Confirmation delay and breadth divergence in a two-average case

What makes a swing forecast-relevant

Dow Theory treats a move as forecast-relevant only when the industrial average and the transportation average confirm each other. Confirmation is the same directional message from both averages before a swing is treated as forecast-relevant. A nonconfirmation is a signal printed by one average and not repeated by the other, so the isolated reading is treated as incomplete and potentially deceptive.

Divergence is the sharper split: one average makes a new extreme while the other fails to follow or makes an opposing extreme. After prices have advanced for several months, especially into new high ground, a new high in one average that the other fails to match is a warning of possible weakening. The method does not estimate how far or how long any decline will run.

Delayed confirmation is still a classification

Joint-confirmation, when both averages confirm at roughly the same time, is uncommon. Delayed-confirmation can still classify the trend when direction, intensity, and the most recent confirmed signal are taken together. Delay alone does not cancel trend classification.

A rule that only names the trend after both averages cross a predetermined prior point is statistical-trend-classification. That rule is a hindsight classification. Directional-trend-classification can support earlier directional inferences from volume and from where prices sit relative to the length and size of the primary-trend and secondary-reaction structure.

A long advance can still be read as a confirmed primary bull market when the two averages repeatedly reconfirm, even if a casual reading keeps finding apparent nonconfirmations.

A historical confirmation failure

In the September to October 1989 case, both averages advanced together into early September. Industrials then made a new October high while transports failed to exceed their early-September peak. That split is a divergence at the intermediate-swing scale: confirmation and divergence at this scale flag strain without stating how large the next move will be.

Breaking recent reaction lows is not automatically a primary-trend reversal. The prior lows must first be judged important enough to count as a secondary-reaction or an intermediate-swing. After an unconfirmed industrial high in early January 1990, the later break of the October 1989 intermediate lows was read as the industrials confirming the already-weak transportation average and as a primary-trend warning. The primary-trend is reclassified only after intermediate or secondary-reaction points are taken out and confirmed.

Split breadth after the warning

After that primary-trend warning, large-capitalization industrials could still occupy a wide trading range and print later highs, while the transportation average and more speculative over-the-counter prices deteriorated. That pattern is split market-breadth rather than a single-index verdict. Participation across the industrial average, the transportation average, and more speculative listings had come out-of-gear.

After the last jointly confirmed high, a sequence of higher industrial highs against lower transportation highs is another nonconfirmation inside a still-classified bull phase. It is treated as a caution about strain, not as a completed trend change.