, but with higher skill level (or in team matches) it’s more common to see these heroes on more “strange” roles. Usually there are specific reasons behind it.

For example, for Dazzle it’s his strength on a one versus one lanes, as well as flash farming (given a good start) and efficiency of team items in his hands. In case of midlane Luna people usually consider magic damage build (aka “Orbital Cannon Luna”). Bounty Hunter and Timbersaw are pretty common niche carries in general, but they require understanding the game on a decent level. Doom was always popular as an answer to enemy Chen picks.

Honorable Mentions

And before we go further, I’d like to note some heroes, who were played on different roles before, but these playstyles are either forgotten, or lost their popularity.

  • Sven Support was popular for a while after a change to his third ability (when it was physical damage shield), but after it got changed back, it was forgotten.
  • Morphling Support (aka “Stun Gun Morphling”) is probably as old as Dota itself. I can’t recall it being very popular at any time, but it was always working and people still tend to remember it. And, well, it’s not surprising: who wouldn’t love to throw some 4 seconds stuns to enemies?
  • Slardar Safelane/Midlane is not really forgotten, and it is picked occasionally at higher ranks, but not in general. The main playstyle is pretty much the same as for offlane in this case.
  • Lifestealer Offlane/Support was popular for about a week, when Infest started to give additional HP to allies, and aghs was changed to upgrade Rage.
  • Axe Support is played in about 5% of Axe’s games and was popping in high MMR game for a long time now. Who knows, maybe we’ll see it in a competitive game one day?
  • Skywrath Midlane used to be popular, but now he’s picked only in 5% of Sky’s games, when stars align.
  • Ancient Apparition Midlane is in the same situation as Skywrath. AA was a popular choice for midlane before, but now it’s considered to be more of an exotic choice. He’s still played occasionally, and AA’s new aghs allow him to scale better into the lategame, but usually it requires way too specific conditions.
  • Io Safelane/Midlane Core is not as wild anymore, and Ana, who made it popular, is nowhere to be found. But this playstyle is still used by 8k+ MMR boosters.
  • Shadow Demon Offlane was never really popular, but was leaking to competitive matches, usually played by iceiceice and 9pasha.
  • Naga Offlane/Support is a weird one. Naga always goes through some sort of a cycle: in one patch she’s usually played as a core, in the other — mostly as a support. Lately utility Naga is not as good, but it’s unique enough and deserves a mention.
  • Dark Willow is almost never played on any role, but support. But it wasn’t always like that: she used to be a Midlane and Offlane hero as well.

I can list even more exotic stuff: support Troll, mid Underlord, mid Earthshaker, support Clinkz, even support Brood. But I guess if we’re going to remember all the exotic stuff, this list will become infinite.

Tier 2

This category is the first one to include my favorite type of heroes: “Panic picks”. Gone to kitchen, came back and there’s a game going on and you need to pick something quickly? You’re playing in party and still haven’t decided who’s playing what role? You can always pick one of these heroes! They have a lot of different playstyles, can be played on various roles, and you can even get more out of them if you know what you’re doing.

The most questionable one is probably Venomancer. Even though he has a distinct division between offlane core and support, it’s a clear division between pubs and pro games — what’s up with the other core roles? The answer is simple: almost every single patch, many teams are actively seeking for a way to add this hero into their drafts, switching him from role to role, training Veno strats in their scrims. The reason for this is him being a disgusting hero to play against. The hero itself may not be as strong, but playing against him is very uncomfortable, and the vision and sieging that Veno provides alongside this mental warfare is always valued in a draft.

The similar reason is behind Nature’s Prophet and Puck. Their playstyles are almost identical on every role, and they are frequently picked, just like veno, because they are uncomfortable to play against. These two are very efficient at split push, elusive and have high base damage values, which can make the laning stage unbearable.

You can also note Viper, who’s on the list for a similar reason. He’s often played as a safelane core as well (but it’s usually more about a lane swap, rather than about role switch). Silencer, like Viper, is able to become a strong additional core for his team, if played as a support. The only detail to note for core Silencer is priority of offlane over safelane: it’s most likely related to lane swaps and early aggression against enemy safelaner.

Snapfire stands out a bit. She is one of the heroes who has different win rates on different roles: 48% as a support and 41–42% on every other role. It’s hard to say what is the exact reason behind it. I think it has more to do with people’s understanding of the hero on these roles: Abed and Topson started the core Snapfire trend and were pretty successful with it in pubs. However, it could’ve been purely because of surprise factor, and once people got used to it, win rates fell off.

The most interesting heroes on the list are probably Chaos Knight and Lycan. Both of these are pretty strong laners, and both of them are rarely seen as supports in ranked pubs. It’s not rare, however, for higher level of play: more and more people are starting to consider these heroes as position 3–4 ones. The reasons for it are their aghanim upgrades, overall utility and ability to either sustain under pressure early on or play aggressively. And while CK as a support popped up years and patches ago, Lycan is quite new to this thing: only after he got the previous aghs upgrade he started to gain popularity as an utility core.

Tier 1

And who are the most flexible heroes? Who can be played on every role?

It’s not hard to notice, that most of them (like some of tier 2 heroes) are quite popular among boosters.

For a common man all these heroes are mainly supports or midlaners. But there’s more to it.

The thing they all have in common is their ability to do pretty much everything. Every each of these heroes has some kind of utility (a reliable stun, an aura, a vision and scout spell, or even everything combined), push potential (and you don’t even need to attack by yourself), surviveability (be it good stats, a spell for tankiness, an escape spell or even movement speed).

It clearly shows, that carry doesn’t need to be “classic”: Lechrac, Pugna or Visage are often played as safelane cores (if they are meta) in competitive games, and do their job — which is surviving and destroying enemy buildings — quite well.

(These aren’t the only things a carry should do in a game: you can be a carry in a lot of different ways; it is, however, the most important thing for these specific heroes to do)

The only hero that stands out from this list is Vengeful Spirit, a hero whose popularity across roles isn’t reflected in these statistics; However, the legends of Carry Venge are as old as the hero itself, and there are few people that still hold doubts about her viability as a midlane or safelane pick at this point.

It’s a bit more interesting with offlane: this playstyle gained popularity after the controllable illusion on death was implemented. After that, Venge, who already had pretty useful spells and great stats, suddenly turned into a strong tank-initiator, who can build all sorts of auras and die in teamfights without fear, while still being useful for the team with all the auras and natural tankiness of the illusion.

Closing words

Of course, flexibility of heroes is a relative thing, and sometimes you may notice completely different playstyles through item builds for the same role (we’ll talk about it later, I promise).

But please, don’t think that just because a hero was mentioned it can’t be played on a different role, or that any deviation from the “standard” is going to lose you the game — you can see in this piece how successful innovation still is in this game.

As I already mentioned, roles are a spectrum, and Dota is a weird game, where everything can work.

And that’s about it! While I’m working on new content and stats hub updates, you may want to subscribe to me on Twitter, YouTube or join my Discord (or even subscribe at VK and Telegram if you’re Russian).

Writing code and stories. https://spectral.gg/