Value bet index yesterday
Prices that beat the model's fair odds by at least 6%, ranked by edge. Expect a lower hit rate than the main picks and bigger returns when they win – a strategy measured over hundreds of bets.
NigeriaNPFLxG1X2O2.5BTTSPick · EV
BrazilSerie AxG1X2O2.5BTTSPick · EV
SpainCopa Del ReyxG1X2O2.5BTTSPick · EV
WorldFriendliesxG1X2O2.5BTTSPick · EV
ColombiaCopa ColombiaxG1X2O2.5BTTSPick · EV
North & Central AmericaCONCACAF Nations LeaguexG1X2O2.5BTTSPick · EV
WorldASEAN Club ChampionshipxG1X2O2.5BTTSPick · EV
AlbaniaCupxG1X2O2.5BTTSPick · EV
ZambiaSuper LeaguexG1X2O2.5BTTSPick · EV
EnglandProfessional Development LeaguexG1X2O2.5BTTSPick · EV
UgandaPremier LeaguexG1X2O2.5BTTSPick · EV
EgyptCupxG1X2O2.5BTTSPick · EV
SpainTercera División RFEF Group 11xG1X2O2.5BTTSPick · EV
SloveniaCupxG1X2O2.5BTTSPick · EV
QatarSecond DivisionxG1X2O2.5BTTSPick · EV
AlgeriaLigue 1xG1X2O2.5BTTSPick · EV
SpainTercera División RFEF Group 13xG1X2O2.5BTTSPick · EV
SpainTercera División RFEF Group 8xG1X2O2.5BTTSPick · EV
SerbiaCupxG1X2O2.5BTTSPick · EV
EuropeUEFA Champions League WomenxG1X2O2.5BTTSPick · EV
ArgentinaReserve LeaguexG1X2O2.5BTTSPick · EV
SpainTercera División RFEF Group 6xG1X2O2.5BTTSPick · EV
EnglandWSL CupxG1X2O2.5BTTSPick · EV
BrazilCopa Do Brasil U20xG1X2O2.5BTTSPick · EV
PeruPrimera DivisiónxG1X2O2.5BTTSPick · EV
Yesterday's value index
44 prices sit above the model's fair odds, averaging 2.32. Largest edge: Iliria Fushë-Krujë vs Luftëtari – Both teams to score at 2.39, rated 57% by the model against the 42% the price implies.
At odds near 2.32, breaking even needs roughly 43% winners. Last 30 days: 1,421/3,254 correct (44%). Value shows up across hundreds of bets, not in a single day.
Betting the value index
- Small level stakes – 1–2% of your bankroll.
- Back them as singles.
- If the price has fallen below fair odds, the edge is gone.
- Expect losing runs; many value picks are under 50%.
Where the edge comes from
Odds imply a probability: 2.50 implies 40%. If our model puts the outcome at 46%, the expected value is 0.46 × 2.50 − 1 = +15%. One such bet can easily lose; a long series of them is how an edge turns into profit.
Value picks are listed separately from the main predictions. Their profit record is on the leaderboard, and price moves on odds movement.
The value bet index: where price and probability disagree
A value bet is a price that pays more than the outcome's probability deserves. If the model gives a team a 50% chance, fair odds are 2.00; any price above 2.00 is value on our numbers. This page ranks Wednesday 18 November's fixtures by that gap, largest first, so you can see where the market and the model disagree most.
Expected value in one line
Expected value is probability × odds − 1. At 50% and odds of 2.30, EV = 0.5 × 2.3 − 1 = +15%. That means a stake of 100 is worth 115 on average over a long series of identical bets. It says nothing about the next single bet, which will either win 130 or lose 100. Value betting is a long-run strategy; its rewards appear over hundreds of bets, not over a weekend.
Why the model and the market disagree
Bookmakers set prices from their own models, their liabilities and the money they receive. Our probability blends independent team ratings with an xG feed and the market itself. When they diverge, one of a few things is usually happening:
- The ratings have picked up a change in form faster than the market has.
- A price has not been updated after a line-up announcement that our blend already reflects through other bookmakers.
- The market is shading a popular team's price to balance its book.
- The model is missing information – an injury, a rotation, a motivation issue – and the market is right.
That last case is why the biggest edges deserve the most scrutiny, not the biggest stakes.
Reading the index
Each row lists the selection, the model probability, our fair odds, the best available price and the edge. Edges between 3% and 12% are typical of genuine value. Edges above 20% are uncommon and often signal missing information, so check the match page before acting. Expected value also depends on probability: a +10% edge on a 25% outsider produces a far more volatile return than a +10% edge on a 60% favourite.
Staking value bets
Because value bets lose often in the short term, staking discipline matters more than selection. A fraction of the Kelly criterion – stake as a share of bankroll equal to the edge divided by the odds minus one, then scaled down to a quarter or less – keeps stakes proportionate. Our match pages show that calculation for every pick, capped at 2% of bankroll.
Has value worked?
The index's own performance is tracked on the leaderboard, with profit at level stakes over thirty and ninety days. If value bets are returning positive yield over hundreds of selections, the edge is real; if not, the model is mispricing something and the numbers will show it. For a broader probability view, use today's predictions; for line movement, see odds movement. Over 18s only.
A note on closing prices
The best test of a value bet is not whether it won, but whether its price shortened before kick-off. If the odds you took were longer than the final market price, you beat the closing line, which is strongly linked with long-run profit. The odds movement page shows how today's prices have moved since we first recorded them.
Frequently asked questions
What counts as value here?
Probability × odds of at least 1.06 – the price beats our fair odds by 6% or more – for odds between 1.60 and 4.50.
Are value picks more accurate?
No – they win less often than the main picks. Their purpose is a positive expected return across many level stakes.
Why do draws appear often?
Draws are under-backed by the public, so their prices can drift above the true chance in evenly matched games.