Divergence is the one setup in this collection that acts before price confirms anything. Every other strategy waits for the market to do something observable; this one acts on a disagreement between price and a derived measure, on the premise that the disagreement precedes a turn.

That makes it the most interesting construction here and the most easily misused. What follows is what the oscillator actually computes, what divergence is, and what acting early costs.

A Disagreement Between Price and Its Own Rate of Change

The relative strength index compresses recent price changes into a bounded number. It compares the average size of upward moves against the average size of downward moves over a lookback window and expresses the result on a nought-to-one-hundred scale. It is a summary of the character of recent movement rather than of position: it says how one-sided the recent changes have been, not whether price is high or low.

Divergence is a disagreement between that measure and price at successive swings. Price prints a lower low than its previous low, while the oscillator prints a higher low than its previous one. Price went further down; the downward moves that took it there were smaller and less one-sided than before.

The strategy reads that as a decline losing force, and enters long in anticipation of a reversal. The target is a swing back up; the stop sits below the low that has just formed, because price continuing through it is the evidence the decline had force after all.

Nothing in this requires price to have turned. The signal is complete while price is still making lows, which is the whole point and the whole risk.

An Exhausted Downswing

The condition is a decline that is running out of participants — each successive push lower achieved with less conviction than the last. When that is genuinely happening, the oscillator registers it before price does, because it measures the size of the moves rather than where they ended up.

What the setup needs is clean swing structure. Divergence is defined between identifiable swing lows, and a market without them offers nothing to compare. Choppy movement without distinct swings produces spurious divergences at whatever points the swing-detection happens to pick.

It also needs the decline to be a decline rather than a collapse. In a fast, one-directional move the oscillator saturates near its lower bound and stays there, and small variations in a saturated measure are not evidence of anything. The strategy is least reliable in exactly the conditions that produce the largest apparent opportunities.

The Price of Acting First

Because the level is identified in advance, entry can rest and pay the maker fee. That is a genuine advantage over the momentum and breakout strategies, and it follows from acting early: the strategy is not chasing anything, so it does not have to cross.

The real cost of a leading signal is not the fee, though. It is that the stop has to be far enough away to survive price making one more low, which the setup explicitly permits. A near stop would be triggered by the ordinary continuation the strategy is expecting to absorb, so the distance to being proven wrong is structurally wider here than in a strategy that waits for confirmation.

A wider stop with a comparable target means fewer trades need to work — but each failure costs more. The arithmetic is not better or worse than a confirming strategy's, it is differently shaped, and the shape is set by the choice to act before price agrees.

Divergence Can Persist

The defining failure is that divergence is not a turn. It can persist for many swings — price making lower lows while the oscillator makes higher lows, again and again — and each occurrence looks like a fresh signal. A decline can lose force and continue for a long time, because losing force is not the same as stopping.

This produces the worst pattern available to a strategy: repeated valid signals into a continuing decline, each entered on sound logic, each stopped out, with the setup appearing more compelling each time as the divergence becomes more pronounced.

There is also a definitional problem. Divergence depends on which swings are compared, and swing identification depends on the parameters used to find them. Different reasonable settings produce different divergences on the same price history, so two implementations can disagree about whether a signal exists at all. It is a less objective observation than it appears.

Finally, the oscillator is a summary and summaries discard information. Two very different price paths can produce the same value, so a divergence between price and the measure is sometimes a fact about the compression rather than about the market. The strategy cannot tell which it is looking at.

Divergence trades a real observation — that a decline's moves are shrinking — with the acknowledged weakness that shrinking moves can continue indefinitely. Acting before confirmation buys a cheaper entry and pays for it with a wider stop and the possibility of being early many times in a row.

Note: This explains how a process works. It is not legal advice, it is not specific to any debt, and it is not a substitute for a licensed attorney in your state. Rules and time limits vary by state and change over time — check the cited sources.