13
Research forecasts trackedCaptured in a manual test. Outcome matching is in progress.
AIONORA / PLAY THE LONG GAME
Every prediction has a story.
See what we research, what the results reveal, and how our models develop. Browse our forecasts, inspect what the results show, and see how the models improve. Every published outcome keeps its research context.
Open researchReal resultsModel learning
Current Daily CORE: Read the latest research edition. The figures below are historic research snapshots, not an independently certified betting record. Official ROI is not yet available.
Inside the research
Archived research snapshot · 25 October 20265
Recent cases reviewedA documented retrospective research review, separate from live forecasting.
5,161
Historical matches studiedBackground dataset used to understand team and market behavior.
917
Out-of-sample test matchesIndependent historical evaluation period for three football markets.
Early-stage research: Our prospective forecast outcomes are being matched to their source results. Historical match results and retrospective case studies are shown separately from published selections. Metrics appear when there is enough comparable evidence.
LATEST REVIEW
What the previous review found
4 / 5Research market cases with favorable outcomes in the 24 October retrospective review
One case went against the research view. That is exactly why AIONORA studies misses as well as hits. The research combination in this review did not succeed.
These five reviewed cases were not established as a complete, prospectively published selection cohort. This ratio is not an official hit rate or demonstrated betting profit.
HOW WE LEARN
From forecast to insight
01
Record the viewMarket, probability, source and pre-match timestamp where available
02
Match the final scoreResults can be source-matched without requiring bookmaker prices
03
Test what changedCompare outcomes and probability calibration over meaningful samples
Forecast journal
Research first · outcomes followFootball — research archive available
We are connecting the initial 13 frozen research forecasts to completed matches. Individual prediction outcomes will be listed only after an event-ID and source-based match. No missing prediction is counted as a win or loss.
Outcome matching in progressMODEL PERFORMANCE
Accuracy, not just outcomes
A hit rate alone can be misleading. Our testing also asks whether forecasts assigned a 70% probability win close to 70% of the time.
917
Separate historical test matchesThree football goal-related markets evaluated retrospectively
↗
Small experimental improvement in Over 2.5Not statistically conclusive; no profitability claim
DISCIPLINED DECISIONS
Knowing when not to pick
We publish research observations more freely than formal betting selections. A research forecast can be tracked without being presented as a safe bet.
The strictest checks remain for investment-like claims: independently tested probabilities, appropriately fresh prices and a defensible advantage.
What do “tracked”, “source-matched” and “official” mean?
Tracked: a forecast was recorded before the event with sufficient identity and market context. Source-matched: a completed event has been matched using a traceable result source; this does not require an executable bookmaker quote. Independently checked: a second independent source or audit supports the result. Official: a published, certified recommendation meeting the stricter release criteria. Historical results, retrospective research and official picks are never silently combined.
The count of completed outcomes refers to source-matched prospective research records; it is not the number of available historical match results. Official selection statistics will appear when qualifying records exist.