What a line is
“Over 1.5 fouls” wins if the player commits two or more; “Over 0.5 shots on target” wins on one. The half means there is no push. The first thing to know about any line is what is being counted. A shot blocked by a defender is not on target; a second yellow that becomes a red is two cards to one provider and something else to another; a foul that also brings a card is still one foul. Per90’s glossary states what its data feed records for each stat — which is not always what a given bookmaker settles on. When the two differ, the bookmaker’s rules win your bet, and the research is only as good as its match to those rules.
Hit rate, average and per 90
Three numbers describe a player’s record against a line, and they answer different questions.
- Hit rate is the share of appearances in which he cleared the exact line — over his last 5, 10 or 20. “62% over 1.5 fouls” means two or more fouls in 62% of the matches counted. It is a count of what happened. It is not a projection and it is not a probability for the next match.
- Average is the mean count. It hides the shape: a player averaging 1.5 shots might take three every other game and none in between, and the line does not care about the average.
- Per 90 scales the total to ninety minutes of playing time. On this site it is minutes-weighted over the last 20 appearances, so a substitute’s twenty-minute cameo cannot inflate it. It is the fairest way to compare two players, and the worst way to predict one who might play thirty minutes.
Use the longer windows. A 5-match hit rate is five coin flips; a 20-match rate says something, and the Props Finder shows all three so a strong recent run can be read against the longer record rather than instead of it.
Minutes are the first risk
A prop needs the player on the pitch. Every count above assumes he plays, and the single biggest reason a well-researched prop loses is that he did not — benched, withdrawn at half-time, sent off, or on for the last quarter of an hour. Before anything else, check the teamsheet state. The model’s expected minutes come in three grades: a confirmed lineup, an expected lineup, or a guess from his historical start rate, and the site labels which one it is using. A line that looks generous on a player whose start is a guess is priced for that uncertainty.
What moves the number
A count is produced by a matchup, not by a player alone. Four things move it, and the research tools each isolate one.
- The opponent, by role. A right winger’s shots depend on what the opposing left-back allows to right wingers, not on the opponent’s defence in general. Position stats shows what a team produces from a role and allows to it.
- The referee. Fouls and cards are the referee’s count as much as the player’s. An official who gives a foul a game more than his league average moves every foul line in the match. Referee records show fouls and cards per game against the league.
- The direct opponent. A full-back facing a winger who draws fouls is a different fouls bet from the same full-back against a winger who cuts inside. Player duels works out who faces whom from the teamsheet’s shape.
- Game state and venue. Favourites shoot more; sides chasing a game foul more; most players’ home and away records differ. The stat pages carry the venue split so it is read, not assumed.
From a count to a fair price
A price implies a probability once the bookmaker’s margin is removed — the de-vig. The question a prop bettor is really asking is whether the chance of clearing the line is higher than the price implies. Per90 answers it two ways, and keeps them apart.
The projection model fits a negative binomial distribution to each market, pulls each player’s rate toward his position and league average by an amount that depends on his sample (shrinkage), adjusts for the referee, and scales to expected minutes. That gives a full distribution of the count, and therefore a probability of clearing any line. The methodology page shows its fit and calibration for every market it prices, including the ones it does not price well.
The value board uses a different reference: the sharp market’s own price, de-vigged, as the fair line, and shows UK prices standing above it. That is a comparison of prices, not a model, and every pick is graded afterwards against the closing price and the result on a public record. An edge against the fair price is not a profit in hand; it is the condition under which a bet is worth having, over a long run, at the right stake, and no outcome is guaranteed.
A working order
- Choose the market and the exact line the bookmaker is offering — not a nearby one.
- Read the 10- and 20-appearance hit rates at that line, then the per-90 rate.
- Check the teamsheet grade and expected minutes.
- Check what the opponent allows to the player’s role, and the referee’s fouls and cards if the line is fouls or cards.
- Compare the price across books, and against the fair price where one is shown.
- Log the bet and, afterwards, the closing price. Closing line value tells you whether you were beating the market long before the results can.
What this does not do
A high hit rate is a description of the past. It makes no claim that a line is mispriced, and the Props Finder makes none either — it filters evidence. The model prices some markets much better than others and says which. And no amount of research moves a player onto the pitch. Research on published prices and public statistics, not advice.
Open the Props Finder → to filter every priced prop, or start from a stat leaderboard.