Betting odds explained: implied probability, margin and value
Betting odds explained: implied probability, the bookmaker’s margin, odds-on and odds-against, expected value, closing line value and how prices move.
Bankroll management and staking: Kelly, plans and variance
Bankroll management explained: setting a fixed fund, choosing a staking plan, what the Kelly criterion does, and why variance makes short runs misleading.
Reading football stats fairly: per-90, sample size and form
Reading football stats fairly: per-90 numbers, sample size, regression to the mean, form, game state, strength of schedule and home advantage explained.
Pressing and territory stats: PPDA, field tilt and xT
Pressing and territory stats explained: PPDA, field tilt, progressive passes, packing, expected threat and possession-adjusted numbers, with what each shows.
Prediction model metrics: Brier score, log loss, calibration
Prediction model metrics explained: the Brier score, log loss and reliability curves, plus overfitting and ensembles, and how to read them on a real record.
How football prediction models work: Poisson to simulation
How football prediction models work: Poisson goals, the Dixon-Coles correction, Elo ratings, Monte Carlo simulation, expected points and in-play probability.
Handicap betting explained: Asian, European and draw no bet
Handicap betting explained: Asian and European handicaps, draw no bet, double chance and void bets, with how each line settles and when stakes are returned.
Goalscorer markets and bet builders explained
Goalscorer markets explained: anytime and first goalscorer bets, how own goals and substitutes are treated, and how bet builders price correlated selections.
Goals markets explained: over/under, clean sheets, correct score
Goals markets explained: over/under lines, clean sheets, correct score and half-time/full-time, with how each settles and what drives the price.
Expected goals metrics explained: npxG, xA, xGOT and more
Expected goals metrics explained in plain English: non-penalty xG, expected assists, xGOT, goals prevented, set-piece xG and shot-creating actions.
Cash out and hedging explained, plus each way and arbitrage
Cash out and hedging explained: what an early settlement costs, how a hedge locks in a position, and how each-way and arbitrage bets work in football.
Why the Premier League table lies
After 28 matchdays the table says Man City are cruising. Our model says three of the top six aren't where they've earned the right to be.
Arsenal's defending is better than it looks
30 goals conceded looks worrying for a title challenger. Dig into the quality of chances they're giving up and a very different picture emerges.
How our match prediction model works
A match prediction model turns team data into a probability for each result. How the TACTIX model works: inputs, Poisson goal matrix, calibration and testing.
Introducing TACTIX: football intelligence
The kind of football analysis you read in The Athletic — passing networks, chance quality breakdowns — was produced using data that cost tens of thousands of
Why Liverpool keep dropping points away
At home Liverpool are near-unstoppable; away from Anfield something shifts. We dig into the possession and pressing data behind the gap.
Football player statistics, explained
TACTIX shows a lot of numbers for each player. This guide explains what the most important ones mean and how to use them when watching a match.
Champions League quarter-final predictions
Eight teams remain in the Champions League. Our model has run each quarter-final tie through its full two-leg simulation 10,000 times. Here is what we found.