Advance-decline line divergences

The advance-decline line is the oldest breadth tool and still one of the best. It is nothing more than a running total of the daily advancers-minus-decliners number, and its whole value comes from one property: every stock counts once, so it cannot be carried by the megacaps.

Building it

Each day, take the number of advancing stocks minus the number of declining stocks and add it to yesterday's total.

The absolute level is meaningless; it depends on when you started counting. Only the shape matters, and only compared to price.

Most traders use the NYSE composite version. A version built from S&P 500 members only, or from Nasdaq stocks, exists too and behaves a little differently, which the last section covers.

Confirmation and divergence

Price:     ___/‾‾\___/‾‾‾\____/‾‾‾‾\        higher highs
A/D line:  ___/‾‾\___/‾‾\_____/‾\____        lower highs   ← negative divergence

Price:     ‾‾‾\___/‾‾\____/‾‾‾\_____/‾‾‾    index breaks out
A/D line:  ‾‾‾\___/‾‾‾\___/‾‾‾‾\___/‾‾‾‾‾   A/D confirms   ← healthy

Confirmation: price makes a new high and the A/D line makes a new high within a few days. The rally is broad. Historically, bull markets spend most of their life in this state.

Negative divergence: price makes a new high and the A/D line fails to. Fewer stocks are participating in each successive push. The rally is narrowing.

Positive divergence: price makes a new low and the A/D line holds above its previous low. Selling is exhausting in the average stock even as the index heavyweights drag the headline lower. Less common and less reliable than the negative version, but worth noting near the end of declines.

Key idea: The A/D line gives every stock one vote. A new high in price without a new high in the A/D line means the vote is narrowing, and narrow rallies are more fragile than broad ones.

What the record actually shows

Negative A/D divergences preceded several of the largest tops on record, and the lead times were long:

Look at the first row carefully. The A/D line topped almost two years before the index. A trader who shorted on the divergence in 1998 sat through one of the strongest rallies in history. That is the honest summary: the divergence identifies a narrowing market; it says almost nothing about when the narrowing will matter.

The reverse error is also common: not every negative divergence leads to a major top. Many resolve when breadth catches up after a short correction. In practice, a divergence of a few weeks is background noise; a divergence of several months that survives a correction and a new price high is the kind that has preceded serious declines.

How to use it without fooling yourself

1. Require a length. Only count a divergence that has persisted through at least one full swing in price: a pullback and a new high with the A/D line still below its own high. Single-day mismatches are noise.
2. Treat it as a regime label, not a trigger. "Breadth narrowing" goes in the morning regime sentence and lowers size on new long breakouts. It does not open shorts.
3. Wait for price. The divergence resolves when the index breaks a swing low. Until then the narrow market is still rising.
4. Combine with the other Module 3 tools. Percent above the 200-day, new highs vs lows and the McClellan summation should be telling the same story. If they are not, the A/D divergence is probably an artefact.

The interest-rate artefact

The NYSE composite includes closed-end bond funds, preferred shares and other rate-sensitive listings. When rates rise sharply these decline together and drag the A/D line down even while common stocks are fine, producing a false negative divergence. When rates fall they inflate it.

Fixes: use a common-stock-only A/D line (some data vendors publish one), or an S&P 500-member A/D line, or check the divergence against the equal-weight ratio from Lesson 4, which contains no bond funds. If the composite line is diverging and the S&P 500 line is not, rates are probably the cause.

Volume-weighted variants

An up-volume minus down-volume cumulative line (the daily version of VOLD, accumulated) weights each day by the money behind it. It tends to be noisier but catches the "broad but thin" rallies the count-based line misses. A few traders keep both and only act on divergences that appear in both.

Weekly view

Plot the A/D line and the index on weekly bars for a slower, cleaner read. Divergences that are visible on a weekly chart over several months are the ones with historical significance; the daily chart is for spotting them early.

Try it: Chart the NYSE A/D line under your index for the past two years. Mark every index high. For each, note whether the A/D line made a new high within five trading days. Count confirmations and divergences, then look at what price did in the following month after each. Keep the counts; they are your own base rate.

Recap

• The A/D line is the cumulative sum of daily advancers minus decliners; level is meaningless, shape vs price is everything.
• Confirmation (new highs in both) is the normal state of a bull market; negative divergence (price high, A/D lower high) means narrowing.
• Divergences have preceded major tops but with lead times from months to two years; they describe fragility, not timing.
• Require persistence through a full swing, treat it as a regime label, and wait for price to break before acting.
• Rate-sensitive NYSE listings can produce false divergences; check a common-stock or S&P 500-member line.