Delayed confirmation is not Dow Theory divergence

Two independent averages must agree

A Dow Theory inference is treated as reliable only when two independent price averages move together. A conclusion drawn from one unconfirmed average is regarded as likely to mislead.

Pairing an industrials average with a transportation average is often criticized as an obsolete production-versus-shipment model. The opposing view is that the two series were built to measure separate market sectors and function as a breadth check that other independent indices can replace.

Editorial note: the practical point of pairing two averages is two-index breadth. Independent market sectors are juxtaposed so a one-average reading can be confirmed or rejected.

Confirmation is directional sync

Confirmation is usefully read as whether the two averages are moving in the same direction, not only as a same-day break of a prior high or low.

Simultaneous new highs are the exception. After a secondary reaction, one average often leads while the other lags. That lag is delayed confirmation rather than bearish divergence.

Delayed confirmation means one average leads to new ground while the other is still advancing but has not yet cleared its nearby swing high. This is a timing lag, not a reversal label.

A split can be nonconfirmation

In December 1988 the transportation average exceeded its October high while the industrials average remained about 60 points below its October high. With both averages still advancing after a completed secondary reaction, that pattern is labeled nonconfirmation, not divergence.

Nonconfirmation is a temporary split in which one average has printed a new swing extreme and the other has not. It is not automatically a bearish divergence.

Divergence after an extended advance

Divergence applies when, after an extended advance, one average continues into new highs and the other fails to follow soon. The reading is that some turbulence or correction is more likely, not that a bear market has begun or that the next decline's size is known.

Not every correction is preceded by a divergence, but after an extensive advance such two-average splits often come before a decline of some degree.

Use the current swing sequence

Key highs and lows should be classified from the current primary-swing, secondary-reaction, and next-swing sequence. Distant historical peaks lose force when too much time or too many points have elapsed.

In the advance that followed the 1987 crash lows, both averages eventually cleared the prior-cycle highs only after about 24 months, showing why remote peaks delay primary-trend classification.

A primary bull classification can be made once both averages clear nearby swing highs even if one average is still below a much older record. Direction and which highs are key matter more than any single distant ceiling.