RSI divergence classes, ratio thresholds, and trendline tests

What this archive article teaches

This archive article teaches the relative strength index as a ratio-shift detector before it is used as a forecast overlay. The oscillator is a bounded reading of average gains versus average losses over a fixed lookback. In the historical workflow, price-versus-RSI mismatches are grouped into three classes, and a straight trendline is fitted to successive highs or lows so a break or a failed breakout can confirm or reject the RSI reading.

Divergence is a mismatch between successive price extremes and the corresponding RSI extremes. It is treated as a repeatable chart condition rather than a standalone order.

RSI as a ratio-shift detector

From RSI 50 to 66.66 the average-gain to average-loss ratio is described as remaining 1:1. Crossing 66.67 is described as shifting that ratio to 2:1. The 33.33 reading is described as a 1:2 shift toward down days. Those two readings are the RSI balance thresholds in this workflow.

The same oscillator is given typical support and resistance bands, treated as typical rather than absolute. In a downtrend, RSI resistance is placed in the 60 to 66.67 range with support near 20. In an uptrend, support is placed at 33.33 and resistance at 80.

Three divergence classes

Price-versus-RSI mismatches are grouped into three classes: simple divergence, hidden divergence, and multiple long-term divergence.

A simple divergence is a single mismatch that can form in either an uptrend or a downtrend. After that counter-move finishes, the prior dominant trend is described as likely to reassert.

Hidden divergence is described as less common. It is described as sometimes appearing after a simple divergence, and as the strongest divergence type for short-term swings.

A trend change is described as requiring at least three divergence points from an initial RSI reference. One bullish case is three successive lower price lows against three successive higher RSI lows. That three-point structure is the multiple long-term divergence used when the argument is that the prevailing trend itself is changing.

An editorial ranking of the three classes

TradersWeek editorial interpretation: rank the three classes by how much evidence they carry before a pause or a trend-change hypothesis is accepted. Simple divergence is the lightest claim, because it is a single mismatch framed as a short-lived counter-move after which the prior dominant trend may resume. Hidden divergence is the stronger short-horizon swing form. Multiple long-term divergence is the only class used here as a threshold for arguing that the prevailing trend itself is changing.

TradersWeek editorial interpretation: a pause hypothesis or a trend-change hypothesis is taken up only when a drawn trendline and the 33.33 or 66.67 balance levels agree with the divergence class already on the chart. The historical workflow supplies the classes, the bands, and the trendline test. The ranking and the joint acceptance rule are editorial.

Typical bands and the trendline test

A trendline in this workflow is a straight boundary fitted to successive highs or lows. A break or a failed breakout is used to confirm or reject the RSI signal. Counter-trend rallies in that framework are described as usually failing to retrace more than 50 percent of the prior decline.

A 10-minute S&P 500 episode

On a 10-minute S&P 500 chart, a rising-channel break near 2 pm Eastern on 27 February 2008 was followed the next session by a 14-period RSI move below 33.33, including a reading of 32.74 with a lower price low.

A later bounce was halted near RSI 60.18 at a declining trendline. An RSI spike to 73.45 was not confirmed by a lasting price breakout above that line.