RECORDED MODEL KPI / ONE REJECTED FORMULATION
Why this candidate stopped.
- Source
- XGCS hand-run research campaign, not the training gyms and not Kairos
- Family
- Cross-sectional divergence validation
- Scope
- Broad US equity panel
- Specimen
public-x2w-divergence-control- Calculation
x2w-divergence-turnover-v1
What this shows. A relationship that looked mildly positive on the years used to find it turned slightly negative on years it had never seen. Why it matters. It was dropped there, and the locked future test was left unopened rather than spent looking for a better answer.
NEVER MEASUREDthe 2026+ test was left unopened
What this evidence does not establish (4)
- This is one rejected cross-sectional divergence formulation, not a conclusion about all field or market-structure research.
- Selection and validation estimates are historical aggregate tests, not live trading results or future-performance claims.
- The locked 2026+ holdout was not evaluated, so no forward estimate exists.
- No implementation internals, raw series, or actionable outputs are included.
Both scores sit within three hundredths of zero, and neither was statistically strong. The vertical scale is expanded so the crossing is visible; the movement itself is small. That is the point. Small, unreliable effects are what most candidates look like.
Recorded answer: a candidate that looked positive in selection reversed in validation, failed the ordered controls, and did not consume the locked forward holdout.
See how Kairos formalizes qualification →