When RSI divergence deserves a second look
Divergence is the most over-marked pattern in retail chart rooms. A price high with a slightly lower RSI peak gets circled, a position is sized, and three sessions later the trend continues without the reader ever asking whether momentum context supported the call.
At our lab we teach divergence as a filter stack — not a standalone trigger. Before you mark a regular bearish divergence (price higher high, RSI lower high), run through these four checks.
1. Trend permission
Divergence against a strong trend fails more often than textbooks admit. In a clean uptrend on the daily frame, we require at least two weeks of RSI rolling over from above 60 before a short bias even gets pencil time. In a range, the same RSI roll from 70 carries more weight because mean reversion is the dominant force.
2. Swing quality
Both price swings and RSI swings must be obvious on print — not hairline differences. We use a simple rule: if you need to zoom below daily resolution to see the RSI peak shift, skip it. Our workbook uses a minimum three-bar separation between RSI peaks.
3. Volume or participation cue
On VN equities we look for declining volume on the second price push. FX readers might substitute a narrowing daily range. Divergence without a participation tell is a drawing exercise, not a setup.
4. Location on the chart
Divergence at fresh all-time highs behaves differently from divergence into a known resistance zone you marked weeks ago. We ask students to write the horizontal level first, then ask whether RSI adds information or merely confirms what structure already showed.
What we do in the lab
Students receive twelve historical examples — six that passed all four filters and six that failed at least one. The failed examples are more instructive. Most bad divergence trades in our alumni surveys came from skipping trend permission.
If you want hands-on marking with instructor feedback, the Momentum & RSI Intensive dedicates all of day two to divergence and its limits.