By David Butler
There was bad news recently for researchers interested in football (soccer) data. FBref.com announced that, as of January 20, its data agreement with Stats Perform would be terminated. Further details were reported in the New York Times. This development represents a significant setback for the analytics, Sports Econ and Sports Management community interested in football, particularly for those working on the women’s game. The platform represented an unprecedented free resource for researchers that has now lost it’s teeth.
Over the past seven years, FBref.com democratised access to advanced performance data. Lone scholars, students, and researchers without institutional budgets were able to access detailed performance metrics that would otherwise sit behind very expensive paywalls (which media companies often pay for). For women’s football in particular, the past two years marked a step change in data availability and research potential. That progress has now been interrupted.
Historically, FBref’s advanced statistics were first powered by StatsBomb and Data Sports Group (DSG). Within the analytics community, StatsBomb’s expected goals (xG) and expected assists (xA) models were often regarded as higher quality than those produced by Stats Perform (Opta), largely due to method. Under the StatsBomb/DSG agreement, FBref provided unusually granular data, including information on pressing actions and their locations, passing by body part and height, carry distances, carries into the final third, and even nutmegs!
When FBref later switched from StatsBomb/DSG to Stats Perform (Opta), many of these variables disappeared. Data quality was traded off with quantity, and coverage expanded in other dimensions: more leagues were included, xG data became more widely available, and, most notably, women’s football data increased substantially.
Perhaps more worryingly, the recent FBref announcement is not an isolated case. Earlier this season, at the start of the 2025/26 campaign, FotMob briefly published physical performance metrics such as sprint counts and distances covered by players. These features were switfly removed. Together, these episodes may suggest a broader tightening of control over performance data distributed through online/public platforms.
Why is this happending? One possible explanation lies in the commercial value of this football data, particularly in gambling markets. Betting firms have historically played a significant role in funding data collection and dissemination, sometimes enabling free availability. Perhaps their incentives may be shifting toward restriction rather than openness. This interpretation is of course speculative.
For researchers the implications are concerning. The erosion of freely accessible data is a step backwards. Women’s football, which has only recently begun to benefit from improved data coverage, may be especially vulnerable. The “good times” of open access football analytics are not fully over as ‘basic statistics’ are still freely available in lots of places, and advanced stats on xG and xA are available elsewhere, but the free data movement is clearly under pressure.
There was bad news recently for researchers interested in football (soccer) data. FBref.com announced that, as of January 20, its data agreement with Stats Perform would be terminated. Further details were reported in the New York Times. This development represents a significant setback for the analytics, Sports Econ and Sports Management community interested in football, particularly for those working on the women’s game. The platform represented an unprecedented free resource for researchers that has now lost it’s teeth.
Over the past seven years, FBref.com democratised access to advanced performance data. Lone scholars, students, and researchers without institutional budgets were able to access detailed performance metrics that would otherwise sit behind very expensive paywalls (which media companies often pay for). For women’s football in particular, the past two years marked a step change in data availability and research potential. That progress has now been interrupted.
Historically, FBref’s advanced statistics were first powered by StatsBomb and Data Sports Group (DSG). Within the analytics community, StatsBomb’s expected goals (xG) and expected assists (xA) models were often regarded as higher quality than those produced by Stats Perform (Opta), largely due to method. Under the StatsBomb/DSG agreement, FBref provided unusually granular data, including information on pressing actions and their locations, passing by body part and height, carry distances, carries into the final third, and even nutmegs!
When FBref later switched from StatsBomb/DSG to Stats Perform (Opta), many of these variables disappeared. Data quality was traded off with quantity, and coverage expanded in other dimensions: more leagues were included, xG data became more widely available, and, most notably, women’s football data increased substantially.
Perhaps more worryingly, the recent FBref announcement is not an isolated case. Earlier this season, at the start of the 2025/26 campaign, FotMob briefly published physical performance metrics such as sprint counts and distances covered by players. These features were switfly removed. Together, these episodes may suggest a broader tightening of control over performance data distributed through online/public platforms.
Why is this happending? One possible explanation lies in the commercial value of this football data, particularly in gambling markets. Betting firms have historically played a significant role in funding data collection and dissemination, sometimes enabling free availability. Perhaps their incentives may be shifting toward restriction rather than openness. This interpretation is of course speculative.
For researchers the implications are concerning. The erosion of freely accessible data is a step backwards. Women’s football, which has only recently begun to benefit from improved data coverage, may be especially vulnerable. The “good times” of open access football analytics are not fully over as ‘basic statistics’ are still freely available in lots of places, and advanced stats on xG and xA are available elsewhere, but the free data movement is clearly under pressure.
RSS Feed
