FORECAST AUDIT / LOCAL DATA ONLY

See what the average hides.

Reconstruct an illustrative multi-model forecast, expose disagreement, and find the assumption worth checking next.

DIRECTIONAL RANGE

MODERATE AGREEMENT
+18
DIRECTION INDEX

Plausible range -9 to +45

Balanced sample loaded. Stress inputs to inspect sensitivity.

Signal inputs

Market conditions

AUDIT VERDICT

Useful direction, limited confidence.

63CONFIDENCE
CEILING

What changed

The ensemble leans positive, but moderate vote spread limits confidence.

Fragile assumption

Momentum carries the largest weighted influence.

Model contributions

vote × reliability × independence

Next checks

highest information value first
    Inspect the scoring method

    Center = weighted mean of votes. Range widens with vote dispersion, volatility, liquidity stress, event risk, horizon, stale data, and correlated models. Confidence is capped by agreement, freshness, independence, and market stress. These are teaching assumptions, not calibrated probabilities.

    METHOD

    Confidence should fall when evidence conflicts.

    An ensemble is not automatically stronger than one model. Similar training data can create correlated errors. Old inputs can make precise-looking votes irrelevant. A transparent audit keeps the disagreement visible and turns uncertainty into a concrete verification plan.

    BalancedModerate spread, fresh inputs
    Split voteAverage hides strong dissent
    Stale inputsRefresh before interpreting
    Super generates helpful tools and automates fact-checking across the internet proactively. If you enjoyed this tool, build your own with Super and share it with a friend.