Divergence, Across Indicators
Price makes a new extreme and an indicator does not. It is the most cited observation in chart reading and one of the weakest, for a reason that follows from the construction.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- Divergence means price and an indicator disagreed about a new extreme.
- Most indicators measure rate of change, so deceleration produces divergence automatically.
- Trends decelerate several times before they end.
- The false positive rate is high and only the last instance is remembered.
- It is stronger when it coincides with something independent of price.
MAD Academy Training Video · 0:45
When Price and Momentum Disagree
Divergence is the one oscillator read that survives a trend — and it warns of a change in pace, never of a reversal.
This lesson is part of a Stock Alerts + Tools plan.
What it is
Price makes a higher high; the oscillator makes a lower one. That is the observation, and it is available on any indicator plotted alongside price.
| Type | Price | Indicator | Conventional reading |
|---|---|---|---|
| Regular bearish | Higher high | Lower high | The advance is weakening |
| Regular bullish | Lower low | Higher low | The decline is weakening |
| Hidden bearish | Lower high | Higher high | Continuation of a downtrend |
| Hidden bullish | Higher low | Lower low | Continuation of an uptrend |
The existence of hidden divergence, which reads the opposite way, means every combination of price and indicator extremes has a named interpretation. A framework with a reading for every case rules nothing out.
Why it appears so often
Most oscillators measure a rate of change or a position within a recent range. A trend that continues while decelerating produces a lower indicator high automatically, because deceleration is precisely what those indicators measure.
Deceleration is not reversal. Most trends decelerate several times without ending, which means most divergences are followed by continuation, and only the final instance before an actual turn is remembered.
Scroll the chart sideways to see all of it.
What strengthens it
- Coinciding with a level that has held before, which is independent of the indicator.
- A change in volume behaviour, which is the one standard input that is not another view of price.
- Occurring on a higher timeframe, where each bar aggregates more trading.
- A failure to make a new high at all, which is a fact about price rather than about a derived series.
Every item on the list adds something the indicator does not contain. Two oscillators diverging simultaneously adds nothing, because they were computed from the same series.
Stating it honestly
The defensible statement is narrow: the rate of advance has slowed. That is a real description, it is worth noticing, and the gap between it and a reversal is where most of the claims made about divergence live.
Counting it honestly
The claim that divergence precedes reversals is testable and the test requires counting every instance rather than the ones that were followed by something.
- 1Define it mechanicallyA specific rule for what counts as a divergence, applied without judgement.
- 2Count every occurrenceOver a long history and across many securities, including the ones nothing followed.
- 3Define the outcome in advanceA stated move within a stated period, rather than eventually.
- 4Compare against the base rateHow often that outcome occurred without a divergence preceding it.
The fourth step is the one that decides it and the one almost never performed. If reversals of the stated size occur at a similar rate without a divergence, then the divergence added nothing however often it appeared to work.
This is the general procedure for evaluating any chart observation, and divergence is the clearest case for it because the observation is frequent, easy to define mechanically, and universally believed.