GoalsProof GoalsProof Join beta

Public proof record

Every public pick is locked before kickoff. Results stay visible.

This is the GoalsProof accountability layer for the priced Over 1.5 model. The first 90-day evidence period is complete; the value-led model built from those results is now being tested publicly in Phase 2.

Loading public proof record.

Phase 2 live proof First 90 days complete 6/7 landed · settled priced picks only
Landed rate 85.7%
ROI +2.0%

Historical proof snapshot

Past results are public. Same-day picks stay inside the beta.

This static summary is included for search engines and AI answer systems. It excludes in-day selections and does not publish today's full board.

Main O1.5 public model 391/471 landed

83.0% of settled historical public Over 1.5 picks.

Average locked odds 1.18

Average recorded O1.5 price across settled public picks.

Worldwide O2.5 lab 399/540 landed

73.9% of settled experimental O2.5 selections landed.

BTTS lab 530/823 landed

64.4% of settled experimental BTTS selections landed.

Historical snapshot updated 17 Nov 2026. Live rolling figures below are calculated from the complete proof logs.

What the first 90 days taught us

High accuracy did not automatically beat the price.

The first complete evidence window landed 84.3% of 127 priced picks but returned -2.0% at the recorded odds. We used that result to retire broad Market Supported selection and replace it with a tighter Realised Value Build.

Phase 2 remains a full 90-day forward test, not a rewrite of the original record. A public-launch review can begin earlier once the sample has at least 100 settled picks, 30 calendar days and positive overall and latest-30 ROI.

Phase 1 · complete -2.0% ROI 127 settled picks · 84.3% landed

19 May to 16 August 2026. This is the benchmark the revised model must beat.

What changed Evidence became rules Market Supported retired

Selection shifted toward the price and probability bands that demonstrated realised value, while the full earlier record stayed visible.

Phase 2 · now testing Calculating ROI Loading forward evidence

15 October 2026 to 12 January 2027 · launch review after at least 100 settled picks and 30 days.

Route out of beta

Launch review starts when sample, time and price performance agree.

Passing all four gates opens a manual review of result QA and league concentration. It does not automatically remove beta status, and the full 90-day record will keep building in public.

  • Waiting100 settled new-model picks
  • WaitingPositive locked-price ROI
  • WaitingPositive latest-30 ROI
  • WaitingAt least 30 days of forward testing

Main priced O1.5 record

Rolling proof performance

Official public selections are model-selected, not personalised. User proof logs stay private; this page tracks the same GoalsProof public picks for everyone.

Main proof record Realised Value O1.5

This is the current trust anchor: evidence-led priced picks, locked before kickoff and tested prospectively while the earlier record stays visible.

Action layer Suggested accas

The same public shortlist is grouped into Clean 3, Balanced 4 and Higher-risk 5 proof streams, measured separately from single picks.

Experimental markets Labs

Worldwide O2.5, BTTS, context and weather tests are tracked below the line and do not change the public proof headline.

The differentiator

No cherry-picked screenshots. No deleted losing runs.

The public proof page is designed to answer the question every user should ask: does the adapted priced Over 1.5 model beat the recorded price when every pick is saved before kickoff and left visible afterwards? The first 90 days remain part of the record while Phase 2 is measured separately. Its launch review can start after the published gates pass, while the full 90-day test continues.

See the full public history
Step 1 Locked before kickoff

Public selections are timestamped before matches start, so the record cannot be rewritten after the result is known.

Step 2 Results stay visible

Wins, losses, pending results and profit/loss all remain on the public page.

Step 3 Performance tracked over time

Hit rate, profit/loss and longer-term performance are tracked across rolling windows.

How to read this page

Four simple labels matter.

Locked

The pick was saved before kickoff.

Landed

The match finished with 2 or more total goals.

Missed

The match finished with 0 or 1 total goals.

ROI

The profit/loss if every public pick was followed equally.

Today’s public snapshot

Today’s public proof status.

Loading today’s main priced public shortlist.

Recent proof summary

The latest settled proof record, in plain view.

Loading the latest settled proof summary.

Performance context

Is the record beating the price?

A high landed rate only matters if it beats the breakeven rate implied by the locked odds.

What does this mean?
  • Landed rate is how often the public picks finished with 2 or more goals.
  • Average locked odds is the average price recorded when the pick was saved before kickoff.
  • Breakeven needed is the hit rate those odds require just to break even.
  • Vs breakeven shows whether the record is ahead or behind that price benchmark.
  • Short prices are sensitive to rounding, so this panel uses the precise average locked odds for the breakeven check.
  • Level-stake return shows the result if every public pick was followed equally.

The key number is Vs breakeven. Positive means the proof record is beating the average price; negative means the hit rate is not yet high enough for the odds being taken.

O1.5 public record

Main priced proof history

Showing official main-model Over 1.5 public selections only. Today’s live shortlist is kept as a small snapshot above.

Filter the public record Loading records
Date Fixture Competition Source Confidence Price Context Result P/L

Suggested acca record

Where the shortlist turns into groups.

Loading suggested acca proof. These groups are built from the same public shortlist and judged separately, so the single-pick proof record stays clean.

Recent suggested groups Locked before results, settled after matches finish.

Experimental evidence below this point

What is proven versus what is still being tested.

The headline record above is the main priced O1.5 public proof. Suggested accas are tracked separately from those same public locks. Below this point, worldwide O2.5 and BTTS are lab streams only.

Worldwide O2.5 / BTTS Lab

Experimental worldwide markets tracked separately

Loading worldwide Over 2.5 and BTTS shadow evidence.

Why this is below the main record

These are worldwide priced-league research streams for possible future markets. They do not count towards the main priced O1.5 public proof headline.

Advanced model checks Open selection rules and diagnostics

Selection rule

How public picks are chosen

Every public pick is selected by a fixed rule before kickoff. The rule is shown here so results cannot be cherry-picked after matches finish.

Probability shortlist Research only

The strongest Over 1.5 candidates remain visible in the app, but are no longer presented as an official value claim.

Value Build Qualifying picks

A narrow, historically validated price cohort with clean context and at least six bookmaker prices. It is now measured forward as the official value record.

Accountability

Simple proof summary

ROI is calculated to flat 1-unit stakes using the recorded Over 1.5 price where available.

Market proof

Did the price move in our favour?

This compares the price recorded when GoalsProof locked a pick against the later market price. If the locked price was better, that is a positive sign over time.

Shortlist lanes

Value Build and legacy lanes

Value Build is the current price-led rule. Older Most Likely and Market Supported records remain visible rather than being rewritten.

Calibration

Do higher-rated picks perform better?

Bucket performance shows whether stronger model confidence has actually converted into better realised outcomes.

Context performance

Do high-risk fixtures underperform?

This tracks whether playoff, relegation, knockout and unusual-context fixtures actually land less often than standard public picks.

Context gate evidence

Excluded does not mean ignored.

Fixtures filtered out by the stricter public proof context gate are still tracked in shadow mode, so we can prove whether excluding them improves the public record.

Heat weather evidence

Does heat slow games down?

Hot-weather fixtures are tracked in shadow mode so GoalsProof can test whether heat watch and heat risk labels actually reduce Over 1.5 performance.

How picks are recorded

How the audit trail is created

The public record is built from repeatable stages: shortlist lock before kickoff, closing price snapshot, and result sweeps after matches complete.

Candidate rule test

Robust value rule under test

This challenger runs beside the public shortlist without changing official customer-facing picks. It remains experimental until its live sample, ROI and closing-line evidence are strong enough to justify a reviewed model change.

Responsible context.

This is a public performance record, not a promise of future results. Football remains uncertain, and even high-confidence selections can lose. The purpose is to make the process accountable over time.

GoalsProof is decision support only. It does not place bets, handle stakes, guarantee returns, or provide personal financial advice.