Per90
Loading live football data…
English Premier League · England
expected lineupBest edge
+14.3%
Marcus Tavernier · shots
Allowed a game
22.8
Manchester City · Concede tackles
Referee
-9%
Jarred Gillett · cards v league
7 live calls — tap one for the price it beat
Manchester City
AFC Bournemouth
expected lineup The expected XI — it changes until the teamsheet is handed in. Manchester City attack left to right. No player prices are standing on this match — the books take them down at kick-off — so the pitch shows the teamsheet alone.
The gap — the model’s fair price is 3.505/2 and bet365 is offering 4.003/1. That distance is the +14.3% edge, and it is what was published — not a price fetched now. Confidence low.
The gap — the model’s fair price is 3.8014/5 and bet365 is offering 4.3310/3. That distance is the +14.0% edge, and it is what was published — not a price fetched now. Confidence low.
The gap — the model’s fair price is 3.8014/5 and Paddy Power is offering 4.3310/3. That distance is the +14.0% edge, and it is what was published — not a price fetched now. Confidence low.
The gap — the model’s fair price is 2.206/5 and Paddy Power is offering 2.503/2. That distance is the +13.6% edge, and it is what was published — not a price fetched now. Confidence low.
Manchester City
22.8
Manchester City
2.9
Jarred Gillett
19.9
Manchester City
9.0
An expected XI is out. Minutes mix whether a player is flagged to start with his own minutes as a starter and as a substitute.
6 withdrawn, still shown.
| Player | Market | Best odds | Fair odds | Edge | Bookmaker | Status |
|---|---|---|---|---|---|---|
| Alex Scott | Over 3.5 tackles | 3.505/2 | 2.6213/8 | +33.6% | Sky Bet | withdrawnlossCLV 0.0% |
| James Hill | Over 1.5 tackles | 1.574/7 | 1.331/3 | +18.1% | bet365 | withdrawnlossCLV 0.0% |
| Marcus Tavernier | Over 2.5 shots | 4.003/1 | 3.505/2 | +14.3% | bet365 | lossCLV +17.6% |
| Alex Scott | Over 2.5 fouls | 4.3310/3 | 3.8014/5 | +14.0% | bet365 | lossCLV 0.0% |
| Erling Haaland | Over 1.5 fouls | 4.3310/3 | 3.8014/5 | +14.0% | Paddy Power | lossCLV +3.2% |
| Adrien Truffert | Over 0.5 shots | 2.503/2 | 2.206/5 | +13.6% | Paddy Power | lossCLV 0.0% |
| Antoine Semenyo | Over 1.5 shots on target | 3.259/4 | 2.8815/8 | +12.8% | bet365 | lossCLV 0.0% |
| Nico O'Reilly | Over 2.5 tackles | 2.503/2 | 2.255/4 | +11.1% | bet365 | withdrawnlossCLV +13.6% |
| Marcus Tavernier | Over 2.5 tackles | 3.002/1 | 2.707/4 | +11.1% | bet365 | withdrawnlossCLV 0.0% |
| Evanilson | Over 2.5 shots | 3.505/2 | 3.2011/5 | +9.4% | bet365 | lossCLV 0.0% |
| Erling Haaland | Over 5.5 shots | 3.002/1 | 2.757/4 | +9.1% | bet365 | lossCLV 0.0% |
| Adrien Truffert | Over 2.5 tackles | 1.538/15 | 1.422/5 | +7.8% | bet365 | withdrawnlossCLV +2.2% |
| Matheus Nunes | Over 0.5 tackles | 1.131/8 | 1.061/16 | +6.1% | bet365 | withdrawnvoidCLV +1.3% |
A withdrawn call is one Per90 no longer stands behind — the rule that published it was later found wrong. Hover it for the reason. Withdrawn calls are not deleted — each keeps its published price and timestamp, is still scored against its closing line, and is still counted in every figure on the record. Why a withdrawal is not a deletion → Published at the price shown and scored against the closing line whatever happens next — the same rows, with staking context, are on the value board and in the track record .
Last six completed matches per side, newest first. Half-time score in brackets where the feed carries one.
No completed matches on record for this club.
No completed matches on record for this club.
Model projections. Nothing here has been compared to a bookmaker.
How this is calculated →No projections for this fixture yet. The projection job runs against fixtures inside the odds window, and only for players with at least 10 completed matches on record.
Form against the lines the books are pricing, what the opposition concede, and the referee against his league.
Facts with sample sizes, not predictions. The market % beside a line is the books’ median price with margin in, and the over market runs hot.
every priced line§5.1: the dominant covariate for fouls, and the reason two identical players price differently on different days.
46 matches on record · full record →
Fouls per game
20.8
-10% vs leagueleague 23.2 fouls
Cards per game
3.8
-9% vs leagueleague 4.1 cards
Cards, home side
1.8
Manchester City
Cards, away side
2.0
AFC Bournemouth
Fouls — every foul committed by both teams in a match he refereed, averaged over his completed matches. The match being previewed is never in it.
Cards — any card, yellow or red, both teams. The same definition the card props settle against, so this figure and the Cards market count the same events. A dismissal that the feed records as a second yellow counts as two.
Both league averages are the mean across every competition Per90 covers, not this fixture's league alone — the fouls comparison has always been that, and the cards comparison is built the same way so the two percentages beside each other mean the same thing.
What the numbers on this page are, and what they are not.
Not tips and not prices. Each row is the model's expected count for one player in this match — how many fouls, tackles or shots it thinks he records — plus the full spread of outcomes around it. Nothing in the projections has been compared to a bookmaker; that comparison is what the Value tab is.
Step 1
His own record
Every completed match in the dataset, as a rate per 90 minutes. The n column is how many matches that rests on.
Step 2
Pulled toward his position
A thin record gets pulled hard toward other players in his position, in his league — a goalkeeper toward goalkeepers, not toward the outfielders around him. How hard is fitted per market, not chosen: tackles pull about four times harder than shots on target.
Step 3
This fixture's conditions
How much of the stat the opponent concedes, what the referee’s matches look like, and home or away. This is where a projection stops being a season average.
Step 4
Scaled to expected minutes
A rate becomes a count only once you say how long he plays. 60 minutes carries two-thirds the exposure of 90, and the distribution is built at that exposure.
How to read each column
Counts are modelled as negative binomial rather than Poisson, because fouls and tackles are overdispersed and a Poisson would systematically underprice the tails — which is exactly where over bets live. Every market on this page cleared its calibration gate before it was allowed to produce a projection: when the model says 60%, it lands within a point or so of 60% out of sample.
The full write-up — every gate, every fitted parameter and what the model refuses to price — is on the methodology page.
