What Is Role Queue MMR? Core vs Support Split Explained

What Is Role Queue in Dota 2?

Role Queue — officially called Ranked Roles — is Dota 2’s ranked matchmaking mode where every player declares their intended position before the game starts. You select one or more roles from positions 1 through 5, and the system builds a lobby where all five positions are covered by players who chose that role for that game.

This stands in contrast to Classic Ranked, where players enter a lobby and fight over roles during the draft or early game. Role Queue costs tokens to access for core positions (more on this below), while Classic Ranked has no token requirement but also has no role guarantees.

The five positions break down like this:

Role Queue was introduced to solve the persistent problem of players claiming support in queue and then stealing farm once the game started. By requiring a declared role and tracking performance within that declaration, Valve created a foundation that eventually led to the core/support MMR split.

The Core / Support MMR Split Explained

When you play Ranked Roles, your MMR does not simply go up or down as one number. The system maintains two separate rating pools:

• Core MMR — tracks your performance when you play positions 1, 2, or 3
• Support MMR — tracks your performance when you play positions 4 or 5

Both numbers are visible on your player profile. The system uses your Core MMR to match you in lobbies when you queue core, and your Support MMR to match you in lobbies when you queue support. They move independently based on results in their respective role category.

This means a player can be Archon in Core MMR and Crusader in Support MMR simultaneously, or the reverse. These are not hypothetical edge cases — the gap between a player’s two pools is often substantial because most players heavily favor one role category over the other.

Core MMR (Positions 1, 2, 3)

Core MMR covers the three farm-priority roles. Carries, mid heroes, and offlaners all contribute to this single pool. Winning as pos 1 and winning as pos 3 both move the same Core MMR number, so if you play across all three core positions, your Core MMR reflects your average performance across all of them.

Core MMR is typically higher for players who have invested more time climbing as a farming role. It is also the MMR that most players talk about when they say “my MMR” because carry and mid roles historically attract more ranked game volume in lower and mid brackets.

If you use a Dota 2 boosting service to raise your rank, the service will typically specify whether they are climbing your Core MMR, your Support MMR, or both, since they require different hero pools and game-sense from the booster.

Support MMR (Positions 4, 5)

Support MMR covers positions 4 and 5. Both soft support and hard support contribute to the same pool. A player who mains pos 5 Witch Doctor and occasionally plays pos 4 Tusk will see both results reflected in one Support MMR figure.

Support MMR tends to be lower than Core MMR for most players for a straightforward reason: supports have lower individual impact on game outcomes in many brackets, and the pool of dedicated support players is smaller, meaning calibration starts from a less-developed baseline for players who later switch to support play.

However, support-maining is a legitimate and increasingly viable strategy for climbing. The MMR boosting guide on this site goes deep on which bracket benefits most from a support-focused climb.

How Your MMR Changes Per Game

Each Ranked Roles game ends with an MMR adjustment to the pool that matches your selected position for that game. The adjustment is typically in the range of 20 to 30 MMR per game, though the exact amount varies based on your uncertainty rating (how many games the system has calibrated you on in that role) and the relative MMR of players in the lobby.

The table below summarizes which pool changes depending on the role you declare:

Classic Ranked still exists as an option and does affect MMR, but Classic Ranked does not use the role-specific split in the same way. Players who queue Classic Ranked have historically seen adjustments to a combined or hybrid MMR figure. If you want to move a specific pool cleanly, Ranked Roles is the correct queue.

Why Valve Split MMR by Role Category

The core/support split did not arrive arbitrarily. It was a direct response to a persistent abuse pattern that emerged when Role Queue was first introduced: role farming.

The Role Farming Problem

Before the Core/Support split, Role Queue used a token system but a single MMR pool. High-MMR players discovered that they could queue as support, play the game as a carry or mid hero, and because their raw individual skill was so far above the lobby average, they could win games easily regardless of the declared role. Each win raised their single MMR number.

The result was lobbies where someone declared pos 5 but built a Battlefury and took every last hit in the safe lane. Their teammates — who had selected carry expecting farm priority — were denied resources and forced into a fundamentally broken game state. The declared role meant nothing because the MMR adjustment was the same either way.

This was called “role farming” and it was widespread enough to degrade the matchmaking experience for large portions of the player base, particularly in high-bracket lobbies where skilled players had the most to gain by gaming the system.

How the Split Solves It

By tying MMR gain to a specific role pool rather than a shared number, the incentive to abuse roles collapses. If a 7,000 Core MMR player queues as pos 5, they win games against a 7,000 Support MMR lobby — but their Support MMR is probably 5,000. They gain Support MMR from wins, which is fine, but their Core MMR does not budge. They cannot use support games to pad the number that reflects their carry performance.

More importantly, because they are matched against other ~5,000 Support MMR players when they queue support, they are no longer placed in lobbies far below their actual skill in that pool. They play against peers in the support context, which produces fair games rather than stomp scenarios.

The split also created a more honest profile: a player’s Core MMR and Support MMR independently reflect how good they actually are at those role categories, rather than one number that mixes different skill expressions into a blurry composite.

The Role Queue Token System

Access to Ranked Roles for core positions is gated by role queue tokens. Here is how the token economy works at a high level:

• Each Ranked Roles game as a core position (1, 2, or 3) consumes one token
• Each Ranked Roles game as a support position (4 or 5) earns one token
• Classic Ranked games have no token cost or reward
• Tokens are capped at a maximum amount to prevent unlimited stockpiling

This token mechanic has a simple purpose: it creates a structural incentive for players to volunteer for support roles. If you want to play carry, you have to have tokens. The reliable source of tokens is playing support. Players who only want to queue core must either play some support games first or exhaust their token reserves over time.

The system is not intended to force anyone to play a role they dislike permanently. A player who has a large token reserve can queue core for extended sessions without touching support. But the design means that the support queue will always have players who are motivated to be there, reducing the frequency of “reluctant supports” who play the role resentfully and perform poorly as a result.

Core MMR vs Support MMR: Which Number Matters More for Your Goals?

The answer depends entirely on what you are trying to achieve.

If You Are Trying to Display the Highest Possible Rank

The medal and rank displayed on your profile is derived from the higher of your two role MMR pools (in most current implementations — confirm in-client as Valve may adjust this). This means if your Core MMR is 5,500 and your Support MMR is 4,200, your displayed badge reflects the 5,500 figure.

Players trying to display a specific rank as efficiently as possible should focus on whichever pool is closer to the next medal threshold. If your Core MMR is 100 MMR below Divine and your Support MMR is 600 below Divine, grinding core is a far faster path to the Divine badge.

If You Are Trying to Climb and Actually Improve

The more meaningful MMR is the one for the role you intend to play long-term. A player with a boosted Core MMR who primarily plays support will struggle every time they queue core because they are matched against players whose game sense genuinely reflects 5,500 core skill. The mismatch creates frustrating lobbies in both directions.

If you are wondering whether it is safe to boost your Dota 2 MMR and are deciding which pool to target, most experienced players recommend targeting the role you actually enjoy and plan to continue playing. Displaying a high Core MMR that does not match your real core ability tends to reverse quickly when you try to maintain it yourself.

If You Want Both Pools High

Some players want competitive numbers in both role categories. This is entirely achievable but requires time investment in both queues. The most efficient approach for players who enjoy multiple roles is to identify which pool is lagging and dedicate a sprint of games to that pool — say, 30 consecutive support games — before returning to the preferred role. The gap narrows faster than you might expect once you are playing against correctly rated opponents in that pool.

How the Core / Support Split Affects Matchmaking Quality

The split’s most direct impact on your game-to-game experience is lobby composition quality. When every player is matched on the MMR pool that corresponds to their declared role, you should theoretically always be in a lobby of peers in that role context.

A 4,500 Core MMR player queuing mid will be in a lobby where the opposing mid also has roughly 4,500 Core MMR. The supports in that lobby have roughly 4,500 Support MMR. The system is running five parallel matchmaking pools simultaneously and finding an intersection that satisfies all five role-specific constraints.

This is more complex than single-pool matchmaking and the reason ranked roles lobbies can sometimes take longer to assemble than Classic Ranked, particularly at high brackets where the eligible player pool is smaller.

The Calibration Phase

When you first start playing Ranked Roles after not having established role-specific MMR — whether you are a new player or are returning after playing only Classic Ranked — the system will have a higher uncertainty about your actual skill level in each role. During this calibration window, MMR swings per game can be larger than the usual 20-30 range because the system is rapidly trying to place you in the correct bracket.

Seasonal recalibrations (when Valve resets ranks at the start of a new ranked season) apply a similar expanded uncertainty window. Core MMR and Support MMR recalibrate on the same seasonal schedule, but they recalibrate independently. A strong recalibration run as carry followed by a poor one as support will end with different adjustments to each pool.

Party Queue Interactions

When you queue as a party in Ranked Roles, the system averages the role-specific MMR values across party members for matchmaking purposes. If you are a 6,000 Core MMR player queuing with a 4,000 Core MMR friend and you both select mid, the effective Core MMR for lobby assembly is roughly the midpoint. This is a significant consideration for players who want to maintain a clean rating in their primary pool. See the solo vs party MMR gap breakdown for a detailed explanation of how the math works.

Strategy: Should You Specialize or Play Both Role Categories?

This is one of the most frequently debated questions in the climbing community. There is no single correct answer, but there are clear patterns based on what successful players do.

The Case for Role Specialization

The strongest argument for picking one role category and grinding it exclusively is consistency. Dota 2 at its core rewards players who deeply understand a narrow set of game situations. A player who plays only pos 2 mid develops pattern recognition for mid matchups, tempo timings, and rotation windows that a player who splits time across all five positions will always be behind on.

Role specialization also produces faster MMR gains in the target pool. Every game refines the same set of decisions, hero interactions, and muscle memory. Improvement compounds when you are not constantly context-switching between farm-priority carry play and ward-placement support play.

Successful boosters at TeamSmurf — players who reliably climb accounts to very high MMR — almost universally specialize on two or three heroes within one role category. The mechanics work across all brackets: focus compounds.

The Case for Role Flexibility

Flexibility has legitimate value at specific points in your ranked journey. Early in a ranked season, when the token reserve is low and the token economy is resetting, having competent support play available means you are never locked out of the ranked queue. Flexibility also helps when you need to fill a role in a party game, or when a specific role category is facing long queue times in your region.

At lower brackets, there is an additional argument: the game sense required to play support well — reading team fight position, understanding the draft, tracking the minimap — transfers directly to core play and often accelerates core MMR gains when the player eventually switches back. Learning support at Guardian or Crusader level before committing to a core main is not a wasted investment.

A Practical Split for Most Players

The most common approach among players who ask us this question is an 80/20 split: roughly 80% of ranked games in your primary role category, 20% in the secondary. This builds enough tokens to sustain the primary queue, keeps the secondary pool from decalibrating too severely, and does not fragment your hero pool to the point where consistency suffers.

If you are evaluating whether a role-specific boost makes sense for your situation, the MMR boosting guide covers how to target a specific role pool efficiently and what results to expect by bracket.

Common Misconceptions About Role Queue MMR

Misconception 1: “Classic Ranked is safer for MMR”

Some players avoid Ranked Roles to dodge the token cost and assume Classic Ranked is a better path for their rating. In practice, Classic Ranked lobbies often have more role chaos, which introduces variance that is independent of your actual skill. Role Queue lobbies — while not perfect — produce more predictable game states where individual skill has a cleaner path to expression.

Misconception 2: “Your overall MMR is the average of Core and Support”

The displayed medal and rating is not a simple 50/50 average of your two pools. Valve has historically used the higher of the two pools for profile display purposes, though the exact formula has evolved. Check your in-client profile for current display logic, as this can change between patches.

Misconception 3: “Playing support in ranked doesn’t contribute to your real MMR”

Support MMR is real MMR. It affects which lobbies you enter when you queue support, it is displayed on your profile, and it contributes to your medal display. The idea that support games “don’t count” is a legacy misconception from before the split existed and no longer reflects how the system works.

Misconception 4: “You can’t boost a specific role pool”

Role-specific boosting is not only possible but common. Services like TeamSmurf can target your Core MMR, your Support MMR, or both depending on what you need. The booster queues the declared role throughout the order, so gains land in the correct pool. If you are considering this, read about boosting safety in 2026 to understand the precautions involved.

Misconception 5: “A high Core MMR means you understand support at the same level”

Core and support require different game knowledge. High Core MMR players who occasionally queue support often find themselves matched against players with comparable support skill who know ward spots, creep pulling timings, and rotations far better. The pools are genuinely independent skill sets, and the split MMR system reflects this accurately.

What Happens to Role Queue MMR at Season Reset?

At the start of each new ranked season, Valve applies a recalibration to both your Core MMR and Support MMR. The recalibration does not wipe your rating entirely — instead, it introduces a softcoded uncertainty boost that makes your rating more volatile for the first several games of the new season, allowing the system to quickly update your position based on current performance rather than results from months ago.

Core MMR and Support MMR recalibrate on the same schedule but move independently. If you crush your first 10 calibration games as carry but play poorly in your first 5 as support, the outcome will be a higher Core MMR and a lower Support MMR than you ended the previous season with. There is no mechanism that ties the two pools together during recalibration.

Practical preparation for recalibration: play the role you want to end higher in for your first games of the new season, when uncertainty is high and wins have extra weight. For players who want to enter a new season at the highest possible bracket, this is the biggest high-leverage window in the entire ranked calendar. The hidden MMR / Glicko system explanation goes into the technical details of how uncertainty affects per-game gains.

Role Queue MMR and the Bigger Picture

Understanding the core/support split matters beyond just knowing what number moves when. It changes how you think about your ranked identity, your token economy, and what “climbing” means in practical terms.

A player who is stuck at Archon in Core MMR but has Legend-level Support MMR is not a Legend player who got unlucky with carry teams. They have a genuine skill gap between the two role categories that will close only through deliberate practice in core games. The split forces this honesty.

Conversely, a player whose Support MMR is far below their Core MMR is not “bad at support.” They have simply not played enough support to calibrate that pool. A sprint of 20 to 30 support games will usually bring that number up quickly — not because the system is generous, but because the initial calibration of any underdeveloped pool moves fast to find the right level.

The role queue system in 2026 is more sophisticated than any version that came before it. The core/support split, the token economy, and the separate calibration cycles combine into a matchmaking architecture that — when used correctly — puts you in lobbies that test and reward your actual skill level rather than a historical average that no longer reflects your game.

Use it deliberately. Know which pool you are targeting on any given ranked session. Track both numbers, not just the one on your badge. And if you want to accelerate either pool with professional Dota 2 boosting, make sure the service you choose understands the role-specific nature of the request — because a boost that ignores your declared role lands in the wrong pool.

Frequently Asked Questions