Player stats markets are the fastest-growing segment of football betting — and among the most persistently mispriced. Unlike match result or Over/Under goals markets, player markets are priced primarily from historical league averages and broad team-level data, with insufficient adjustment for individual player quality, positional role, and fixture-specific factors. The result is a market ecosystem where elite shot-volume strikers like Haaland and Mbappé are consistently underpriced in shots on target markets, top penalty takers like Lewandowski are underpriced in anytime scorer markets when penalties are expected, and defensive midfielders with high individual card rates like Casemiro are underpriced in player card markets against strict referees. The edge in player stats betting is found by knowing individual player rates — not team rates — and matching them against the specific fixture context.

In this guide
  1. Anytime scorer — how to model individual scoring probability
  2. Shots on target — the most persistently mispriced player market
  3. Assists — the hardest player market to model accurately
  4. Player cards — how individual rates and referee data combine
  5. Settlement rules — what counts and what doesn't
  6. Frequently asked questions

Anytime Scorer — How to Model Individual Scoring Probability

Anytime scorer probability is driven by five factors in descending order of importance: the player's xG per 90 minutes in the current season (the most predictive individual metric); whether the player is the primary penalty taker for their team (penalty takers receive an approximately 7–10% uplift in scorer probability relative to their open-play rate); the match's expected goal total (a higher-scoring match benefits all players' individual scoring probabilities); the opponent's defensive quality (goals conceded per game, clean sheet rate, and specific vulnerability to the player's position and style); and the player's H2H scoring record against the specific opponent.

The most important distinction in anytime scorer modelling is between non-penalty xG and total xG. Lewandowski's total xG of 0.78 per 90 includes penalty kicks — his non-penalty xG is approximately 0.58. The difference is significant for matches where penalties are unlikely. Always identify whether the player is a penalty taker and adjust the xG model to include or exclude penalty probability based on the specific fixture's expected penalty rate (approximately 0.25 penalties per game on average, with wide variance by fixture type).

Shots on Target — The Most Persistently Mispriced Player Market

Shots on target is the player stats market with the largest and most consistent model edges across all European football. The reason is structural: bookmakers price SoT markets from team-level shooting data averaged across all players, then apply a multiplier for the named player's share of total team shots. This methodology massively underestimates elite shot-volume strikers whose individual SoT rates (2.0–3.5 per 90) are 3–5× the team-average share that the bookmaker's model assigns them.

Today's Mbappé Over 2.5 SoT at 2.50 (implied 40%, model 64%, edge +24.0pt) is the most extreme example: Mbappé's 3.1 SoT per 90 means Over 2.5 SoT is his structural base rate — yet the bookmaker prices it at 2.50 as if it were a below-50% probability outcome. Haaland's Over 1.5 SoT at 2.10 (implied 47.6%, model 68%, edge +20.4pt) shows the same systematic pattern. These are not isolated cases — elite strikers' SoT markets are persistently underpriced across all seasons and all bookmakers because the pricing methodology fails to incorporate individual volume data properly.

Assists — The Hardest Player Market to Model Accurately

Assists are the most difficult player market to model because they depend on both the assist provider performing their role (crossing, through-ball, set piece delivery) and a teammate converting the resulting chance. The model correctly identifies high-xA players (Trent Alexander-Arnold at 0.45 xA per 90, Bellingham at 0.38 xA per 90) — but converting expected assists into actual assists requires a teammate to score, adding a conversion variance layer that is absent from pure player performance markets.

The practical implication: assist markets should only be backed at odds of 3.00+ where the model edge is at least 5 percentage points — the higher baseline variance of the assist market requires a larger edge to produce positive expected value at the same confidence level as goal or SoT markets. Today's assist selections (Bellingham at 3.40, TAA at 3.20) both land above the 3.00 threshold with genuine model edges of +2.6pt and +3.7pt respectively.

Player Cards — How Individual Rates and Referee Data Combine

Player card markets require two inputs: the individual player's card rate per start, and the assigned referee's booking rate per game. Neither input alone is sufficient — a player with a 40% card rate in league games will have a materially different card probability under a 6.0 cards/game referee versus a 2.8 cards/game referee. The correct player card probability is calculated as: individual foul rate × (referee's booking probability per foul in this competition).

Today's Casemiro selection illustrates the methodology: his 40% season card rate combined with a 3.8 cards/game referee (above the 3.4 Premier League average) pushes his card probability to 34% — above the 30.8% implied by 3.25 odds. The Kanté selection uses the same framework in La Liga: 37.5% base card rate, 5.1 cards/game referee (the strictest in La Liga), producing a 36% model probability against the 33.3% implied by 3.00 odds.

Settlement Rules — What Counts and What Doesn't

All player stats markets settle on events in 90 minutes plus stoppage time. Performances in extra time or penalty shootouts do not count. For anytime scorer markets: penalty goals count as goals for the player; own goals do not count as goals for any player; if a player does not start the match, most bookmakers void the bet (some settle at no-goal if the player doesn't play at all — always verify the bookmaker's specific non-starter rules before placing the bet). For shots on target markets: the definition of a shot on target varies slightly between bookmakers — most use the official match data provider's definition (on-target shots that would have gone in if not saved or that hit the post), but some include saved crosses. Always check the bookmaker's specific shots on target definition.

Frequently Asked Questions

If the player starts the match but is substituted off before the final whistle, most bookmakers settle the anytime scorer bet based on whether the player scored before coming off — if they scored, the bet wins; if they didn't score and were substituted, the bet loses. The bet is not voided simply because the player didn't complete the 90 minutes. Only if the player fails to take any part in the match (does not start and does not come on as a substitute) will most bookmakers void the bet. Some bookmakers void if the player is substituted before a certain minute — always check the specific bookmaker's anytime scorer settlement rules.
Yes — penalty goals count as anytime scorer goals at all major bookmakers. A player who scores only from the penalty spot wins the anytime scorer market. This is why identifying primary penalty takers is an important component of anytime scorer modelling — players like Lewandowski and Osimhen, who are their team's first-choice penalty takers, receive a meaningful probability uplift in fixtures where penalties are expected. The penalty taker advantage is most significant in fixtures where the attacking team is a heavy favourite and expected to earn multiple penalty-area challenges.
xG (Expected Goals) is a statistical measure of shot quality — it assigns a probability between 0 and 1 to each shot based on its location, the preceding action, the body part used, and other factors. A shot with xG of 0.4 is expected to result in a goal 40% of the time based on historical data from similar shots. xG per 90 minutes measures a player's average shot quality across a full game. A player with xG of 0.70 per 90 is expected to score 0.70 goals per 90 minutes from open play — meaning in a 90-minute match, their expected goals involvement is 0.70. For anytime scorer modelling, xG per 90 is the most predictive single metric — significantly more accurate than raw goal totals, which are subject to finishing luck variance.
Shots on target markets are underpriced for elite volume shooters because bookmakers price them from team-level shot data rather than individual player rates. When the team averages 6 shots on target per game and there are 11 players contributing, the naive per-player share is less than 1 shot on target per player — leading to conservative pricing on individual player SoT markets. In reality, shot volume is highly concentrated among a small number of players: Haaland, Mbappé, and Lewandowski each register 2.4–3.1 shots on target per 90 — 3–5× their naive expected share of the team total. Until bookmakers fully incorporate individual player SoT rates into their pricing models (rather than distributing team totals), this structural mispricing will persist.