How much do betting lines actually move?

11.7 million odds snapshots · 52 bookmakers, 13+ sports, 29 June – 3 August 2026 · Published 3 August 2026 · Companion dataset free under CC BY 4.0

Everyone in betting talks about line movement — steam, closing line value, "the market knows." Almost nobody measures it. The most-cited public evidence that closing lines are sharper than opening lines remains a single bookmaker's blog study from years ago. So we measured it: every pre-match price in our snapshot archive, roughly hourly, across 52 bookmakers — 11.7 million observations — tracking each market from the first time we saw it to its final pre-match price, with the bookmaker's margin removed so we measure real price relocation, not vig noise.

Finding one: movement is a property of the sport, not of betting

The average open-to-close move in a market's de-vigged win probability differs by nearly an order of magnitude across sports:

The pattern tracks information and liquidity, in opposite directions. UFC cards and AFL games carry concentrated news flow (weigh-ins, team sheets, weather) into thinner markets, so prices travel. MLB is the opposite: a deep, daily, heavily modelled market whose openers are already close to the truth. If you hear "the line moved, that means something" — in MLB it usually moved 0.6 points; in a UFC fight it moved four times as far.

Finding two: price discovery happens days out, not at the death

For in-season sports, we measured how far each price sits from its own closing value at different times before the start:

By 24 hours out, the market has already done most of its learning: prices sit within 0.6 points of where they will close. The final six hours — the window where most recreational money arrives — move prices by a third of a point. The market's opinion is essentially formed a day early; late bettors are price-takers.

Finding three: favourites shorten

Movement has a direction. In almost every sport, the opening favourite's de-vigged probability drifts up toward the close: +0.80 points in UFC, +0.50 in AFL, +0.44 in WNBA, +0.2 to +0.3 at Wimbledon. Openers systematically overprice longshots, and the close corrects it. This is the dynamic twin of what our calibration study found in settled prices — the bookmaker's margin and the market's early errors both live on the longshot side. A bettor who likes the favourite is, on average, better off early; a bettor who likes the underdog gets a better number late.

Finding four: the lines themselves move where scoring is high

For spreads and totals, the quoted line (not the price) is the market's forecast, and its movement is measured in points of scoreboard:

High-scoring, high-variance sports reprice in whole points; low-scoring markets hold their number and move the juice instead. And pre-season markets barely move at all — NFL and NHL lines posted months ahead moved 0.3–0.4 points across our whole window. Lines move when information flows, not with the passage of time.

What this means for a bettor, in plain English

• "Beating the close" is sport-dependent. In UFC or AFL there are real points to capture between open and close. In MLB the open and close are nearly the same number — there is far less closing-line value to win.
• The market forms its opinion 1–7 days out. If you bet in the last hours, you are accepting the market's settled price, for better and worse.
• Backing favourites? Earlier is better on average. Favourites shortened toward the close in nearly every sport we measured.

Method

All pre-match snapshots (roughly hourly) for head-to-head, spread and totals markets across 52 bookmakers, 29 June – 3 August 2026. Prices were converted to implied probabilities and de-vigged across each market's outcomes, so movement reflects price relocation rather than margin changes; spread/total movement is measured on the quoted line itself. "Open" is the first snapshot in the window — a floor on true movement for markets posted earlier (NFL, NHL). Exchanges excluded. The published dataset contains aggregates only (sport × market × time bucket): no odds, prices, bookmaker identities, teams, players or event identifiers, and no underlying feed can be reconstructed from it.

Citing this study

Bet Better (2026). How Much Do Betting Lines Actually Move? Line movement measured across 11.7 million odds snapshots. https://betbetter.world/studies/line-movement

Companion research: the Bookmaker Margin Study · Do betting odds tell the truth? · The 5% price and the 10% hold · the free model API.