Continuation Play
The historical sample suggests the initial reaction tended to follow through across the measured horizons.
Methodology & Evidence
How the decision reads are built — and what they don't claim.
Intellucent turns raw event history into structured decision context — with scheduled economic releases as the flagship dataset today. This page documents where the data comes from, how it is classified, how it is validated against forward outcomes, and the limits you should hold it to. No black boxes, no performance promises.
01 · Data sources
What goes in.
The foundation is a historical record of scheduled economic releases and the market reactions around them, aligned to specific currency pairs and time horizons.
Economic event calendar
Scheduled economic releases across 10 event categories — CPI, NFP, PMI, retail sales, GDP, rate decisions and more.
Price reaction history
Post-release behaviour for the major pairs: EURUSD, USDJPY, GBPUSD and AUDUSD.
Horizon windows
Reactions are measured at fixed horizons — H1, H3 and H24 — so comparisons are like-for-like.
Event pairing
Each release is paired with an instrument to form an event-pair situation, the atomic unit of analysis.
02 · Classification
How raw reactions become a decision read.
The Decision Engine groups each new event into the set of historically similar situations, then summarises how those situations behaved into one of four decision states.
-
1Find similar situations
Match the event-pair to comparable historical releases by category, surprise and context.
-
2Measure follow-through
Summarise how those situations moved across H1, H3 and H24 horizons.
-
3Assess trap risk
Quantify how often the initial reaction historically reversed.
-
4Assign a decision state
Combine edge, follow-through and trap risk into one of four labelled states.
03 · Validation
Historical edge, checked against forward data.
A pattern is only useful if it holds up out of sample. The Validation Engine compares the historical edge to how events behaved in a forward window, so the read reflects what actually happened next — not just a curve fitted to the past.
- Historical edge is computed on the historical sample.
- A 2026 forward window is used to validate that the edge persisted.
- Where forward behaviour diverges, evidence quality is lowered.
~80.7%
continuation rate observed in high-edge situations
~16%
trap rate — reactions that reversed
2026
forward validation window
233
similar situations behind a typical read
04 · Edge calibration
What an edge score means.
Edge scores map to evidence quality bands. A higher score reflects a larger, more consistent historical sample — not a prediction or a guarantee.
| Edge score | Evidence quality | Interpretation |
|---|---|---|
| 70–100 | High | Large, consistent historical sample with clear follow-through tendency. |
| 40–69 | Mixed | Some historical signal, but with conflicting or thinner evidence. |
| 0–39 | Low | Weak or noisy history — typically a No Trade / Stand Aside read. |
05 · Coverage
The size of the evidence base.
96,925
historical events reviewed
294,710
event-pair observations
13,655
event-pair situations
1,361
selected decision contexts
10
event categories
4
currency pairs
Update frequency: the decision context is refreshed as new releases are reviewed and the forward-validation window advances.
06 · Decision states
Four states, defined.
Wait for Confirmation
Evidence is mixed or timing-sensitive; the history favours waiting for the move to confirm.
No Trade / Stand Aside
Weak or noisy history with no reliable edge — the read is to stand aside.
Trap Risk
The initial reaction historically reversed often enough that chasing it carried elevated trap risk.
07 · Engines
Six engines behind every read.
Market Memory Engine
Surfaces the historically similar situations behind an event.
Decision Engine
Assigns the four-state decision read.
Risk Engine
Quantifies trap risk and weak-expression awareness.
Replay Engine
Reconstructs how similar releases actually behaved.
Validation Engine
Checks historical edge against forward data.
Event Intelligence Engine
Maintains the market-memory record across reviewed events.
08 · Limitations
What the data does not capture.
Being explicit about the limits is part of the methodology. Intellucent is historical context — not a forecast, and not a complete picture of the market.
Real-time sentiment
Live positioning shifts and crowd sentiment are not in the historical record.
Order-book liquidity
Intraday liquidity and depth conditions are not modelled.
Positioning data
Aggregate trader positioning around an event is outside scope.
Unscheduled shocks
Geopolitical or off-calendar shocks are not anticipated by the engines.
Intellucent provides historical market context and decision support. It does not provide investment advice, trade signals, portfolio management or execution services. Historical behaviour does not guarantee future outcomes. Built and operated by DeltaCore.
See the evidence in the product.
Every decision read links back to the historical situations behind it.
Intellucent provides historical market context and decision support. It does not provide investment advice, trade signals, portfolio management or execution services. Historical behaviour does not guarantee future outcomes.