Predictive intelligence
A forecasting system that scores itself against the market, building toward a decision intelligence consultancy. We test in predictive markets, where closing-line value is one of our KPIs. Honest math. Transparent track. Slow on purpose.
148
audited bets
since 2026-03-12
since 2026-03-12
51.0%
strike rate
+10.1%
ROI on stakes
75-72
settled
win-loss record
win-loss record
Every bet logged. Every result settled. Closing-line value tracked. The numbers are not 97%. They are real.
The difference
What we do, and what we don’t
What we do
- Publish single bets above gate, sized to edge
- Document the reasoning per fire and per sit
- Re-check at T-45 when lineups land
- Settle and publish the result, win or lose
- Publish closing-odds for every bet (CLV currently being re-verified)
What we don’t
- Accumulator bait or “banker of the day”
- 96% accuracy claims
- Celebrity endorsements (real or AI-faked)
- “The AI does the work” mystique. A model produces numbers, a human decides
- Tipping. We're not tipsters. We're a multi-predictive intelligence system reading predictive markets
Where to follow
Analysis
Which football stats actually mean anything
Volume and behaviour carry over. A player who takes people on this season
takes people on next season, at a correlation of 0.89.
Passing, duels and tackles all hold up.
Efficiency does not. Shot accuracy keeps 12% of its separation and conversion keeps 12%. A screen that ranks forwards on how many of their shots go in is ranking them largely on last season's variance.
Read the full piece → · All research →
Efficiency does not. Shot accuracy keeps 12% of its separation and conversion keeps 12%. A screen that ranks forwards on how many of their shots go in is ranking them largely on last season's variance.
Read the full piece → · All research →
Why sport
The proving ground, not the destination
Sport is the harshest scoring environment we could find. Every forecast gets a price
attached, a closing line to measure against, and a result inside ninety minutes.
Nothing else marks your homework that fast.
So we bet. Not because tipping is the business, but because a bet is a forecast with money and a timestamp on it, and closing-line value is the only honest scoreboard here.
Where this goes is quantitative analytics in sport, and the same scoring discipline applied to decisions well outside it. This is a working track record, not a finished one. 148 bets is suggestive; 500+ is what significance looks like. We publish as we go so you can audit the build in real time.
So we bet. Not because tipping is the business, but because a bet is a forecast with money and a timestamp on it, and closing-line value is the only honest scoreboard here.
Where this goes is quantitative analytics in sport, and the same scoring discipline applied to decisions well outside it. This is a working track record, not a finished one. 148 bets is suggestive; 500+ is what significance looks like. We publish as we go so you can audit the build in real time.