PROJECT SCOPE
QUANTUM is an independent quantitative analysis project applied to football.
QUANTUM integrates mathematical, statistical and physics-based models within a proprietary system developed specifically for football analysis.
The objective is not to predict every match or necessarily produce selections every day. QUANTUM is designed to operate selectively: some sessions may generate multiple opportunities, while others may end without any officially published selection.
The internal architecture of the model remains proprietary.
PRIVATE DEVELOPMENT & TESTING
Before public tracking began, the current model configuration underwent approximately six months of intensive private testing and calibration.
Across the internal sample used during this phase, the system produced a strike rate above 60%.
This figure is provided solely as background on the development process. The private tests predate the public ledger, cannot be independently verified, and are not included in the official Public Alpha statistics.
Public tracking therefore starts from zero and represents the verifiable reference point for the future evaluation of the system.
MODEL OUTPUT
For each market analysed, the model produces a probabilistic assessment expressed as a percentage, publicly shown as Model Output.
This percentage is not derived from a single statistic and is not produced by applying the same uniform structure to every match.
The system uses complex quantitative modelling in which the specific context of the competition, the teams and the match influences the final assessment through proprietary parameterisation.
A Model Output of 70%, for example, represents the model's final probabilistic estimate for that specific event under the conditions analysed. It does not represent a guarantee of the outcome.
MARKET ODDS & EDGE
The odds used in the analysis represent the available market price or a reference average observed at the time of analysis and publication.
Each price incorporates an implied market probability.
QUANTUM compares that implied probability with the probability expressed by its Model Output.
The Edge represents the quantitative added value identified when the probability estimated by the model is more favourable than the probability implied by the market price.
In other words, Edge measures the theoretical advantage identified before the event, not the result achieved afterwards. A positive Edge does not imply that the selection must win and does not represent a guaranteed return.
SELECTION REVIEW & PUBLICATION APPROVAL
Outputs produced by the model are determined by the quantitative system and are not modified during the review stage.
The probability generated by the model, the associated Edge and the other analytical values remain exactly as originally produced by the system.
The review takes place only at a subsequent level: it determines whether a candidate should be promoted to an Official Position and therefore enter the public record.
This second decision layer also draws on more than 25 years of direct football observation and thousands of matches watched and analysed.
Human experience is not used to rewrite or correct the mathematical output. It is used exclusively as a publication filter.
An approved candidate is published before the start of the match and becomes an Official Position. A candidate that is not approved remains outside the official tracking record, regardless of the result subsequently produced by the match.
The classification is established before the event and cannot be changed retroactively.
RAW OUTPUTS & POSSIBLE FUTURE DEVELOPMENT
During the Public Alpha, the official record contains only selections that have been approved and published.
In the future, QUANTUM may choose to make the model's complete outputs available as well, clearly separating Raw Model Outputs from Reviewed / Official Positions.
The two categories would remain separate, allowing the performance of the unfiltered model to be compared over time with the performance of Official Positions that passed the Publication Review.
Any Raw Model Outputs published in the future would not be used retroactively to alter statistics that had already been recorded as official.
Candidates generated directly by the system.
Selections approved after Publication Review.
PUBLICATION & TRACEABILITY
Every Official Position is published before the start of the match.
Each selection is assigned a permanent identifier and a publication timestamp.
The timestamp makes it possible to verify when the position entered the public record.
Once published, the selection remains tracked regardless of the final outcome.
PUBLIC TRACKING
All published Official Positions contribute to the public tracking record.
Wins and losses are recorded under the same rules and contribute to the aggregate metrics displayed in the Performance and Archive sections.
Sessions in which no position is published are recorded as NO OFFICIAL SELECTION.
This label means only that the session ended without an officially published position. It does not necessarily describe what the system may have processed internally.
HOW TO INTERPRET RESULTS
A probabilistic model cannot be evaluated correctly on the basis of a single prediction or a very small number of events.
For this reason, QUANTUM uses progressive and permanent public tracking.
The objective of the Public Alpha is to build a growing public sample through which strike rate, result distribution, model-output distribution, probability calibration and the overall behaviour of the system can be observed.
Historical performance describes what has already occurred and does not guarantee future performance.
TRANSPARENCY PRINCIPLE
The core principle of QUANTUM is simple:
Transparency does not require disclosure of formulas, parameterisation or the internal architecture of the system.
It requires official decisions to be published before events, associated with verifiable timestamps, preserved in the historical record and evaluated under the same rules regardless of the outcome.
PUBLIC ALPHA // FREE ACCESS
QUANTUM is currently operating in Public Alpha.
During this phase, all published content, Official Positions, performance tracking and the Archive are available free of charge.
The purpose of the Public Alpha is to allow anyone to observe the system under real conditions and verify over time the relationship between ex-ante published selections and subsequent results.
Free access applies to the current public evaluation phase and does not necessarily represent a permanent commitment regarding the future structure of the project.