MODEL ADOPTION #1634: REJECT — TG AUTO: EXTRACTED METHOD: e8_graph_embedding — E8 Intelligence Research
Year: 2026
Type: preprint
Abstract: DECISION: REJECT ELEMENT: No element survives. The "8-fold symmetry score" from sign co-occurrence graphs is a narrative artifact, not a tradable edge. It fails the hostility test: it is not concrete enough to code into a 15s–7m scalp without introducing a spectral-clustering black box that will overfit to the script corpus and add 200–500ms latency to a fast pipeline. The claimed "filter" on the 4 losses is hindsight—there is no evidence the symmetry score would have been low on those specific trades without fitting the threshold to the outcome. E8 geometry timing is already encoded in entry targets; adding a graph-projection layer is redundant and untestable at scale. EXPECTED IMPACT: 0% win-rate improvement; likely -0.1R to -0.2R per trade due to whipsaw filtering and missed entries from noisy graph projections. The 68% confidence is not backed by a single live-microstructure test, only a lab narrative. RISK: Overfitting to historical narrative patterns; spectral clustering on Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
Author: Caldin, Andrew Stewart
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7204121908
Cites:
Cited by:
Related to: