The price you see when you open Sportsbet or Bet365 to bet on an AFL match didn't appear by magic. It's the output of a specific process — part model, part market management, part response to observed betting behaviour. Understanding that process doesn't require insider information. Enough of it can be reconstructed from observable behaviour and published sports betting literature to form a reasonably accurate picture.
This guide covers the three-step process AU bookmakers use to set odds, what makes lines move after they're released, and why different bookmakers end up with different prices on the same event.
The three-step pricing process
Step 1: the model produces a baseline
Every bookmaker starts with a probabilistic model that produces an estimate of true outcome probabilities. For an AFL match between Collingwood and Essendon, the model might output:
- Collingwood probability of winning: 62%
- Essendon probability of winning: 38%
At fair no-vig odds, this would translate to Collingwood $1.61 and Essendon $2.63.
The inputs to AU bookmaker models typically include:
- Historical team performance — win/loss records, scoring margins, recent form across some weighted window.
- Player-level data — injuries, lineup changes, individual player performance metrics where available.
- Situational factors — home/away, travel, rest days, weather, venue (MCG vs smaller grounds have different scoring patterns).
- Market signals — opening lines and moves at other bookmakers, particularly sharp books like Pinnacle (not licensed in AU but monitored by AU books).
- Seasonal adjustments — models update throughout the year as teams' true strengths become clearer.
Smaller bookmakers often license pricing from third-party providers rather than building in-house models. Several specialist companies provide odds feeds that smaller AU books consume with minor customisations.
Step 2: margin is added
The model's fair probabilities sum to 100%. Bookmakers don't offer fair prices — they offer prices that include built-in margin (the vig). On the Collingwood example, a 4.5% vig produces:
- Collingwood probability (including margin): 64.8%
- Essendon probability (including margin): 39.7%
- Sum: 104.5% (book percentage)
- Collingwood displayed odds: 1 / 0.648 = $1.54
- Essendon displayed odds: 1 / 0.397 = $2.52
How much margin is added depends on several factors. Main-market AFL H2H typically gets 3-6% vig at AU bookmakers. Less-liquid markets get more. Novelty markets (first goal scorer, exact margin) get 15%+. The vig is calibrated to a combination of competition from other books, perceived risk of sharp action, and commercial considerations about which customers the book wants to attract.
Step 3: lines move based on action and signals
The opening price is rarely the closing price. Once the market is live, the bookmaker adjusts based on:
Observed betting action at their own book. If disproportionate money lands on Collingwood at $1.54, the book has accumulated exposure to Collingwood losing. To balance this, they shorten Collingwood further (say, to $1.48) and lengthen Essendon (to $2.65). This attracts Essendon bets to offset the Collingwood exposure.
Line moves at other bookmakers. AU bookmakers watch each other's prices constantly. When Bet365 moves Collingwood from $1.54 to $1.48, Sportsbet and Ladbrokes see this and typically follow within minutes. Nobody wants to be the outlier offering the best price on a side that the sharper books have decided is mispriced.
New information. Injury news, weather updates, late mail, confirmed team lists. The bookmaker's model re-runs with the new inputs and lines adjust to reflect updated true probabilities.
Sharp money patterns. When experienced bettors (sharps) concentrate on one side, the bookmaker's models flag this and may move the line further than the observed dollar amount would suggest. Sharp money carries informational weight beyond its size. See the NRL line movement piece for how to read these signals from outside the book.
Why different AU bookmakers have different odds
If the process above were deterministic, every AU bookmaker would converge on the same price. They don't, and the reasons are structural.
Different models. Bet365's model isn't identical to Sportsbet's. They weight inputs differently, have access to different historical data, use different algorithms. The resulting probability estimates for the same event can vary by 2-5 percentage points.
Different customer pools. Sportsbet's customers skew recreational AU punters with strong team preferences for Melbourne-based AFL clubs. Bet365's customer base is different. These differences affect the flow of money on each side, which affects how the book adjusts lines.
Different risk tolerances. Some books are more willing to carry one-sided exposure if they believe their model over the market. Some aggressively balance every position. These choices produce different line behaviour.
Different margin levels. AU bookmakers compete on vig for high-profile events and relax on niche ones. Bet365 typically runs tighter vig on AFL H2H than TAB does. Same model output, different margin added, different displayed price.
Different reaction speeds. When sharp money moves a line at one book, fast books (Bet365, Sportsbet) react within minutes. Slower books (TAB, Dabble) can lag 30-90 minutes. During those lag windows, price dispersion across AU bookmakers opens up, which is where arbitrage opportunities come from.
What a trading team actually does
Larger AU bookmakers employ trading teams (or “pricing teams”) who supervise the pricing models and make human judgements where automation falls short. Typical responsibilities:
- Managing the model. Ensuring inputs are current, tuning model weights, adding new markets as they become available.
- Monitoring exposure. Watching the book's running position on every live market, flagging when exposure exceeds tolerances.
- Manual line adjustments. When the model is clearly wrong, or when new information emerges faster than the model can incorporate it, traders override the automated price.
- Responding to sharp action. When specific customers' betting patterns suggest informed positions, traders adjust lines (and often restrict the customer accounts) to manage the exposure.
- Market opening and closing. Deciding when to open markets on upcoming fixtures, when to suspend markets mid-event, and when to resume them.
The human element matters most on low-liquidity markets and edge cases. For high-volume main markets, the trading team mostly supervises — the model handles the routine pricing and the line moves are automated.
The opening line vs the closing line
The opening line is the first published price. The closing line is the final price before the market closes (typically at kickoff). Between those two is where most of the interesting pricing behaviour happens.
Opening lines are generally the loosest. The book knows less, the exposure is minimal, and the price reflects the model's initial estimate. Early sharp action often finds value on opening lines because the book hasn't yet incorporated sharp signals.
Closing lines are the most informationally dense. They've absorbed every piece of information, every pound of sharp money, every line move at competing books. Closing lines are the most accurate outcome predictors sports betting has ever produced. This is why closing line value is the primary metric for measuring whether a bettor's process is actually sharp.
Why the bookmaker's price isn't the truth
It's tempting to treat the bookmaker's price as an authoritative probability estimate. It isn't — it's the bookmaker's estimate, combined with margin, adjusted for exposure, adapted to competitive signals. It's a business output, not a scientific measurement.
Several things can cause systematic mispricing:
Public-money bias. Popular teams get over-shortened. See the overrated teams piece.
Model blind spots. AU bookmaker models excel at main AFL and NRL markets but are consistently weaker at niche markets (NBL player props, A-League goals totals) because model investment has been smaller.
Recency bias. After a team has a surprising result, AU markets tend to over-weight that result in the following week's lines. Systematic fading of recency-biased markets has been a durable edge.
Information gaps. Late breaking news that hasn't reached the book's pricing team yet, or specialised information available to a small subset of bettors, creates temporary gaps between the offered price and the true probability.
These gaps are where advantage bettors extract value. The +EV guide covers the systematic approach to exploiting them.
Frequently asked questions
Do bookmakers use AI to set odds?
All major AU bookmakers use statistical models. Some use machine learning models specifically, particularly for large-liquidity markets. “AI” as a marketing term is often overstated; the underlying maths is largely classical statistics combined with modern compute.
Can punters actually beat the bookmakers at setting odds?
Punters don't set odds, so the question is whether you can identify mispriced odds and bet them profitably. Yes — arbitrage and +EV betting are both practical approaches. See the guides for the methodology.
Why do lines move differently at different books?
Different books have different customer pools, different internal exposure, and different models. Line moves reflect all three. When Sportsbet's customers pile onto Collingwood, Sportsbet shortens Collingwood faster than Bet365 does, so the two books' lines diverge until they converge again through broader market pressure.
How long does it take a bookmaker to open a market?
AU bookmakers typically open AFL and NRL markets a few days to a week before kickoff. NBA markets can open within hours of the prior night's games. Futures markets (AFL premiership, NRL grand final) run year-round with updates after each round.
Can a bookmaker refuse to take a bet?
Yes. Every AU bookmaker's terms and conditions reserve the right to refuse or void bets at their discretion. This is how account limiting works, and also how palpable-error clauses operate. See the bookmaker limiting piecefor how this plays out for winning customers.
James covers the AU bookmaker market — pricing mechanics, line movement, promotional structures, and how the corporate books actually operate. Previously worked in financial markets before moving to sports analytics.