ADX awards its point to bulls and bears simultaneously, because it has no direction to give.

The asymmetry

The strength condition is identical on both sides of the calculation.

Structural position, momentum acceleration, participation, gear alignment and volatility direction all have separate bullish and bearish forms — price above the cloud or below it, the histogram expanding upward or downward, the oscillator beyond sixty or beneath forty. Trend strength has one form. The condition is that the reading is above its floor and rising, and that condition is direction-blind by construction. When it holds, it holds for both evaluations, and both totals receive a point.

Why it is in the stack anyway

It is not there to pick a side. It is there to say whether sides are meaningful.

A directional read in a market with no trend strength behind it is a description of drift. The strength component's job is to establish that the market is doing something decisive enough for a continuation thesis to apply at all, and that question genuinely has no direction in it. Removing the component because it cannot discriminate would remove the only test of whether the other five are describing a real move or a slow shuffle that happens to be pointed somewhere. It is a precondition wearing the same uniform as the contributors.

The consequence for the scale

One point of every score is shared, so the effective discriminating range is five.

In a strongly trending market both scores carry the strength point automatically. The gap between the two totals is therefore produced entirely by the other five components, and the shared point sits underneath both like a floor. This is worth internalising because it slightly flatters the low end: a score of one in a strong trend may consist of nothing but the strength point, which is a reading with no directional content whatsoever. A score of one in a quiet market means something different again. The number alone does not distinguish them; the row does.

The reading it makes possible

Strong strength with split direction is the most specific warning the stack gives.

The configuration to watch for is the strength row passing while the directional rows are divided. The market is moving decisively and the components cannot agree on which way — which is the signature of a fast reversal or a violent range rather than a continuation. It is also a state that generates conviction, because the movement is real and obvious. The stack's response is to hold both scores in the middle, and the strength row is what identifies the middle as dangerous rather than merely dull.

• Strength passing plus split direction is not a weak trend; it is an unresolved one.
• Strength failing with clean directional rows is the drift case — right idea, wrong conditions.
• The two cases look similar in the total and are immediately distinguishable in the rows.

The general form

A precondition folded into a score stops behaving like a score.

There is a design tension here worth naming. Elsewhere the system treats preconditions as gates, precisely because a precondition scored alongside contributors can be traded off against them — a strong enough combination of other components could otherwise purchase the missing requirement. The continuation stack accepts that tension deliberately: it scores the strength component rather than gating it, so that its state stays visible in the recorded number instead of vanishing into a filter. The cost is that the number needs its rows, which is already true for every other reason.

The key idea

Know which of your inputs cannot take a side before you read agreement as consensus.

Agreement between components only counts when the components were capable of disagreeing. One member of this stack is structurally incapable of disagreeing with itself across directions, so its contribution appears on whichever side eventually wins and never distinguishes between them. That does not make it worthless — it makes it a different kind of input, and the reading improves as soon as it is treated as one. The same test applies to any composite: separate the parts that could have said no from the parts that were only ever going to say yes.