Inconclusive · Filters · 2026-10-10
Do RSI, MA or RSI-divergence conditions help on any timeframe? (catalogue re-check)
What this means
Of 16 checks (four conditions on four timeframes), 15 showed no clear difference. The exception was daily setups with RSI at entry ≥ 48, which earned +0.32R against +0.08R for the rest, and the gap held in both halves of the decade. But with 16 unadjusted tests, one apparent hit is about what chance alone would produce. It is a lead worth watching in forward testing, not a finding.
The explanation is our best reading; the data shows the result but not the cause directly.
By timeframe
| Timeframe | Verdict | n | Result |
|---|---|---|---|
| 15-minute | Not supported | 16,496 | all four conditions within +/-0.01R |
| 1-hour | Not supported | 13,208 | all CIs span 0 (divergence -0.08) |
| 4-hour | Not supported | 3,990 | all CIs span 0 |
| Daily | Inconclusive | 2,098 | RSI at entry >= 48: +0.25R (+0.09, +0.38); others span 0 |
- Hypothesis
- RSI at entry >= 48 (the old chart setting), RSI at D <= 40, price below the 100 MA, or bullish RSI divergence (rd=1) improves mean setup R.
- Method
- Re-run on the saved TradingView trade logs (features tagged per trade): 4 conditions x 4 timeframes (15m/1h/4h/1d), in vs out, cluster-bootstrap CI. No multiple-testing correction.
- Result
- 15 of 16 differences have CIs spanning 0 (e.g. 1h divergence -0.08 (-0.17, +0.01); 15m all within +/-0.01). One exception: daily RSI at entry >= 48 +0.324R (n=1,669) vs +0.075R (n=429), diff +0.25 (+0.09, +0.38); it holds in 2016-20 (+0.53 vs +0.25) and 2021-26 (+0.17 vs -0.06).
- How much to trust it
- Medium: TV fills, but 16 post-hoc tests, so one 'hit' is expected by chance
- Did it change the live rule?
- No.
- Notes
- The daily RSI>=48 result is a lead for forward testing, not a finding. 'Price < MA100' is measured at entry, not the Pine 'Point D below 100 MA' option, which was never swept.
Scripts and data for this study will be linked here when the research code is published on GitHub.
Education and research only. Not financial advice. Backtests use today's index members (survivorship bias).