By MaJ — Gaming systems analyst with 18+ years of industry experience. Full bio
“The database is the game.” That is what every veteran FM player will tell you. They are not exaggerating. The game’s database, compiled by thousands of volunteer researchers, covers over 500,000 players across more than 150 countries. Each player is represented by attributes spanning technical, mental and physical categories.
Football Manager 2026 is the first entry in the franchise’s history to hold an official FIFA license, granting Sports Interactive access to major tournaments. This is not merely a branding change. It restructures what the game can simulate and, more importantly, what it can predict.
What Hundreds of Thousands of Players Actually Means
FM2026’s database covers an enormous number of players across leagues in many countries. Each player is represented by multiple attributes spanning technical, mental and physical categories, all rated on a scale. The compilation process is run by a network of volunteer researchers who attend matches, review match footage, and maintain local contacts to produce and maintain attribute assessments.
Researchers are assigned geographic and league territories. A researcher covering a specific league will follow multiple clubs, attending matches where possible and cross-referencing observations against statistical outputs. This is not dissimilar from the work a professional scout does, except that the researcher is producing attributes for every player in their territory.
A study tested this by comparing FM’s attribute predictions against actual performance outcomes for players who transferred clubs. The study found a statistically significant correlation between FM’s predicted career trajectories and real-world outcomes. Several commercial scouting platforms scored lower on the same metric. FM outperformed real scouting tools.
The key insight is that the mental attributes, which have no direct real-world equivalent, function as composite predictors that aggregate observable behaviors across time. A player with high Composure and Decisions consistently overperforms expected output under pressure in actual match data.
The Volunteer Research Network
The scale of Sports Interactive’s volunteer research operation is difficult to comprehend unless you have seen it from the inside. The database that powers Football Manager does not come from a licensing deal or a data vendor. It comes from a global network of over 1,400 volunteer researchers who cover every professional league on Earth. This is not a team of paid employees. This is a community of football obsessives who treat FM’s database as a serious civic responsibility.
The recruitment process for researchers is selective and structured. Applicants must demonstrate detailed knowledge of their target league, submit writing samples, and complete a trial period where their attribute assessments are cross-checked against existing data and match footage. The rejection rate is substantial. Sports Interactive is not looking for fans. They are looking for people who can produce scouting-grade assessments without payment, consistently, across multiple seasons.
Each researcher is assigned a territory—usually a league or a geographic region—and is responsible for maintaining the data for every player in that territory. The work involves attending matches in person when possible, reviewing broadcast footage when not, and maintaining a network of local contacts who can provide information about training performances, dressing room dynamics, and injuries that never make the official injury reports. A researcher covering the Polish Ekstraklasa, for example, might be filing weekly reports on thirty different players across multiple clubs, updating attributes as new information becomes available.
The attribute system itself deserves explanation. FM uses a 20-point scale across technical, mental, and physical categories. Technical attributes are relatively straightforward: Finishing, Passing, Tackling. Physical attributes are measurable: Pace, Strength, Stamina. The mental attributes are where the system becomes genuinely sophisticated. Composure, Decisions, Anticipation, and Vision do not correspond to anything you can measure with a stopwatch or a radar gun. They are inferred from pattern recognition—from watching how a player behaves under repeated situational stress and encoding that behavior into a number.
The validation process is continuous. Senior researchers and SI staff periodically audit attribute assignments by cross-referencing researcher reports against match data and, when possible, against actual scouting reports from professional clubs. The system is not perfect. Researchers have blind spots. A researcher who supports a specific club in their territory may unconsciously overrate players from that club. SI manages this through overlapping assignments and periodic re-audits, but the fundamental problem of human bias in data collection remains.
The database’s growth trajectory tells its own story. In FM2005, the database contained approximately 100,000 players. By FM2026, that number has grown to over 500,000 players across more than 150 countries. The expansion was not driven by licensing agreements. It was driven by the research network’s gradual extension into lower divisions, women’s football, and non-league systems that no commercial scouting service covers. If you are managing a semi-professional club in the Swedish fourth tier, FM has data on your players. No other football database on Earth has that coverage.
There is a darker side to this system that deserves mention. The volunteer researchers are not employees. They have no job security, no benefits, and no formal negotiating power with Sports Interactive. When SI introduced changes to the research platform or the attribute system, researchers had no formal mechanism to contest those changes beyond community forums. Some researchers have raised concerns about data ownership—the thousands of hours they have invested in building the database belong to SI, which monetizes that data in a commercial product. Sports Interactive has historically relied on the intrinsic motivation of its research community: the pride of contributing to the most comprehensive football database ever built. Whether that is sustainable as the commercial value of the database increases is an open question.
The research network also creates geographic inequalities in data quality. A league in Western Europe with multiple researchers and extensive broadcast coverage will have significantly more accurate player data than a league in Southeast Asia covered by a single researcher who may not be able to attend matches regularly. The FM database is often criticized for underrating players from leagues with less research coverage, and those criticisms are sometimes valid. The game’s scouting system, which mirrors the research network’s structure, can produce misleading assessments of players from under-covered regions.
Despite these limitations, the volunteer research network remains the single most impressive achievement in sports gaming. No other sports simulation game has anything comparable. EA Sports FC has licensing deals and official data partnerships, but those partnerships do not produce the granular, attribute-level detail that FM’s researchers generate. The researchers are the game. Without them, FM is a tactical engine with no players to run it.
The Compounding Architecture of Tactical Depth
FM2026’s tactical system offers many player roles per position group, multiple mentalities, and numerous tactical instructions that can be applied globally or per-unit. The combinatorial space is enormous.
When running controlled experiments with identical squads, each individual tactical adjustment produced a marginal but measurable win-rate improvement. Adding certain roles improved win rate. Adding multiple adjustments pushed it higher. The lesson here is about how tactical understanding compounds. A player who grasps why certain roles work together gains more than someone who simply applies one role in isolation.
The structural problem is that the game does not teach this. The tactic creator offers preset combinations but does not explain the spatial logic beneath them. Players who do not engage with external community resources will not discover the compounding effect.
Youth Development: The Infrastructure Dependency
FM2026’s youth development system is built on a hidden attribute that determines the maximum Current Ability a player can reach. This attribute is assigned at generation and is fixed. The game engine determines whether a player reaches their ceiling through a combination of game time, training quality, coaching staff attributes, and age.
Across tracked youth intakes over multiple seasons, the data revealed findings that challenge how most players approach youth recruitment. First, elite youth players had the lowest rate of actually reaching their ceiling. This is not a flaw in the game. It is encoding a real phenomenon: elite talent development is infrastructure-dependent. A player with high potential born into a lower-league academy with average facilities will likely plateau well below their potential. The same player at a top-tier academy with elite coaching reaches much higher.
Second, the data showed that the optimal development window runs from approximately age 16 to 21. Players who received regular first-team minutes at younger ages developed faster than players who were loaned out or kept primarily in the reserves.
The Transfer Market’s Structural Inefficiencies
FM2026’s AI transfer market uses a valuation model where a player’s asking price is a function of multiple factors. The formula creates consistent market inefficiencies that an attentive manager can exploit.
The reputation modifier is the key variable. A player at a top club carries a market valuation significantly higher than the same player at a lower-league club. The AI does not aggressively scout lower-league or non-league talent. This means that clubs in the lower tiers of the football pyramid consistently undervalue hidden gems in their squads.
The exploitable strategy is direct: specialize in scouting lower leagues, build relationships with clubs outside your league structure, and acquire talent at well below market rate.
The AI Transfer Market Exploits
If you have spent more than a few transfer windows in FM, you have likely noticed that the AI managers behave in ways that no human manager ever would. This is not a偶然 observation. The AI’s transfer logic has systematic vulnerabilities that experienced players exploit as a matter of routine. Some of these vulnerabilities are amusing. Others suggest deeper problems with how the game simulates football management.
The most reliably exploitable behavior involves the AI’s valuation of players with high potential but low current ability. The AI evaluates a young player’s transfer value primarily on current ability, reputation, and contract length. It does not adequately price in potential. This means that a 17-year-old with 150 potential and 80 current ability can often be purchased for a fraction of his actual market value if you move quickly, before his reputation inflates. The window to exploit this is usually the first 12 to 18 months after a player’s regen appears in the database. After that, the AI adjusts its valuation upward, sometimes dramatically.
A related exploit involves contract expiry. The AI is notoriously poor at managing contract negotiations for players who are not first-team regulars. A player entering the final six months of his contract becomes eligible to sign a pre-contract agreement with another club. Human players routinely monitor the contract expiry list and cherry-pick valuable players for free. The AI rarely proactively extends contracts for squad players until it is too late. This is not a偶尔 occurring bug. It is a consistent pattern across multiple AI clubs in multiple leagues. In one observed save, a top-five league club let a starting-caliber center-back walk for free because the AI failed to initiate contract renewal negotiations until the player was already in advanced talks with a rival club.
The wage structure problem is more subtle and more damaging to the game’s long-term simulation quality. AI clubs do not manage wage inflation effectively. If you sell a player to an AI club for an inflated fee, that club’s wage budget does not adjust proportionally. The AI will routinely offer wages that exceed the player’s actual market value, creating cascading wage inflation across the league. In the medium term, this produces unrealistic wage structures where mediocre players earn salaries that should be reserved for elite talents. The problem compounds over multiple seasons. By year five or six of a long-term save, wage inflation in some leagues can become so severe that the league’s financial sustainability rating degrades significantly.
Then there is the instant resignation behavior. AI managers have a propensity to resign immediately after a string of poor results, often after as few as four or five consecutive winless matches. This creates a managerial carousel that does not reflect real-world football, where managers are typically given more time—or at least are fired rather than resigning. The frequency of AI manager resignations increases as the save progresses, suggesting that the game’s morale and board confidence systems become more volatile over time. Community forums have documented cases where a mid-table club changed managers three times in a single season because each newly hired manager failed to immediately improve results and resigned in frustration.
The loan market is another area where AI logic fails. AI clubs routinely reject loan offers for players who are not getting first-team minutes, even when those players would clearly benefit from regular playing time. The AI appears to prioritize squad depth over player development in its loan decision-making, which leads to promising youngsters stagnating in the reserves of top clubs. Human players who manually manage their loan lists can develop players much faster than the AI does, creating a competitive advantage that is not based on tactical skill or recruitment acumen but on exploiting the AI’s loan logic.
Some of these exploits have been partially addressed in patches. Sports Interactive has acknowledged specific issues with AI contract renewals and wage offers. But the fundamental problem is architectural. The AI’s decision-making is based on a set of heuristics that do not fully capture the complexity of football management. The game can simulate a match with remarkable fidelity, but it cannot yet simulate the reasoning process of a human manager navigating a transfer market. Until it can, the AI transfer market will remain a collection of systematic vulnerabilities that experienced players can exploit.
The community’s response to these exploits has been mixed. Some players view them as features—opportunities to outsmart the game. Others argue that the exploits undermine the simulation’s credibility. SI has historically taken a pragmatic approach: patch the most egregious exploits, leave the rest as emergent gameplay. The result is a transfer market that feels realistic at a distance but reveals its mechanical nature the closer you look.
The Database That Scouts Use
In recent years, multiple football clubs have acknowledged using FM data as a reference point in recruitment processes. The game’s database—compiled by thousands of volunteers—is functionally cheaper and, in some assessments, more current than professional scouting subscriptions.
This is worth sitting with. The world’s most sophisticated sports simulation game has become a tool that professional football clubs use in their actual recruitment pipelines. The reason is not complicated: nobody else has coverage like it. Sports Interactive’s researcher network reaches every professional league on the planet.