How Conviction Levels Are Calculated (High, Medium, Low)

Conviction in one sentence

Conviction answers the question 'how much should I trust this setup?' — not 'how big is the move?'. A pair can have a large directional score and still be Low conviction if the signals contradict each other or the data behind them is thin.

StraviaX surfaces three levels: High, Medium and Low. They are decided by a transparent rule set, not a black box, and you can audit every input on the Pair Detail page.

The three blocks that feed conviction

Every pair is graded by three independent blocks: Macro (growth, inflation, employment, central bank stance), Sentiment (a fusion of COT institutional positioning and retail crowd positioning), and Seasonality (the pair's historical tendency for the current week/month).

Each block ends up with two attributes: a direction (bullish, bearish or neutral) and a strength (strong, moderate, weak or unavailable). Conviction is then a function of how those six attributes line up.

Step 1 — Start from block agreement

If all three blocks point the same way (Macro, Sentiment and Seasonality all bullish, or all bearish), you start at High conviction. This is the cleanest possible setup.

If two of the three blocks agree on a direction, you start at Medium. The third block is either neutral or pointing the other way.

If there is no clear majority direction — for example, one bullish, one bearish, one neutral — you start at Low. The market is not telling a consistent story.

Step 2 — Strength can promote you to High

A Medium setup gets bumped up to High if the strongest block is genuinely strong and the other aligned block agrees with it. Specifically: if Macro is rated 'strong' and at least one of Sentiment or Seasonality points the same way, conviction is promoted to High.

It also gets promoted if two of the three blocks are both rated 'strong' and both point in the dominant direction. The logic is simple: when the loudest evidence in the room is unusually loud, partial agreement counts more.

Step 3 — Penalties that knock conviction down

A block that is moderate-or-stronger and points against the dominant direction reduces conviction by one level. So a Medium setup with a credible contradicting block drops to Low.

Low data coverage on the macro block (fewer than 40% of applicable indicators have a real value) also costs you a level. You cannot have High conviction on a half-empty dataset.

Each block that is completely unavailable (for example, no COT data on a currency CFTC does not cover) costs you another level. Missing data never gets faked — it gets penalised.

What 'strong' actually means in each block

Macro is 'strong' when the average per-indicator score is at or above 1.5 on a -2 to +2 scale, with at least 40% data coverage and no severe internal conflict. In plain English: most macro indicators are firmly leaning the same way.

Sentiment is 'strong' when either COT or retail positioning is at the maximum of its scale (for example, a COT reading in an extreme percentile, or a crowded retail position near 70%+ on one side), and the other side is not contradicting it.

Seasonality is 'strong' when the historical edge for the current window is large in absolute terms — roughly 0.8 or higher on the internal -1 to +1 scale.

A worked example

Say GBPUSD shows Macro bearish, Sentiment bearish, Seasonality neutral. Two blocks agree → you start at Medium.

Now the Macro block is rated 'strong' (most UK and US indicators are firmly favouring USD) and Sentiment is at least 'moderate' bearish. The promotion rule fires: Medium → High.

If instead the macro coverage was below 40%, the engine would refuse to promote and the setup would stay at Medium — even with the same headline direction.

Why conviction is separate from the score

The headline confluence score tells you how strong the directional pull is. Conviction tells you how much to trust that pull. A +6 score with Low conviction is the engine saying 'the math says bullish, but the evidence underneath is thin or contradicted'.

Disciplined traders weight position size by conviction, not by score size. The best setups are usually the ones where High conviction and a meaningful score appear together.

How to use it day to day

Filter the dashboard to High conviction first. That is your A-list. Only consider Medium when no High setups exist and you understand which block is missing.

Avoid Low conviction altogether for new positions. It is the engine's way of telling you the market is not paying you to take a view today.