METHOD · VALIDATION SNAPSHOT

How Jodie turns a market move into evidence you can inspect.

Jodie combines observed market behaviour with relationships extracted from company filings. It shows what is moving together, why those companies may be connected and when that reaches a name you follow.

THE PRODUCT LOOP

Two maps, joined around your companies.

Neither market behaviour nor filing data is enough on its own. The useful product is the connection between them.

01

Observe the market

Jodie removes the broad market move, then looks for companies whose remaining behaviour repeatedly converges. Established groups and early groups are kept separate.

You see a group marked Confirmed or Watching—not a generic movers list.
02

Read the business structure

Company filings are turned into traceable relationships: customers, suppliers, partners, competitors and companies whose annual reports use similar operating language.

You see the relationship, filing date and original excerpt.
03

Join the two maps

Every observed group is checked against the filing map and the companies you follow. That is how a market-wide event becomes relevant to your own list.

You see when a group reaches one of your names or a connected business.

WHAT THE LABELS MEAN

Confidence is visible, not hidden in a score.

Confirmed
An established co-movement group that cleared Jodie’s evidence threshold.
Watching
An early group that may fade. It is deliberately quieter and never presented as established.
Action
Trading activity is unusually concentrated. Direction remains unknown.
Filing receipt
The source excerpt and filing date behind a business relationship or operating claim.

THE VALIDATION LEDGER

The measurements behind the product language.

These are evaluation snapshots, not a performance record. Each figure is paired with the conclusion it supports and the boundary it does not cross.

THE LIVE MAP52%forward co-movement precision

Confirmed groups continued to behave as groups more often than the previous detector.

In the comparable evaluation window, 52% of 50 confirmed episodes retained the measured group behaviour. The earlier detector measured 30% across 4,025 episodes.

Boundary: This measures group persistence—not whether members rose, fell or produced a return.

THE STRUCTURAL MAP42.9%versus a 30% baseline

Companies with unusually similar filings moved together more often.

The filing-similarity gate reached a 42.9% co-movement rate against the 30% comparison rate. The measured difference was statistically distinguishable from noise (p = 0.0015).

Boundary: Similar language is supporting evidence of shared exposure, not proof of causation.

THE JOIN+0.046sector-adjusted difference

Filing-linked companies moved with active groups beyond sector alone.

The read-through study found a positive co-movement difference after controlling for shared sector membership. Its 95% interval was +0.017 to +0.075.

Boundary: The result supports connected-market context; it does not identify a leader or predict direction.

ACTIVITY, NOT DIRECTION1.25–1.33×near-term realised movement

Concentrated activity preceded more movement, but not a reliable direction.

Across 14,767 evaluated events, higher-intensity observations were followed by elevated realised movement over the measured horizon.

Boundary: Up versus down remained approximately coin-flip. Jodie therefore describes action, never a directional call.

VALIDATION SNAPSHOT · JULY 2026 · DEFINITIONS AND WINDOWS HELD CONSTANT WITHIN EACH COMPARISON

EVIDENCE CHAIN

Where a Jodie explanation comes from.

MARKET DATAObserved price and volume behaviour

Adjusted for broad market movement before companies are grouped.

+
PUBLIC FILINGSDisclosed relationships and operating language

Stored with the source document, date and supporting excerpt.

JODIEA connected event with receipts

Relevant to a group, company or name that the user follows.

KNOWN LIMITS

What can make the picture incomplete.

Not a newswire

A single company may move first on a headline. Jodie’s job is to identify the wider group and connect it to disclosed business structure.

Less reliable after shocks

For roughly five trading days after volatility or credit shocks, group behaviour weakens. Jodie suppresses weaker observations and shows less.

Relationships are bounded by disclosure

A filing map reflects what companies disclose and what can be verified. Private, omitted or outdated relationships may not appear.

THE HARD BOUNDARY

No forecasts. No buy or sell calls. No price targets.

Concentrated activity and connected movement are context. They are never presented as a recommendation or a promise of investment performance.