Predictive Geometry is a Myt Technologies research programme exploring whether the structure of relationships between events can support predictive reasoning.
The approach represents events as nodes and their relationships as connections in a graph. It investigates how patterns emerge through construction, linking and the information that remains unused. Lottery histories provide a controlled experimental testbed, alongside the deterministic gameplay underlying Myt Puzzle.
Research development
The programme spans six complementary strands:
- Foundations: translating the original geometry framework into a modern Python research environment.
- Geometry Lab: building a deterministic graph engine with visual exploration, verified replay, reproducible experiments and early machine learning.
- Maze and Webs: exploring linear paths, branching networks and interconnected structures, then examining the information collected at their boundaries.
- Saucer: constructing linked rings and layered structures, with visual inspection of connections and unused numbers.
- Inverse Geometry: reconstructing structures around known historical outcomes, then testing whether models can propose new structures using earlier information alone.
- Rings and Chains: studying patterns in the historical events supporting a structure and testing whether those patterns improve number recovery.
How the research is tested
Construction and prediction are evaluated separately. Reconstructing a known outcome demonstrates a structural relationship; predicting an unseen outcome requires independent evidence.
Experiments retain source provenance, explicit search limits and replayable construction records. Forecasting studies use chronological evaluation, exclude the target and future observations from inputs, and compare results against simpler reference methods.
Current position
The programme has produced working research applications, validated graph constructions and reproducible experimental pipelines. A narrowly scoped learner has also reproduced a deterministic expert’s legal direction choices under its defined evaluation protocol.
These achievements establish research capabilities. The forecasting studies reviewed have not established a reliable predictive advantage attributable to the geometric structures. Further work focuses on identifying which relationships, if any, contribute useful information beyond simpler methods and testing them on unseen outcomes.
An ongoing research initiative by Myt Technologies.