Foreline
The fair price of every market, every line, every bet
Foreline turns the sharpest football market into a complete, verifiable fair-price reference — and puts it to work for you: pricing the lines nobody quotes, scoring the flow you accept, guarding the prices you post and the parlays you build. B2B data and API for operators, pricing vendors and market-makers.
One reference market, read honestly — and completed
Everything on this page stands on a single asset: a real-time, de-vigged, internally consistent reading of the sharpest football market — extended to the prices it doesn't quote, tagged with uncertainty, and cryptographically committed as it happens.
Model internals and data sources are proprietary. The output and the proof are what we publish — verifiable by you, not narrated by us.
Without this reference, money leaks in four places
Each leak below is a daily operational reality — and each has a product built on the foundation above to stop it.
You price beyond the reference — blind
The reference market quotes ~10 prices per match. Every alternate line and derivative you post beyond them has no honest benchmark behind it.
You learn who's sharp after they've won
Judging accounts by P&L takes thousands of bets. UK regulator data: 46.8% of limited accounts were already in profit when limited — cut late, and often cut wrong.
You move stale prices after they're hit
Across 3,000 daily markets some prices always drift from fair. Without a live benchmark you find out which ones from your losses, not your tools.
Your bet-builder misprices correlations
Parlays and builders drive 60–72% of book revenue — and one correlation failure cost an operator $934k under a single regulatory order (Massachusetts, 2025).
Four products, one foundation
The same fair-price surface, applied four ways. You don't have to change your pricing stack — each product plugs in as a reference, a score or an alarm.
Every fair price of the match — including the ones nobody quotes
The foundation itself, delivered as a feed. Where the reference market shows ~10 prices, the surface returns the whole picture — every rung of the handicap and totals ladders, 1X2 and computed derivatives — one consistent, de-vigged, uncertainty-tagged snapshot, refreshed continuously until kick-off.
- Extension beyond the quoted window, measured: on 5,781 first-listings of new lines our pre-listing price was unbiased to +0.03pp — accuracy comparable to the protective spread the reference book itself applies to those lines.
- Mutual consistency: all prices derive from one calibrated score matrix — no cross-market contradictions, and robust to feed staleness (measured: even day-old marginals price outcomes as well as closing ones).
- Main-line switch fact-stream: every reference main-line change as a typed, timestamped event, with the new ladder attached.
- As-of-time history for backtests — provably free of backfill, thanks to the public proof ledger.
- Format: REST/stream, normalised IDs, seconds of latency. Licence excludes bettor-facing use.
A future professional shows in the first 10–20 bets
The surface gives every bet an honest benchmark: the fair closing price of exactly the line the player took. A player who systematically beats that close is a professional — even while his balance with you is still negative. Judging by price instead of outcomes is why 10–20 bets are enough where P&L-based tools need thousands.
- Real-time account scoring via API: a continuous score plus an anytime-valid confidence level you can read after every bet and threshold yourself — no fixed number of bets to wait for. Family-wise false-positive rate 1.32%, measured on real recreational flow against a 2% budget set in advance.
- Any line, not just the main one — bets on lines the reference never quoted are scored through the same surface.
- No personal data — an anonymised account ID and the parameters of the bets.
- Limit weeks earlier — and stop cutting profitable ordinary players (today almost half of limited accounts are in the red with you).
- The betting footprint of bonus abuse (risk-free paired patterns) is flagged by the same score.
A fire alarm for stale prices
The surface knows fair for all 3,000 daily markets — so instead of "pricing more accurately", you simply get told which 20–40 of yours have drifted and are being eaten, prioritised by your own money at risk.
- Every alert is prioritised by money at risk (gap × your limit).
- Dual trigger: the gap is observably widening, or early "smart money" markers fire in the fastest books while the reference is quiet — a measured 2.4–2.8× elevation in the probability of an imminent big move.
- Direction-free by design: defensive action needs no direction — raise margin, trim the limit, or reprice to fair. We alert on risk; we don't predict the market.
- Delivery: SSE stream for latency, catch-up poll for completeness.
60–72% of your revenue is parlays. We find the holes in them
The same score matrix that keeps the surface consistent also prices joint outcomes honestly — which is exactly what a typical bet-builder does crudely, and sharp players milk.
- Step 1 — audit: we run yesterday's builder prices through the joint model → a report on which combinations cost you, and how much, in money.
- Step 2 — a live feed of consistent joint prices for your marginal combinations.
- Calibrated on 555,000+ settled outcomes — and robust: joint prices hold their accuracy even on stale marginals (measured, not assumed).
Player props (goals/shots for a specific footballer) are on the roadmap, phase 3.
Every number above is verifiable
The foundation is committed cryptographically as it is built — so quality is something you check, not something we narrate.
- A public cryptographic ledger. Every published row is hash-chained; every pass is
Merkle-committed; roots are anchored in Bitcoin via OpenTimestamps daily. The proofs and an
independent verifier are public:
github.com/foreline-io/foreline-proofs.
One command —
python3 verify.py— checks the daily roots, the commitment chain and the Bitcoin stamps. Selective disclosure and after-the-fact edits are cryptographically impossible. - A retro-audit on your data is the first commercial step: our historical price axis and account scores against your real betting history. You verify it, not us.
- CLV reconciliation from day one of the pilot: every published price is automatically reconciled against the actual closing price — with no right to cherry-pick.
Data and delivery
Data sources and model internals are proprietary and are not disclosed. Quality is evidenced by the public proof ledger, CLV reconciliation and a retro-audit on your data.
A pilot in a month, with zero risk
Request a retro-audit or pilot
Email us with your company and the markets you run — we'll reply with the next step. Starter and Pro are self-serve: you buy them yourself in your account (sign in with an email code, no password). Enterprise is per contract, and a trial key is issued on request from the same account.
Tell us who you are and what you'd like to explore. We usually reply within one business day.