Bet Better Open Model API
A free, open, read-only JSON feed of our sports model's win probabilities and
fair odds across nine sports. No API key, no sign-up, no rate limit worth worrying about.
No key required
CC BY 4.0
CORS enabled
HTTPS
9 sports
Quick start
Append ?format=json to any picks page on the site. That is the whole API.
curl "https://betbetter.world/afl/picks.aspx?format=json"
What the model publishes
Every entry is our model's own estimate for one selection: how likely we think it is to
happen, and what odds would make that a fair price.
| Field | Type | Meaning |
game | string | Fixture, formatted Away @ Home. |
gameTimeUtc | ISO 8601 | Scheduled start, UTC. |
market | string | Market type, e.g. Head to Head, Spread, Total Points. |
selection | string | The team, player or outcome being rated. |
line | number|null | Handicap or total the selection is measured against. |
modelProbabilityPct | number | Our estimated chance the selection lands, 0–100. |
fairOdds | number|null | Decimal odds implied by our probability — the break-even price. |
confidence | string | HIGH, LEAN or LONG-SHOT. |
verdict | string | One plain-English sentence summarising the above. |
Example response
{
"site": "Bet Better",
"sport": "AFL",
"type": "picks",
"updatedUtc": "2026-07-26T04:10:00.0000000Z",
"licence": "CC BY 4.0 — free to use with attribution to Bet Better (https://betbetter.world)",
"attribution": "Bet Better — https://betbetter.world",
"docs": "https://betbetter.world/api/",
"disclaimer": "Model estimates for research. Not a guarantee. 18+. Please gamble responsibly.",
"count": 1,
"picks": [
{
"game": "Geelong Cats @ Collingwood Magpies",
"gameTimeUtc": "2026-07-29T23:30:00.0000000Z",
"market": "Head to Head",
"selection": "Geelong Cats",
"line": null,
"modelProbabilityPct": 54.9,
"fairOdds": 1.95,
"confidence": "LEAN",
"verdict": "As of Jul 26, 2026, Bet Better's model rates Geelong Cats a 54.9% chance to land, which is fair value at odds of 1.95. Confidence: LEAN."
}
]
}
Endpoints
Three feed shapes exist for each league. best-bets returns game lines only,
prop-bets returns player props only, and picks returns both,
ranked. Add ?format=json to any of them.
| League | Base path | Feeds |
| AFL | /afl/ | best-bets, prop-bets, picks |
| MLB | /mlb/ | best-bets, prop-bets, picks |
| NBA | /nba/ | best-bets, prop-bets, picks |
| NFL | /nfl/ | best-bets, prop-bets, picks |
| NHL | /nhl/ | best-bets, prop-bets, picks |
| NCAAF | /ncaaf/ | best-bets, prop-bets, picks |
| UFC | /ufc/ | best-bets, prop-bets, picks |
| WNBA | /wnba/ | best-bets, prop-bets, picks |
| Tennis (WTA) | /tennis/wta/ | prop-bets, picks |
| Soccer (EPL, La Liga, Serie A, Bundesliga, Ligue 1, World Cup) | /soccer/{league}/ | best-bets, prop-bets, picks |
What this API deliberately does not include
No bookmaker prices. You will not find a bookmaker name, a market
price, or an implied probability anywhere in this feed, and there is no plan to add
them. Our odds data is licensed from a third-party provider under terms that permit
us to publish work derived from it but not to redistribute the feed itself.
Everything here is our own model's output. If you need live market
odds, license them directly from an odds provider.
Licence and attribution
Published under CC BY 4.0.
You may use, transform and redistribute it, including commercially, provided you credit
the source. A link back to betbetter.world is the
simplest way to do that.
Caching and fair use
- Responses are cacheable for 15 minutes; underlying model output refreshes several times a day.
- No authentication and no hard rate limit. Please cache rather than polling in a tight loop.
- Fixtures already started are excluded, so a feed is empty out of season. That is expected, not an error.
- On a data error the feed returns a valid document with an
error field rather than a 500.
Open data
We also publish a free dataset of 206,722 AFL player-match records from 2000 to
2023, with weather, attendance and advanced tracking statistics, under the same
CC BY 4.0 licence. It contains no odds data.
Model context
Probabilities come from per-sport models retrained on historical results. They are
estimates, not predictions of certainty — a 70% selection is meant to lose roughly
three times in ten. Read how the model works or check the
published results record before relying on any of it.
18+. Gambling involves risk and most people lose money.
This API is provided for research and analysis. It is not betting advice and not a
guarantee of any outcome. If gambling is causing you harm, support is available at
Gambling Help Online (AU)
or 1-800-GAMBLER (US).