Who this entry is for — Readers who use divergences between indexes or sectors and want to distinguish historical Dow confirmation from a generic modern notion of “confluence”.
Confirmation between the averages is a cornerstone of William Peter Hamilton's systematisation. The industrial and railroad averages had to corroborate each other because they represented economically complementary activities: producing goods and transporting them.
Dow constructed an average made up mostly of railroad stocks in 1884 and the Dow Jones Industrial Average in 1896. Hamilton turned the joint observation of the two series into an explicit interpretive rule.
What confirmation means
In plain language — If only one average indicates a new movement, the information remains incomplete. When the other takes the same direction, the reading gains weight.
Hamilton specifies three points:
- the averages must move in the same direction;
- they do not have to break the level on the same day or in the same week;
- they do not have to produce highs or lows on the same date.
Confirmation concerns direction and structure, not numerical equality between the movements.
| Situation | Cautious reading |
|---|---|
| both exceed their respective resistance levels | movement corroborated |
| one breaks while the other remains in its previous range | signal not yet confirmed |
| one makes a new extreme that the other does not follow | non-confirmation to monitor |
| both reverse their structure in the same direction | stronger evidence of a turn |
What non-confirmation does not mean
A non-confirmation is not automatically a signal in the opposite direction. It means that the aggregate market has not yet provided the required corroboration.
Treating “absence of confirmation” as “bearish confirmation” or “bullish confirmation” is a logical leap. To identify a contrary movement, the averages must build that structure; it is not enough for one of them to lag behind.
Hamilton also accepts time lags. One average may confirm after the other; the delay becomes problematic only if subsequent behaviour negates the initial reading.
Confirmation of a line
The rule is especially important when the averages leave a narrow range. Hamilton describes a move above or below the respective limits as significant when both averages take the same direction.
The numerical levels of the two ranges may be different. What is compared is:
- the direction of the breakout;
- the structure relative to previous highs and lows;
- its place within the primary or secondary movement.
The role of volume
Popular versions of Dow Theory often add that “volume must confirm the trend”. The sources require a more precise formulation.
Hamilton observes that volume tends to expand in bull markets and contract in bear markets, with important exceptions during secondary movements. In the chapter devoted to lines, he also states that volume is relative to the state of the market and less significant than commonly assumed.
Therefore:
- volume may provide context for a movement;
- it does not replace confirmation between the averages;
- a universal volume threshold does not belong to the original theory;
- the modern formula “price and volume must confirm each other” is a later systematisation.
Contemporary adaptations
Today, industrials and transportation do not exhaust the structure of the economy. An analyst may compare indexes, sectors, breadth, futures, or related assets, but must state that this is an adaptation.
A reproducible test should specify:
- which series play the role of the two averages;
- what constitutes a breakout;
- how much delay is allowed;
- how a non-confirmation is classified;
- costs, data and evaluation period.
Without these rules, “confirmation” risks becoming a label assigned after the fact.
Sources
- William Peter Hamilton, The Stock Market Barometer, 1922, ch. I, pp. 4–8; ch. XIV, pp. 158–159; ch. XV, pp. 172–179, Internet Archive.
- S. A. Nelson, The A B C of Stock Speculation, 1903, chs. VII–IX, Internet Archive.
- S&P Dow Jones Indices, “The Dow — Creating an Icon”, history of the industrial and railroad averages.