Limitations
Known limitations
Everything that could make these results wrong or overstated. Stated here rather than left for a reader to discover. Three of these are severe enough to change how the numbers should be read.
The bottom line
Nothing on this site is a validated trading strategy. The findings most likely to be real are the descriptive ones — the divergence between the index and the median stock, and the extent of sector rotation. The findings most likely to be noise or artefacts are the predictive ones — the early-mover edge and the rate-channel relationship.
What would be needed to do this properly
- Point-in-time index membership instead of the 2026 constituent list — this is the single biggest improvement available and it requires historical constituent data, which is not freely available.
- A real total-return series (dividends reinvested) rather than price returns.
- Daily-frequency macro data to test the rate channel properly — annual data with n=10 simply cannot identify it.
- Explicit corporate-action records from an exchange source, rather than inferring unadjusted bonuses and demergers from price gaps.
- Out-of-sample validation on a different market or period. Nothing here has been tested outside Indian equities over a structural uptrend.