Fact-checked by Grok 2 weeks ago

Skill-based matchmaking

Skill-based matchmaking (SBMM) is a computational system employed in online multiplayer video games to pair players against opponents of comparable skill levels, thereby promoting balanced, competitive, and enjoyable gameplay experiences. This approach relies on algorithms that estimate player proficiency through performance metrics such as win rates, kills, and scores, adjusting matches to minimize skill disparities across teams or individuals. The foundations of SBMM trace back to early systems like the rating method, originally developed for chess in the , which was later adapted for to handle team-based and multi-player scenarios. advanced this with TrueSkill in 2006, a Bayesian skill rating model implemented in Xbox Live titles such as Halo 3 and Gears of War, enabling rapid skill estimation even for new players by accounting for uncertainties and team dynamics. An improved version, TrueSkill 2, introduced in 2018, incorporates additional factors like player experience, squad composition, and quit behavior to enhance prediction accuracy to 68% for match outcomes, compared to 52% in the original. In practice, SBMM operates by calculating a player's rating post-match based on expected versus actual performance, using metrics like kills per minute (KPM) or score per minute (SPM), and then queuing similar-rated players together while balancing for and wait times. For instance, in , is derived from relative performance across the player population, with updates weighted by outcome predictability to reward adaptability and resilience. Advanced implementations, such as those in Ghost Recon Online, extend beyond raw to include behavioral data and preferences for holistic match quality. The primary benefits of SBMM include reduced match blowouts, higher player retention, and decreased quit rates, with data from early implementations showing 80-90% of players achieving better end-of-match placements and extended session times when skill is prioritized. In simulations using beta data, demonstrated faster convergence to stable skill distributions than after just 50 games, fostering competitive incentives and overall engagement. However, tighter skill matching can intensify competition for high-skill players, potentially affecting their enjoyment, though it disproportionately aids novices by preventing discouraging mismatches. By the mid-2020s, SBMM faced substantial criticism in titles like , prompting developers to introduce options for less skill-focused as of 2025.

Overview

Definition

Skill-based matchmaking (SBMM) is a system employed in multiplayer to pair with opponents or teammates of comparable estimated levels, thereby fostering balanced and competitive matches. This approach relies on algorithmic evaluation of player performance to minimize disparities in match outcomes, promoting fairness by reducing the likelihood of one-sided encounters. Unlike random , which assigns without regard to , SBMM prioritizes equity in competition over expediency or chance. The core components of SBMM include input data derived from player statistics, such as win-loss records, kill-to-death ratios, and overall performance scores accumulated across sessions. These metrics feed into a estimation , typically involving systems that quantify a player's proficiency, though the precise methods vary by and are often proprietary. The output consists of assembled lobbies or teams where participants' estimated skills are closely aligned, aiming for an approximate 50% for each side to enhance engagement. SBMM is commonly implemented in genres requiring precise coordination and competition, such as first-person shooters (e.g., ), where balanced groupings help maintain challenge without overwhelming novices or underutilizing experts. Skill rating systems, such as those akin to , provide the foundational estimates but are referenced here only as enablers of the matching process.

Objectives

Skill-based matchmaking (SBMM) aims to promote fair competition by pairing players with opponents of similar levels, thereby creating balanced matches that minimize one-sided outcomes and enhance overall equity. A core objective is to reduce player frustration from mismatched encounters, such as frequent losses against much stronger opponents, which can otherwise lead to disengagement. By fostering environments where players can compete meaningfully, SBMM encourages improvement through challenging yet achievable interactions, allowing individuals to refine abilities without overwhelming discouragement. Ultimately, these goals contribute to maintaining player retention, as balanced experiences keep a broader audience engaged over time. Key principles guiding SBMM include targeting approximate 50% win rates for to ensure consistent challenge without predictable dominance, achieved through algorithms that balance team compositions based on estimates. This involves adaptive difficulty adjustments, where constraints may loosen if wait times extend, prioritizing over perfect to avoid excessive queue delays. Such principles support equitable participation across spectra, with 80-90% of reporting improved placements and reduced quit rates when is factored in. On a broader scale, SBMM enhances for beginners by shielding them from elite players, while simultaneously providing experts with sufficiently competitive lobbies to sustain interest, thus broadening the game's appeal. This sustained engagement indirectly bolsters strategies, as retained players are more likely to invest in in-game purchases and long-term play. A notable tension in SBMM design lies between upholding competitive integrity—through strict skill alignment—and preserving casual fun, where overly rigid matching can alienate relaxed seeking low-pressure sessions. Objectives often vary by mode; for instance, ranked modes emphasize precise balancing to support progression and tournaments, whereas unranked modes adopt looser criteria to prioritize quick entry and social enjoyment.

Historical Development

Origins

The origins of skill-based matchmaking can be traced to the chess world, where the challenge of creating balanced tournament pairings necessitated a reliable method for assessing player strengths. In 1960, , a Hungarian-American physicist, chess player, and professor, developed a specifically designed to improve upon the existing Harkness method used by the (USCF). This innovation provided a numerical value representing each player's relative skill as a probabilistic estimate derived from historical game outcomes, enabling organizers to pair competitors with similar ratings for more equitable matches. The system's foundation in statistical modeling allowed it to predict expected results between players, adjusting ratings post-game to reflect actual performance and maintain accuracy over time. The quickly proved its value and was formally adopted by the USCF in , marking the first widespread implementation for organization. By 1970, the International Chess Federation () had endorsed it as the global standard, applying it to international competitions and player classifications. personally oversaw rating calculations for until the mid-1980s, ensuring the system's reliability during its formative years. This adoption underscored the system's role in promoting fair play, as equal-skill pairings minimized lopsided games and maximized competitive integrity. In the 1970s and 1980s, the system extended into early computing applications beyond human tournaments, influencing simulations and evaluations. Computer chess programs, competing in events like the North American Computer Chess Championship starting in 1970, were assigned ratings to gauge their performance against each other and human opponents, facilitating structured pairings in these nascent digital contests. The brought a pivotal shift to fully networked digital platforms, adapting ratings for real-time online matchmaking. The (ICS), established in 1992, was among the first to integrate the system for player rankings, automatically pairing users based on their ratings during internet-based games. This development transformed the analog rating concept into a cornerstone of virtual competition, enabling global access to skill-balanced play.

Adoption in Video Games

Skill-based matchmaking in video games began to gain traction in the early 2000s, building on foundational rating systems like , which originated in chess and influenced adaptations for online multiplayer environments. introduced , a Bayesian skill rating system, in 2006 through its research division, deploying it for Xbox Live matchmaking to enhance player pairing beyond the simpler 1-50 ranking used in multiplayer. This marked an early milestone in integrating probabilistic models for more accurate skill assessment in console gaming. In 2008, id Software's highlighted the importance of skill-based matchmaking during development discussions for , emphasizing its role in pairing players of comparable ability to retain newcomers against seasoned competitors. By the mid-2010s, adoption expanded across major titles leveraging online platforms. implemented skill-based queues in upon its 2009 launch, using a hidden matchmaking rating (MMR) derived from principles to form balanced teams in normal and ranked modes. integrated SBMM into in 2013 via its Steamworks platform, introducing ranked matchmaking that required about 150 games to unlock and relied on MMR for equitable opponent selection in modes like All Pick and Captain's Mode. followed suit with in 2016, launching Competitive Play in Season 2 that incorporated skill tiers and MMR-based matchmaking to group players by relative proficiency. Key events underscored evolving priorities in implementation. opted against full SBMM in (2010), focusing instead on connection quality to minimize queue times and prioritize accessible multiplayer sessions over strict skill balancing. This approach shifted in Call of Duty: Black Ops II (2012), where and revamped matchmaking to emphasize and exclusively, moving away from locks to ensure low-latency games while deprioritizing as the primary factor. Recent developments through 2025 reflect ongoing experimentation and refinements. In Pokémon UNITE's Season 23 (late 2024), The Pokémon Company announced a shift away from traditional MMR-based , transitioning to a system reliant solely on Master rank ratings starting in Season 24 to streamline high-level play and reduce wait times. revamped Deadlock's MMR system in December 2024, unifying matchmaking pools across normal and ranked modes while incorporating hero-specific strengths and weaknesses for more nuanced pairings. In contrast, introduced OG's Expert Duos mode in 2024, deliberately excluding SBMM and bots to recreate the unstructured, high-stakes lobbies of early seasons.

Technical Mechanisms

Skill Rating Systems

Skill rating systems form the foundational component of skill-based matchmaking by assigning numerical values to players' abilities, enabling fair opponent pairing. These systems update ratings based on game outcomes, incorporating factors such as win/loss results, opponent strength, and uncertainty to reflect skill more accurately over time. Primary models like and Glicko provide straightforward mechanisms for individual player assessment, while advanced variants such as and the Cornell model extend capabilities to handle team dynamics, draws, and contextual intransitivities. The , developed by in the for chess, represents one of the earliest and most influential approaches. It operates on a logistic model where a player's rating adjusts after each game based on the actual outcome compared to the expected outcome against the opponent. The core update formula is: R_{\text{new}} = R_{\text{old}} + K \times (S - E) Here, R_{\text{new}} and R_{\text{old}} are the updated and prior ratings, K is a constant factor determining adjustment sensitivity (typically 32 for beginners and lower for experts), S is the actual score (1 for win, 0.5 for draw, 0 for loss), and E is the expected score calculated as: E = \frac{1}{1 + 10^{(R_b - R_a)/400}} where R_a and R_b are the ratings of the player and opponent, respectively. This system assumes binary or ternary outcomes and promotes convergence toward stable ratings through iterative updates, though it does not explicitly model uncertainty in new or infrequent players. Building on Elo's limitations, the Glicko system, introduced by Mark Glickman in 1995, incorporates rating deviation (RD) to quantify uncertainty and volatility in skill estimates. RD starts high for new players (e.g., 350 points) and decreases with consistent performance, allowing conservative adjustments for those with limited data. The update process uses a Bayesian framework to revise both the rating and RD based on game results, opponents' ratings, and their deviations, making it more robust for sparse match histories. Glicko-2, an adaptation, introduces a per-player volatility parameter to model potential skill changes over time and extends handling of non-binary outcomes, such as multi-player scores. TrueSkill, developed by in 2006, advances rating through a fully Bayesian probabilistic model suitable for team-based games and draws. Each player is represented by a distribution with \mu (average ) and standard deviation \sigma (uncertainty), initialized at \mu = 25 and \sigma = 25/3 for conservatism. After a match, posterior distributions are computed using approximate in a , accounting for team performance and partial outcomes without requiring exact differences. This enables handling of multi-player scenarios and draws by modeling win probabilities via a Gaussian on differences. TrueSkill's design ensures scalability for online platforms like Xbox Live, emphasizing uncertainty reduction over time. The Cornell model, proposed by researchers at Cornell University in 2016, addresses intransitivity in matchups by representing skill as a vector rather than a scalar, capturing context-specific strengths such as performance on particular maps or against certain playstyles. Using a Bradley-Terry-like framework extended to vectors, it learns pairwise preference relations from matchup data, allowing predictions of non-transitive outcomes (e.g., rock-paper-scissors dynamics in games). This vector approach models skill as a low-dimensional embedding, enabling more nuanced ratings that adapt to environmental factors without assuming total orderings.

Matching Algorithms

Skill-based matchmaking algorithms utilize player skill ratings, typically derived from systems like or , to pair individuals or teams in online s. The core process begins with queue formation, where players entering a matchmaking are categorized by mode, region, and availability. Algorithms then apply skill bucketing to group players within narrow rating ranges, such as ±50 points, to ensure competitive balance while minimizing wait times. For team-based s, balancing occurs by sorting queued players by skill and assigning them alternately to opposing teams, often using a greedy approach that adds each subsequent player to the team with the current lowest aggregate skill sum to prevent imbalances. Beyond primary skill considerations, algorithms incorporate secondary factors to enhance match quality and player retention. Latency () is frequently prioritized to avoid network disadvantages, with thresholds like under 100 enforced for pairings. Other elements, such as playstyle preferences or capabilities, may influence selections in specialized systems, though skill remains dominant. Trade-offs between match fairness and wait times are managed dynamically; for instance, broader skill buckets expand during low population periods to reduce queues from minutes to seconds, as modeled by formulas estimating wait time based on online player count and bucket granularity. Common algorithms include matching, which iteratively pairs the highest-rated available player with a suitable opponent within a predefined gap, enabling rapid resolution suitable for high-volume games. For more optimized outcomes, techniques formulate matchmaking as a minimum weight problem on a of player nodes, where edges represent predicted match quality (e.g., or engagement risk), solved in polynomial time to maximize overall balance across multiple pairs. A simple example for Elo-based pairing in a 1v1 might proceed as follows:
function greedyPairing(queue):
    sort queue by Elo descending
    matches = []
    i = 0
    while i < queue.length - 1:
        player1 = queue[i]
        for j = i+1 to queue.length:
            player2 = queue[j]
            if abs(player1.Elo - player2.Elo) <= threshold:
                matches.add((player1, player2))
                remove player1 and player2 from queue
                break
        i += 1
    return matches
This approach ensures pairs stay within skill thresholds but may leave outliers unpaired if queues are sparse. To handle performance variability, algorithms apply volatility adjustments, such as higher K-factors in systems for new players (e.g., K=32 for beginners versus K=16 for veterans), accelerating convergence toward true without overreacting to noise. New player protections mitigate initial rating uncertainty by confining matches to provisional pools or inflating buffers, allowing novices several games (e.g., 5-10) before full integration, as uncertainty in skill estimates diminishes rapidly with experience. These mechanisms, implemented in platforms like Xbox Live, promote equitable progression while sustaining engagement.

Implementations

In Esports Titles

Skill-based matchmaking (SBMM) plays a pivotal in titles, where competitive integrity is paramount for fair ladder progression, tournament qualification, and professional play. In these environments, SBMM systems pair based on metrics to ensure balanced that reflect true ability rather than luck or mismatched opponents, often integrating with broader features like leaderboards and anti-cheat mechanisms. League of Legends, since the introduction of its ranked queues in late 2009 shortly after the game's launch, has employed SBMM through a combination of visible and a hidden Matchmaking Rating (MMR). Players earn or lose LP based on match outcomes, with MMR adjusting dynamically to form teams of comparable skill within tiered divisions, from Iron to , facilitating precise for millions of competitive users. This system underpins the game's scene, including the , by providing reliable skill assessments for draft picks and seeding. Similarly, introduced its MMR system in December 2013 to enhance ranked matchmaking, assigning numerical ratings that increase with wins and decrease with losses to create equitable games. In esports contexts, such as The International tournaments, MMR serves as a key metric for player seeding and team composition, ensuring high-stakes matches feature appropriately skilled competitors and reducing variance in outcomes. The system uses Glicko-2 methodology for variance estimation, which helps in calibrating new players accurately for tournament eligibility. Overwatch's Competitive mode, launched on June 28, 2016, utilizes tiers ranging from to , with SBMM grouping players by SR to promote skill-congruent teams in 6v6 matches. This setup supports the by enabling precise role-based and progression tracking, where SR adjustments reflect performance modifiers beyond mere wins. In , released in June 2015, SBMM adaptations emphasize rapid queue times while targeting balanced engagements, resulting in approximate 50% win rates as players converge to their true skill level through iterative MMR adjustments. Blizzard's design encourages this equilibrium without explicit forcing, prioritizing fair compositions in hero-based battles for competitive ladders. Esports SBMM implementations have faced exploits, as seen in Destiny 2's PvP modes around 2022, where players manipulated connection drops to force rematches with lower-skill lobbies, leading to patches that tightened SBMM parameters and introduced penalties for intentional disconnects to preserve match integrity. Unique to esports, SBMM integrates with spectator tools by generating predictably competitive matches that heighten viewer engagement, as balanced games reduce blowouts and highlight strategic depth in broadcasts. Anti-smurfing measures, such as mandatory account verification and behavioral analysis in titles like League of Legends and Dota 2, detect and penalize alternate accounts to maintain ladder purity. Regional server balancing further refines global tournaments by routing players to low-latency data centers based on skill pools, minimizing ping disparities during events like Worlds or majors. A notable recent development occurred in Pokémon's competitive ecosystem in December 2024, when updates to Pokémon TCG Live's Seasonal shifted higher tiers to an Elo-based ranking system for the Arceus League, aiming to create fairer and more accurate stratification for qualifiers and regional championships.

In Casual Games

In casual games, -based (SBMM) is implemented to foster engaging experiences for broad audiences by pairing players of comparable abilities in non-competitive modes, thereby supporting retention without the intensity of ranked play. For instance, in : Warzone's public matches launched in 2020, SBMM evaluates player performance metrics such as kills, deaths, and win rates to form balanced lobbies, prioritizing fair competition alongside low connections. Similarly, Hearthstone's casual queues, introduced with the game's 2014 launch, utilize a hidden matchmaking rating (MMR) system to match opponents of similar levels, incorporating factors like win streaks to adjust pairings and prevent prolonged dominance. Fortnite's core battle royale modes, such as solos and duos, have employed SBMM since a 2019 update to create equitable matches based on eliminations, survival times, and overall performance, with the system dynamically incorporating bots for newer players to ease entry. These implementations often feature lighter variations to emphasize fun and social interaction; for example, Roblox's matchmaking framework, enhanced in 2023, allows developers to weight social connections—such as friends or preferred groups—over strict skill metrics in user-generated experiences, promoting collaborative play in casual environments. Some titles provide opt-out mechanisms or relaxed SBMM in casual playlists, as seen in XDefiant's unranked modes, where developers explicitly omitted skill considerations to prioritize quick, varied matches. To balance accessibility, SBMM in party modes typically averages the group's skill levels rather than enforcing individual precision, enabling mixed-ability friends to enjoy casual sessions without severe imbalances. In , the October 2025 update introduced split matchmaking for multiplayer battles, with casual Regular Battles focused on loot collection and no trophy loss, alongside a separate Ranked Battles option (from 7) that uses levels for skill-based matching in competitive play. This approach contrasts with modes designed for variety, such as Fortnite's OG Expert Duos introduced in July 2025, which eliminates SBMM and bots to deliver unpredictable, old-school matchmaking for players seeking diversity beyond balanced queues.

Advantages and Challenges

Benefits

Skill-based matchmaking (SBMM) enhances match quality by pairing players of comparable skill levels, resulting in more balanced games and approximately 50% win rates for most participants. This balance minimizes one-sided matches, or "blowouts," where score differentials exceed 30 points, fostering a sense of fairness that encourages continued play. In games like , tightening skill constraints in matchmaking leads to 80-90% of players achieving better end-of-match placements and lower quit rates, directly boosting retention across skill distributions. SBMM also provides clear visibility into skill progression through accurate rating systems, allowing players to track improvements and set achievable goals. Microsoft's algorithm, deployed in titles, achieves 68% accuracy in predicting match outcomes, enabling precise skill assessments that highlight personal growth and motivate ongoing participation. By creating equitable lobbies, SBMM contributes to lower churn rates when opponent skill variance is minimized, as balanced matches reduce player dropout compared to highly mismatched games. For newcomers, this setup accelerates learning curves, as players face appropriate challenges that build competence without discouragement, leading to faster skill acquisition and higher engagement. These dynamics sustain by extending play sessions; balanced matches encourage longer , with studies showing players stick with titles longer when skill is prioritized, supporting the long-term health of online communities.

Criticisms

One prominent criticism of skill-based matchmaking (SBMM) is that it creates overly competitive "sweaty" environments for skilled players, denying them occasional easy wins and turning casual sessions into high-stakes battles. In games like : III, the system's focus on "perfect matches" results in lobbies with minimal skill variance, often leading to narrow score margins and increased stress without the relaxation of mismatched opponents. Similarly, top-tier players report constant intensity, as the algorithm prioritizes even team compositions over diverse experiences. SBMM also incentivizes smurfing, where experienced players create alternate accounts to bypass the system and access easier lobbies. This behavior arises because high-skill individuals face relentless challenges in standard queues, prompting them to seek casual play through new profiles, which undermines fair for beginners. Prior to 2025, the absence of non-SBMM casual modes in many titles exacerbated this, as players resorted to smurfing for variety rather than enduring perpetual try-hard encounters. Longer queue times, particularly in low-population regions or for elite , represent another key drawback. High-skill users often wait extended periods due to the scarcity of suitable opponents, with former Halo multiplayer lead Max Hoberman describing this segregation as "a form of " that isolates top performers from the broader base. In titles like , prioritizing skill over availability can inflate waits, leading to abandon queues and fragment the further. Controversies have highlighted these issues, notably streamer backlash in Call of Duty: Warzone during its 2020 launch. Influencer publicly demanded "bot lobbies" and a ranked mode, arguing that SBMM's strict enforcement made streaming untenable by forcing constant high-level competition without casual alternatives. His complaints amplified community frustration, dividing players over whether SBMM stifles content creation and enjoyment. In , perceived manipulation via SBMM exploits emerged around 2022-2023, allowing skilled fireteams to infiltrate low-skill PvP lobbies through matchmaking flaws introduced in Season 18, resulting in one-sided stomps that frustrated newcomers. Technical flaws include inaccurate skill ratings for new players, as algorithms rely on limited performance data, potentially misplacing them in mismatched games until sufficient matches accumulate. Matching algorithms can also force trade-offs between skill balance and connection quality, prioritizing low ping but occasionally pairing players with distant servers, leading to laggy matches when ideal skill pools are unavailable. Privacy concerns arise from the extensive tracking of stats required for SBMM, raising fears of misuse. Activision's 2021 shutdown of the SBMM Warzone stat-tracking site cited violations of laws in the and , emphasizing risks from unauthorized access to detailed performance metrics. Demands for SBMM toggles have grown, as seen in 2024 OG discussions, where criticized bot-adjusted lobbies for eroding nostalgic play and called for options to customize matchmaking rigor. In response to ongoing criticisms, as of November 2025, 7 has implemented changes to SBMM, making it no longer the default. The game features open playlists for casual multiplayer without skill-based considerations, alongside standard modes that retain SBMM, aiming to provide variety, reduce queue times for high-skill players, and mitigate sweaty lobbies while preserving competitive balance where desired.

Community Response

The community response to skill-based matchmaking (SBMM) has been notably polarized, with casual players often praising it for fostering balanced and engaging matches that reduce from mismatched opponents, while experienced or players frequently criticize it for creating overly competitive "sweaty" lobbies that diminish the fun of casual play. This divide became particularly evident in the late and early , as SBMM's implementation in major titles amplified debates over player retention versus enjoyment. A prominent example of frustration among veterans occurred in 2020 with Call of Duty titles, where players launched petitions demanding the removal of SBMM, arguing it forced reliance on meta strategies and hindered social play with friends of differing skill levels. Social media trends on platforms like and further highlighted this discontent, with hashtags and threads decrying SBMM as overly punitive, leading developers to publicly address the feedback by explaining its role in combining skill with factors like . In contrast, the esports community around has shown stronger support for SBMM, particularly in ranked queues, where it is viewed as essential for fair competition and skill progression, with ' matchmaking system routinely praised in forums for maintaining balanced 50% win rates. Data from Activision's experiments underscores the approval among casuals, revealing that deprioritizing led to an 80% increase in quit rates across 80% of players, with reduced retention for the bottom 90% of skilled players (primarily lower-skilled ones), indicating broad retention benefits in non-esports contexts. Streamers have amplified these criticisms, influencing public discourse and contributing to viral trends that pressure developers, though quantitative polls remain limited. Cultural differences also shape responses, with Asian gaming communities—particularly in titles like —embracing SBMM's ranked focus due to a stronger emphasis on competitive progression and higher overall skill levels compared to Western audiences, where casual relaxation often takes precedence. This evolution reflects a shift from relative acceptance in the , when SBMM was seen as a novel fairness tool in early implementations like : Advanced Warfare, to intensified 2020s debates fueled by larger player bases, influencer voices, and data-driven defenses from publishers.

Emerging Developments

In 2024, significantly overhauled the and MMR system in its MOBA game to address imbalances and improve fairness, introducing AI-enhanced predictions that account for individual player strengths and weaknesses across different heroes. This update unified matchmaking pools into a single mode, eliminating separate normal and ranked queues, and incorporated more nuanced skill assessments to create balanced teams without time restrictions on queuing. Hybrid systems blending strict SBMM with random elements have gained traction as developers seek to balance competitive integrity with gameplay variety. For instance, Call of Duty: Black Ops 7, released on November 14, 2025, launched with "open " as the default in most playlists, minimizing skill-based considerations to allow for more randomized lobbies while offering persistent lobbies and optional ranked modes (with standard SBMM) for those preferring structured matches. This approach represents a shift toward hybrid models that reduce the rigidity of traditional SBMM, responding to player feedback on overly predictable experiences. Emerging trends include greater player control through opt-in or features for SBMM intensity, as seen in Black Ops 7's default open system that effectively opts players out of heavy skill weighting unless they choose ranked play. Cross-platform unification efforts continue to evolve, with games like maintaining SBMM across console and PC pools to ensure consistent experiences despite hardware differences, though challenges in unifying skill metrics persist. Additionally, Pokémon Unite's Season 23 update in December 2024 shifted from pure MMR-based SBMM to a rank-only system using rank ratings, aiming to simplify and reduce over-reliance on hidden skill calculations for more transparent competition. Looking ahead, future developments may see a decline in strict SBMM to prioritize variety, exemplified by Fortnite's 2024 OG mode update, which reintroduced bots into lobbies for lower-skilled players alongside SBMM to foster more diverse and less punishing matches. Ethical integration for bias reduction is also on the horizon, with 2025 research highlighting the need for fair algorithms in SBMM to mitigate discriminatory outcomes in player pairing, particularly in AI-driven systems used in fantasy sports and multiplayer games. Projections for 2025 suggest continued experimentation, such as potential behavioral metric expansions beyond traditional stats—like engagement patterns and decision-making styles—to refine matchmaking in titles like .

References

  1. [1]
    [PDF] TrueSkill 2: An improved Bayesian skill rating system - Microsoft
    Mar 22, 2018 · Abstract. Online multiplayer games, such as Gears of War and Halo, use skill-based matchmaking to give players fair and enjoyable matches.
  2. [2]
    [PDF] The Role of Skill in Matchmaking - Activision
    Skill is used in matchmaking to give players a chance to impact matches, and is based on expected performance against the population, calculated after each ...
  3. [3]
    [PDF] Ranking and Matchmaking - Ralf Herbrich
    Skill based ranking and matchmaking are crucial elements for designing competitive yet enjoyable online gaming experiences. They are essential tools of game ...
  4. [4]
    [PDF] Beyond Skill Rating: Advanced Matchmaking in Ghost Recon Online
    Advanced matchmaking uses more than just skill, considering player behavior, preferences, and multiple skills, to maximize player fun and balance, not just  ...
  5. [5]
    What is skill-based matchmaking and why do streamers hate it?
    May 27, 2022 · Skill-based matchmaking is a system multiplayer games typically use to place players of similar skill levels in matches against each other to fairly balance ...
  6. [6]
    (PDF) Achieving fairness in team-based FPS games - ResearchGate
    To pursue a fair user experience, matchmaking mechanisms typically try to put players with similar skill levels into the same game. The traditional process ...
  7. [7]
    [PDF] Skill-Based Matchmaking for Competitive Two-Player Games
    In this paper, we present methods for skill-based matchmaking for competitive two-player games that involve two distinct components: (1) A novel rating ...
  8. [8]
    Call of Duty's Skill-Based Matchmaking - Insights for Multiplayer ...
    Activision has found that integrating skill into matchmaking improves player retention, with 80-90% of players experiencing a better match outcome.
  9. [9]
    Who Was Chess Master Arpad Elo, and What is the Elo Rating ...
    Nov 5, 2021 · The Elo system was adopted in 1960 by the US Chess Federation in Saint Louis. The International Chess Federation (FIDE) adopted it in 1970.
  10. [10]
  11. [11]
    [PDF] A Comprehensive Guide To Chess Ratings Prof. Mark E. Glickman ...
    The method Elo laid out for adjusting ratings was adopted by the USCF in 1960, and subsequently adopted by FIDE in 1970. Through the years, various ...<|control11|><|separator|>
  12. [12]
    1 July 1971: FIDE introduces Elo ratings - ChessBase
    Jul 1, 2025 · The new Elo system was adopted by the USCF in 1960 and by the World Chess Federation (FIDE) in 1970. For many years, Arpad Elo manually ...
  13. [13]
    Anniversary of Arpad Elo – rating system that changed chess world
    Aug 25, 2025 · That same year, FIDE officially adopted Elo's system for calculating chess ratings. Elo pioneered this rating system, which the US Chess ...
  14. [14]
    Engine Rating Lists - Chessprogramming wiki
    The classical SSDF rating list, sanctioned by the ICGA, and already started in the 80s with dedicated chess computers is the most established and acknowledged.
  15. [15]
    Historic trends in chess AI - AI Impacts
    The Elo rating of the best chess program measured by the Swedish Chess Computer Association did not contain any greater than 10-year discontinuities between ...
  16. [16]
    The History of Internet Chess Servers
    Feb 8, 2022 · The first internet chess server was creatively named Internet Chess Server or ICS for short and was programmed by Michael Moore and Richard Nash ...
  17. [17]
    QuakeCon 2008: Quake Live First Look - GameSpot
    Jul 31, 2008 · A matchmaking system will ensure new players are matched with players of similar skill, so they don't go running for the hills when pro players ...
  18. [18]
    Matchmaking and Autofill - League of Legends Support - Riot Games
    Jul 19, 2023 · To the matchmaker, a “fair” match is one where each team has a 50% +/-1% chance of winning, which we create by pairing off teams of roughly the ...Matchmaking · High-Rank Opponents · Autofill
  19. [19]
    Ranked Matchmaking Coming to Dota 2 - IGN
    The next big update to Dota 2 will introduce ranked matchmaking features to the game. Valve said on its official blog the mode is aimed at experienced players.Missing: launch | Show results with:launch
  20. [20]
    Overwatch PTR Patch Notes – August 17, 2016
    Aug 24, 2016 · The most noticeable change is the introduction of skill tiers, which we hope will better communicate players' relative skill levels.
  21. [21]
    Call of Duty: Black Ops multiplayer takes aim at cheaters, looks to ...
    Sep 3, 2010 · Another feature that Treyarch discussed, but didn't include, in Black Ops was a clan mode. Like skill-based matchmaking, Vonderhaar said clan ...
  22. [22]
    Call of Duty: Black Ops 2 uses ping and latency "exclusively" for ...
    Oct 26, 2012 · Treyarch Design Director David Vonderhaar revealed that Black Ops 2 deviates from standard procedure and matches players via ping and latency exclusively.
  23. [23]
    Season 23 - Matchmaking Changes - Pokémon Forums
    Dec 3, 2024 · We're moving away from MMR-based matchmaking and instead switching to a system that only uses Master rank ratings. This aims to make things more ...
  24. [24]
    Deadlock - Valve Revamps Ranking and MMR System - EGW-News
    Dec 8, 2024 · Valve announced the update on the testers' forum for the new game. The most striking change is the removal of traditional ranked matches.
  25. [25]
    Fortnite OG Gets New Version Without SBMM Or Bots - GameSpot
    Jul 29, 2025 · And with this latest patch, OG got a retro matchmaking mode with no skill-based matchmaking and no bots. OG Duos just went Expert mode! Wanna ...
  26. [26]
    [PDF] The Glicko system - Mark Glickman
    In 1995, I created the Glicko rating system in response to a particular deficiency in the Elo system which I describe be- low. My system was derived by ...
  27. [27]
    [PDF] TrueSkillTM: A Bayesian Skill Rating System - Microsoft
    In this paper we present a new rating system, TrueSkill, that addresses both these challenges in a principled Bayesian framework. We express the model as a ...
  28. [28]
    [PDF] Modeling Intransitivity in Matchup and Comparison Data
    Feb 25, 2016 · ABSTRACT. We present a method for learning potentially intransitive prefer- ence relations from pairwise comparison and matchup data. Un-.
  29. [29]
    [PDF] EOMM: An Engagement Optimized Matchmaking Framework
    In this paper, we propose an Engagement Optimized Matchmak- ing (EOMM) framework that maximizes overall player engage- ment. We prove that equal-skill based ...<|separator|>
  30. [30]
    Skill Based Matchmaking Explained - PubNub
    Nov 26, 2024 · Skill-based matchmaking (SBMM) pairs players of similar skill levels, using performance metrics to group players with comparable abilities and ...
  31. [31]
    Ranked Tiers, Divisions, and Queues - League of Legends Support
    Aug 27, 2024 · Ranked players are divided into ten tiers, shown above in ascending order. Each one houses players of similar skill levels and follows a simple rule: the more ...
  32. [32]
    League Of Legends Ranking System Explained - Red Bull
    May 10, 2020 · Matchmaking Rank (MMR) is a number Riot assigns to each player based on their skill level that determines how much LP that player gains or loses ...
  33. [33]
    How Does Dota 2's Ranked Matchmaking Work? - Esports Edition
    Nov 24, 2017 · MMR was introduced to Dota 2 in December 2013, and it has served as the primary in-game status symbol for the community ever since. Some Dota ...
  34. [34]
    Dota 2 Ranks Explained (2025): Complete MMR & Ranking System ...
    Oct 23, 2025 · Learn everything about Dota 2 ranks, MMR, and the ranking system, including the different rank medals, calibration, Glicko, and tips to ...Missing: seeding | Show results with:seeding
  35. [35]
    Competitive Play - Overwatch Wiki - Fandom
    Per-Hero Skill Rating. Starting with Season 18, players can start gaining per-Hero SR to present their skill at individual heroes. The rules are outlined ...
  36. [36]
    Overwatch's Skill Rating System - A Mistake? - YouTube
    Jul 21, 2016 · Is Overwatch's competitive play skill rating system a mistake? It's ... 2016/07/19/overwatch-competitives-percentile-based-skill-rating ...<|separator|>
  37. [37]
    Heroes of the Storm doesn't force a 50% winrate, merely encourages it
    it won't match you with lesser-skilled players just to drag you down. However, it ...
  38. [38]
    Destiny 2 PvP exploit sees skill-based matchmaking completely ...
    Aug 29, 2022 · Some lower-ELO players are having a rough time with a Destiny 2 PvP exploit, seeing high-level opponents flood their games due to a SBMM ...
  39. [39]
    Letter to the Community - December 3, 2024 - Pokémon Forums
    Oct 9, 2024 · Starting with this update, players who reach Arceus League in a Seasonal Ladder will transition from earning Rank Points to competing in an Elo- ...More feedback about the new update to rankedMajor complaint with the Ranked LadderMore results from community.pokemon.com
  40. [40]
    FAQ: Ranked Mode Overhaul Matchmaking - Blizzard Forums
    Apr 1, 2020 · We put together a general matchmaking FAQ here for our new ranked system to help answer questions you may have.
  41. [41]
    Fortnite Matchmaking Update Battle Royale
    Sep 23, 2019 · In the v10.40 update, we're introducing improved matchmaking logic to Battle Royale core modes to create fairer matches.
  42. [42]
    Inside the Tech - Solving for Matchmaking on Roblox
    Oct 11, 2023 · Matchmaking builds the services that match Roblox users to an experience server in the join process. When someone wants to visit a Roblox ...
  43. [43]
    [[UPDATED]] Call of Duty: An Inside Look at Matchmaking
    Apr 4, 2024 · This blog focuses on how matchmaking works across Multiplayer only. We will be continuing the conversation about matchmaking in other modes ( ...
  44. [44]
    Get Ready for Ranked! - Update × Clash of Clans - Supercell
    Oct 6, 2025 · Multiplayer matchmaking has been split into two modes so you can battle your way! Test your skills and earn Trophies in Ranked Battles or ...
  45. [45]
    Match experiences affect interest: Impacts of matchmaking and ... - NIH
    Jan 20, 2024 · We found that matching players with highly disproportionate skills caused greater churn. Better performance, characterized by higher win rates ...
  46. [46]
    Admit It, You Don't Understand Skill-Based Matchmaking ... - Kotaku
    Nov 21, 2023 · Skill-based matchmaking and the many side effects it has on everyone's multiplayer sessions is not a simple issue. It's not just that top-tiers ...Missing: video | Show results with:video
  47. [47]
    How Call of Duty's SBMM Controversy Divided a Toxic Multiplayer ...
    Oct 20, 2020 · A seemingly simple way to help people play together has instead divided an increasingly hostile Call of Duty community.
  48. [48]
    former Halo multiplayer lead on the 'failure' of SBMM in modern games
    Nov 20, 2023 · Former Bungie multiplayer lead Max Hoberman argues modern multiplayer games do a "disservice" to players by only serving them even matchups.
  49. [49]
    Nadeshot Frustrated With Warzone's Skill Based Matchmaking
    Mar 26, 2020 · Nadeshot took to Twitter to voice his desire for ranked in Call of Duty: Warzone and a distaste for skill based matchmaking.
  50. [50]
    Destiny 2 Exploit Lets Top Players Troll Newbies - TechRaptor
    Aug 29, 2022 · A Destiny 2 exploit has seen low-skill fireteam lobbies get obliterated by higher-ranked fireteams gaining access to low-ranked games.
  51. [51]
    Activision shuts down Warzone's largest stat tracking site - PC Gamer
    Mar 29, 2021 · Activision's complaint seems, primarily, to be a concern with a potential breach of privacy. Monetisation is not mentioned, but SBMM Warzone ...
  52. [52]
    Fortnite OG's SBMM Controversy Explained - Game Rant
    Dec 13, 2024 · A lot of controversy currently surrounds Fortnite OG thanks to how skill-based matchmaking has impacted the mode, sparking debate around how the game has ...
  53. [53]
    Why Players Blame Skill-Based Matchmaking for Losing in Call of Duty
    Dec 2, 2020 · Skill-based matchmaking, as you can guess, is a type of multiplayer matchmaking system in which players' are pitted against other players of similar skill ...
  54. [54]
    Petition · Remove SBMM Before Launch - United States · Change.org
    Sep 19, 2020 · Hello, I'm starting a petition to remove SBMM, in favor of having unranked and ranked gameplay in the next Call of Duty in the series; ...
  55. [55]
    Call of Duty Devs Finally Comment on Skill-Based Matchmaking ...
    Nov 30, 2023 · Players have long said that the system makes it hard for solo players to find enjoyable matches, and that it makes it hard to find good games ...Missing: criticisms | Show results with:criticisms
  56. [56]
    The Case For and Against Skill-Based Matchmaking in Destiny 2
    Oct 13, 2023 · SBMM is defined as a type of matchmaking that takes relative player skill into consideration first and foremost. Likewise, connection-based ...
  57. [57]
    Activision Secretly Turned Off Skill-Based Matchmaking in Call ... - IGN
    Jul 30, 2024 · With deprioritized skill, returning player rate was down significantly for 90% of players. The 10% of highest skilled players came back in ...Missing: criticisms | Show results with:criticisms
  58. [58]
    Activision report claims players don't hate SBMM as much as they ...
    Jul 29, 2024 · Meanwhile, quit rates in matches spiked by about 80 percent during the test. Activision quit rate chart from its deprioritized SBMM test.
  59. [59]
    Gaming Culture in Asia vs. The West - MouseOne
    Jun 21, 2023 · One striking difference between Asian and Western gaming culture is the influence of cultural aesthetics and preferences on game genres. Asia ...
  60. [60]
    Do SBMM Algorithms Ruin Gaming? - Data Behind Skill ... - YouTube
    Oct 29, 2020 · SBMM is one of the most debated issues in gaming today, but does the data back it up? Subscribe for more in-depth gaming videos ...
  61. [61]
    Valve gets closer to perfecting Deadlock's matchmaking, which will ...
    Nov 22, 2024 · The matchmaking pools are no longer split between normal and ranked, there is only one primary matchmaking mode, and there are no limited hours.
  62. [62]
    Activision finally gives in and confirms Call of Duty Black Ops 7 will ...
    Oct 9, 2025 · Activision finally gives in and confirms Call of Duty Black Ops 7 will not use SBMM: "At launch, Open Matchmaking with minimal skill ...
  63. [63]
    Black Ops 7 will launch with open matchmaking. Minimal skill ...
    Oct 9, 2025 · Removing the SBMM turns it into a more casual experience, which is probably better for a game with such fundamental imbalances. Let's just say ...
  64. [64]
    So what's the verdict on SBMM now that the game has been out for a ...
    Oct 14, 2025 · There's obviously sbmm in this game. It's not as strong as CoD but it's noticeable. I play on PS5 and when I enable crossplay it's a sweatfest ...
  65. [65]