How Bookmakers Set Player Prop Lines (and Where the Averages Lie): 286,990 Player-Matches Analysed

A forward has averaged 1.5 shots over his last 10 games, and the line is Over 1.5. It looks like a coin flip, so anything above evens feels like value. It isn't: players with that profile went over 38.4% of the time. Here's how prop lines are built, and what 286,990 player-matches say about where averages mislead.

Data: StatPair match database · 20 European leagues · 2024/25, 2025/26 and 2026/27 to date · Updated: 13 October 2026

TL;DR

• An average isn't a 50% line. Players averaging 1.5 shots over 10 appearances went Over 1.5 in 38.4% of next matches (20,637 cases). Tackles: 38.5%. Fouls committed: 33.7%.
• Why: counts are skewed (Poisson already says 43.9%), hot averages fall back (1.49 shots became 1.37), and minutes vary.
• Minutes matter most. That 1.5-shot group hit 46.4% as starters, 51.0% in full 90s and 18.4% off the bench.
• Regular starters still slip. Players with 10 starts in 10 played under 60 minutes in 11.7% of next appearances.
• Per 90, most props are Poisson-like. Shots on target and fouls committed have a dispersion index of 1.00 in full 90s; tackles (1.16) and saves (1.25) are more volatile.
• "8 of the last 10" isn't 80%. Fouls Over 0.5 hit 65.2% after an 8/10 run.

How bookmakers build a prop line

Bookmakers don't publish their models, so this describes the general industry approach, not any one operator's method.

1. Project the player's mean. A rate such as shots per 90 blends recent form with longer-term numbers, adjusted for opponent, venue and team strength.

2. Multiply by expected minutes. A player expected to play 75 minutes gets about five-sixths of his per-90 rate. Start uncertainty is the hardest part to price before team sheets.

3. Turn the mean into a distribution. The model needs the probability of 0, 1, 2 or more, often from a Poisson distribution with adjustments. Lines usually sit at 0.5, 1.5 or 2.5 so there are no pushes, and the price does most of the work.

4. Add the margin. One public study, The Price of a Bet, recorded prices from 52 bookmakers in June-August 2026 and measured an average 7.93% margin on two-sided player props, against 4.77% on two-way match markets and 7.67% on three-way match betting. Its sample covers several sports, not only football. A June 2026 comparison put the average Premier League 1X2 margin at 5.4% across nine bookmakers.

Same-game multis add correlation. Legs on the same player aren't independent: a player with 2+ shots also had a shot on target in 64.9% of cases, against a 23.1% base rate. Bet builders price that link in. See our bet builder hit-rate guide.

How we measured it

We took every player appearance in 20 European leagues (listed in the method notes) and computed the player's average over his previous 10 appearances, keeping only appearances with 10 earlier ones on record. That leaves 286,990 player-matches from 9,643 players in 11,495 matches (268,857 outfield, 18,133 goalkeeper).

We then checked how often the player went over 0.5, 1.5 and 2.5 in the next match, against what a Poisson distribution predicts from his 10-game average. Start data is complete from 2025/26, so the minutes analysis uses 152,574 player-matches from 2025/26 and 2026/27.

The average is not the line

Here is what happened next for outfield players grouped by their 10-game average. Each group covers an average within ±0.1 of the value shown.

Three separate effects push the hit rate below 50%.

Skew. Counts can't go below zero but can run to five or six, so the median sits below the mean. Even a perfect Poisson model with a mean of 1.5 gives Over 1.5 only about 44%.

Regression to the mean. A high 10-game average is partly luck. Players averaging 1.49 shots produced 1.37 next time, and those averaging 2.49 produced 2.12. Shots on target fell further: 0.98 became 0.75.

Extra zeros. Within each group, the variance was mostly 1.15 to 1.40 times the mean (Poisson gives 1.00), and it shows up as blanks. Players averaging 1.0 shot blanked in 43.6% of next matches, where Poisson at their real next-match average predicts 37.9%.

For the 1.5-shot group, skew takes 50% down to 43.9%, the lower real average to 39.8%, and extra variance to 38.4%. Low averages regress too, upwards: when Poisson gave a player under 10% to commit a foul, he did so 31.9% of the time.

Minutes: the biggest risk in any prop

A prop on a player averaging 1.5 shots is really two bets: will he play a full role, and will he shoot? Here are next-match hit rates split by role (2025/26 onwards):

(Player-matches in brackets.)

A 1.5-shot average is close to a coin flip only when the player plays the full 90. A substitute appearance cuts the hit rate by more than half, and at many bookmakers the bet still stands.

How reliable are "regular starters"? We grouped players by starts in their last 10 appearances:

A player who started all of his last 10 appearances still played under an hour in 11.7% of next appearances, and only 71.3% of his starts went the full 90. For an "8 or 9 out of 10" starter, it was a quarter. Substitutes averaged 17 to 25 minutes. (Matches where the player didn't get on aren't in our data; those bets are usually voided.)

Which stats behave like Poisson?

To separate the stat from minutes, we measured each player's own variance-to-mean ratio (dispersion index) for players with at least 15 matches. 1.00 is pure Poisson; above it means more blanks and more big games.

(Full 90: 3,416-3,552 outfield players and 551 goalkeepers. All appearances: up to 8,244 players.)

• Shots on target and fouls are near-textbook Poisson once minutes are fixed: a per-90 rate plus expected minutes prices them well. See our shots on target analysis and fouls guide.
• Shots are slightly over-dispersed, and more so once minutes vary (1.26).
• Tackles are the most volatile outfield stat, because they depend on how much the opponent has the ball.
• Goalkeeper saves follow the opponent, not the keeper. Keepers averaging 2.0 saves went Over 1.5 in 71.1% of next matches (1,580 cases); Poisson from their average said 59.8%. Price saves from the opponent's shots on target.
• Yellow cards are under-dispersed: a second yellow means a red, so Poisson overstates the chance of two.

For which stats repeat best from game to game, see our most predictable football stats study.

"He's hit it in 8 of his last 10"

Bet builder apps show recent hit rates. Here's how those runs held up next time:

Recent hit rates carry information but overstate the true chance. An "8 out of 10" fouls run was really a 65% shot. Price from the raw hit rate and short prices will look like value when they aren't.

From hit rate to fair odds

Fair decimal odds are 1 ÷ probability. The margins below are illustrative (6% and 8%, spread evenly). Real margins vary, and a companion study suggests bookmakers load more of it onto the less likely side, especially in props.

Say the over is 2.20. On the 50/50 story it looks like a 10% edge. At the real 38.4%, expected value is 0.384 × 2.20 − 1 = −15.5%. With the player confirmed in the XI, it's 0.464 × 2.20 − 1 = +2.1%. The team sheet can turn a bad bet into a fair one.

To do it yourself:

1. Remove the margin: fair probability of the over = (1 ÷ over odds) ÷ (1 ÷ over odds + 1 ÷ under odds).
2. Estimate your own probability from per-90 rates, expected minutes and the opponent, not the raw 10-game average.
3. Expected value = your probability × decimal odds − 1. Our expected value calculator does the maths, and our odds explained guide covers conversions.

What this means for betting

A checklist for spotting a soft prop line

• Minutes first. Confirmed starter? Recently subbed around the hour? A line built on a full-90 rate is soft against a likely early sub, and vice versa.
• Opponent. Shots depend on what the other team concedes, tackles on how much it has the ball, saves almost entirely on its attack.
• Referee, for fouls and cards. Foul counts vary a lot by official. See our referee impact analysis.
• Home or away. Many players produce more at home. Our home and away study shows how much.
• Role change. A new position, manager or a teammate's injury changes a rate faster than a 10-game average can follow. That's when lines lag.
• Regression. Hot streaks cool off. Be wary of overs after a run well above the long-run average.

Bet builders: correlated legs

For outfield players, 2+ shots and a shot on target landed together 14.4% of the time, against 5.1% if independent (2.81 times as often). 2+ fouls and a yellow card: 6.5% against 2.9% (2.19 times). A shot and a foul won: 27.6% against 23.9% (1.15 times), falling to 1.06 in full 90-minute games, so that link is mostly minutes. Linked legs pay less than the product of their single prices, and that's fair.

Method notes

• Leagues: Premier League, La Liga, Serie A, Bundesliga, Ligue 1, Championship, Segunda División, Serie B, Eredivisie, Primeira Liga, Belgian First Division A, Scottish Premiership, Süper Lig, Greek Super League 1, Danish Superliga, Liga I, Ekstraklasa, HNL, Czech Liga and Ligat Ha'al. League matches only, July 2024 to 11 October 2026.
• Rolling averages use the previous 10 league appearances in these leagues, so cup and European games aren't counted, and a run can span two seasons or clubs.
• Appearances only. Matches where a player didn't get on aren't in our data, so figures describe the next appearance.
• The Poisson comparison uses the raw 10-game average as the mean. Bookmakers use richer inputs, so this shows where averages mislead, not how accurate bookmakers are.
• Dispersion is measured within each player across his whole sample, so it includes some real changes in form.
• Margins come from the cited public sources. We don't hold bookmaker odds, so we can't test whether real prop lines were mispriced.

How to check this with StatPair

1. Open the player's match-by-match history and check minutes and starts, not just the average.
2. Use the home/away and half filters to see if output depends on venue or fades late.
3. Use the opponent filter to see what that team usually allows, and referee stats for fouls and cards.
4. Use the Trends Finder to find players clearing a line, then discount the run: an 8/10 streak was worth 57-65% next time in our data.
5. Ask the AI Analyzer for a matchup summary, and save players with changing roles to your Notebook.

Try StatPair free for 7 days → · Bet builder hit rates → · Most predictable stats →

FAQ

Q: How do bookmakers set player prop lines?
A: In general, they project the player's average for the stat from recent and longer-term rates, scale it by expected minutes, adjust for the opponent, turn that into a probability distribution and add a margin. Lines usually sit at 0.5, 1.5 or 2.5, and the odds do most of the work. One 2026 study of 52 bookmakers measured an average 7.93% margin on two-sided player props.

Q: If a player averages 1.5 shots, is Over 1.5 shots a 50/50 bet?
A: No. In our data, players averaging 1.5 shots over their last 10 appearances went over 1.5 in 38.4% of next matches (20,637 cases). The gap comes from skew, regression to the mean and minutes. It reached 51.0% only when the player played the full 90.

Q: What happens to player props when a player comes off the bench?
A: Hit rates collapse. Players averaging 1.5 shots hit Over 1.5 in 18.4% of substitute appearances, against 46.4% when they started. At many bookmakers the bet stands once the player comes on, so check line-ups and settlement rules.

Q: Is the Poisson distribution good for player props?
A: For shots on target and fouls, it fits very well once minutes are fixed (dispersion index 1.00 to 1.02 in full 90-minute games). It's less accurate for tackles (1.16) and goalkeeper saves (1.25), and it overstates the chance of two yellow cards. Most of the error comes from minutes and from using a raw 10-game average as the mean.

Q: How do I calculate fair odds for a player prop?
A: Fair decimal odds are 1 divided by the probability. A 38.4% chance is fair at 2.60. To remove the margin from a two-way price, divide each side's implied probability (1 ÷ odds) by the sum of both. Expected value is your probability times the decimal odds, minus 1.

Bottom line

An average is not a line. Count stats are skewed, hot streaks cool off and minutes swing results more than anything else. Players averaging 1.5 shots hit Over 1.5 38.4% of the time, and about half the time only when they played 90. Price props from per-90 rates and expected minutes, discount recent hit rates, and wait for the team sheet when the start is uncertain.