Frequently Asked Questions

SoScout is an AI-powered football analytics platform that helps fans, journalists, and scouts discover players through advanced statistics. We provide detailed performance analysis, player similarity matching, and market value predictions for over 98,000 players across Europe's top leagues.

SoScout is designed for football fans who want deeper insights than basic stats, journalists researching player profiles, amateur scouts and coaches looking for talent, and anyone curious about player comparisons and performance analytics.

Our similarity engine uses a Variational Autoencoder (VAE) neural network trained on 177 performance features. It analyzes playing style, statistics, physical attributes, and positional data to find players with comparable profiles. The algorithm considers factors like goals per 90 minutes, pass completion rates, defensive actions, and many more metrics.

Yes, SoScout is currently free to use. We offer full access to player profiles, similarity search, and comparison tools. Premium features may be introduced in the future.

Our models are trained on millions of data points from professional matches. The similarity algorithm achieves high accuracy in matching playing styles, and our market value predictions are calibrated against real transfer fees. However, football is complex - our tools provide data-driven insights to complement your own judgment.

SoScout covers over 98,000 professional football players with detailed statistics. Our database includes 6.7 million individual statistics spanning from 2003 to present day.

We cover 30 top European leagues including the Premier League (England), La Liga (Spain), Serie A (Italy), Bundesliga (Germany), Ligue 1 (France), Primeira Liga (Portugal), Eredivisie (Netherlands), Belgian Pro League, Scottish Premiership, Austrian Bundesliga, Swiss Super League, and many more.

Player statistics are updated daily with the latest match data. Our automated pipeline processes new fixtures, transfers, and performance metrics every night to ensure you always have access to current information.

We source our data from SportMonks, a professional football data provider that supplies statistics to major sports platforms worldwide. This ensures accuracy and reliability of all player metrics.

Yes, we include youth players who have made appearances in the top divisions of covered leagues. This makes SoScout useful for identifying emerging talents before they become household names.

We cover players from 30 European top divisions. If a player hasn't appeared in these leagues, they may not be in our database yet. We're continuously expanding our coverage.

Simply search for any player using the search bar, open their profile, and click the 'Find Similar' button. Our AI will instantly show you players with comparable playing styles and statistics, ranked by similarity score from 0-100.

Badges indicate a player's ranking compared to others in their position. Legendary (💎) means top 1%, Elite (🔥) is top 5%, Star (⭐) is top 10%, and Key Player (🔑) is top 20%. Badges can be league-specific or worldwide, helping you understand how a player ranks both locally and globally.

Market values are predicted using an XGBoost machine learning model trained on historical transfer data and player performance metrics. The model considers age, current statistics, league level, international appearances, and recent form to estimate a player's current market value in euros.

Yes! From any player profile, click 'Compare' and search for another player. You'll see a detailed comparison with radar charts, statistics breakdown, and performance metrics for both players across all categories.

We provide 177 different metrics including goals, assists, passes, tackles, interceptions, aerial duels, dribbles, shots, key passes, crosses, and many per-90-minute normalized statistics. Players are also rated with percentile rankings compared to others in their position and league.

Per-90 stats normalize performance metrics to a 90-minute match basis. Instead of total goals, you see goals per 90 minutes played. This allows fair comparison between players with different playing time - a substitute scoring 5 goals in 500 minutes is actually outperforming a starter with 8 goals in 2000 minutes.

Percentile rankings show how a player compares to others in the same position and league. A 90th percentile in 'Goals per 90' means the player scores more than 90% of players in their position. We calculate percentiles for each statistic, making it easy to identify a player's strengths and weaknesses.

SoScout offers three time periods: Current Season (this year's performance), Last 3 Seasons (medium-term consistency), and All-Time Career (complete history). You can switch between these views to understand both current form and long-term quality.

Use the similarity search with age filters. Find a top young player you like, click 'Find Similar', then filter by age (under 21 or under 23) to discover comparable emerging talents who might be flying under the radar.

Look for players with high performance badges (Elite or Star level) but playing in smaller leagues. Our market value estimates combined with performance percentiles help identify players whose stats exceed their current valuation.

Search for Mbappé on SoScout and click 'Find Similar' to see forwards with comparable speed, goal-scoring ability, and playing style. Our AI analyzes his complete statistical profile to find the closest matches across all covered leagues.

Use the position filter in similarity search. Find a player who matches your target profile, then filter similar players by position to see only defenders, midfielders, or forwards as needed.

Absolutely. If a team loses a key player, search for that player on SoScout and find similar profiles. You can filter by market value and age to find realistic transfer targets with comparable playing styles.

A VAE is a type of neural network that learns to compress complex data into a smaller representation. For SoScout, the VAE converts 177 player statistics into a 64-dimensional 'fingerprint' that captures playing style. Players with similar fingerprints have similar playing styles, even if their raw stats differ.

After the VAE creates player embeddings, we use cosine similarity to measure how close two players are in the 64-dimensional space. The score is then adjusted for position compatibility and league level, giving a final 0-100 score where 100 means identical profiles.

We use a PyTorch Variational Autoencoder for similarity matching and an XGBoost gradient boosting model for market value prediction. Both models are retrained monthly to incorporate the latest player data and transfer market trends.