Why a trader cares
A model is most useful when it explains its horizon, evidence coverage, drivers, and prior outcomes—and admits when the setup is not strong enough.
PlatformModel Signal
Statistical contextLive
Statistical confirmation without forcing a trade.
Enerlytics Model Signals can return Bullish, Bearish, Mixed, or No Edge. Abstention is intentional when coverage, agreement, or reliability is insufficient.
Question answered
Does the statistical evidence confirm the setup, contradict it, or show no defensible edge?
Built for more than one level of market experience
Start simple. Keep the professional context.
Every screen preserves the instrument, period, unit, source, and limitation while explaining why the evidence matters.
Start here01
Identify the observation before interpreting it.
Read the instrument, unit, timestamp, period, and source first. A weekly stock estimate and a five-minute futures quote answer different questions.
Desk read02
Compare level, change, and historical context.
The current value matters less without its direction of travel, seasonal baseline, related markets, and the catalyst that moved expectations.
Professional use03
Demand cross-source confirmation and preserve uncertainty.
Treat every dataset as evidence with a release lag, methodology, revision risk, and known blind spots. The workflow should be reproducible after the fact.
Component guide
What each component means—and how a desk reads it.
Definitions stay plain enough to learn from, while the desk read preserves the mechanism, cadence, and source a professional user expects.
Horizon-specific state
In plain English
Each configured horizon is evaluated separately instead of stretching one forecast across every holding period.
Professional read
A one-day directional lean can coexist with a mixed five-day outlook; disagreement is information, not a formatting problem.
- Cadence
- Scheduled official inference plus preview passes
- Source
- Enerlytics feature and inference pipeline
Feature coverage
In plain English
Coverage reports how much of the expected evidence was present at inference time.
Professional read
High model confidence with weak input coverage deserves less trust than a fully observed setup with consistent drivers.
- Cadence
- Recorded with every inference
- Source
- Point-in-time feature snapshots
Driver diagnostics
In plain English
The screen exposes the evidence families most responsible for the current state without publishing a proprietary recipe.
Professional read
Use drivers to test whether the call is supported by curve, inventories, flows, news, or merely correlated price behavior.
- Cadence
- Recorded with every inference
- Source
- Model diagnostics
Resolved outcomes
In plain English
Predictions remain pending until the full forward path exists, then record realized return, direction, and miss reason.
Professional read
Recent hit rate is an audit statistic—not the probability that today's call will succeed.
- Cadence
- Daily outcome resolution
- Source
- Official inference and settlement history
No Edge
In plain English
The model may abstain when evidence coverage, horizon agreement, or reliability is insufficient.
Professional read
Abstention reduces forced directional calls and is often the most honest state during regime shifts or sparse data.
- Cadence
- Every inference
- Source
- Deterministic decision policy
What data it uses
The evidence underneath the screen.
- Price and return features
- M1-M2 and M1-M6 curve structure
- Inventory and physical-flow evidence
- Market-impact news features
- Horizon-specific resolved outcomes
How Enerlytics interprets it
Method before conclusion.
- 01
Shows signal state separately for configured horizons
- 02
Exposes drivers, quality, and reliability adjustments
- 03
Keeps recent hits, misses, and unresolved predictions auditable
How it connects
Evidence moves through a workflow.
- 01Feature coverage
- 02Model state
- 03Reliability context
- 04Physical confirmation
- 05Outcome review
What not to infer
Historical hit rate is not today’s probability of price increasing. A model state is context, not individualized investment advice.
Worked market workflow
Use a model as a challenged witness
The signal earns weight through coverage, stable drivers, and realized accountability—not through a confident label alone.
- 1
Read the state
Bullish, Bearish, Mixed, or No Edge is the output, not the entire explanation.
- 2
Inspect coverage
Check whether the expected evidence was actually present at inference time.
- 3
Challenge drivers
Open the curve, EIA, flow, and news screens behind the strongest contributors.
- 4
Review history
Separate pending observations from resolved outcomes and inspect misses as closely as hits.
Methodology and limitations
Trust comes from showing the seams.
Point-in-time discipline
Official records preserve the feature date, prediction start date, horizon, model version, and evidence available at the time.
Reliability is empirical
Rolling performance and calibration are shown as historical context; they are not converted into a promise about the next observation.
Verified production coverage
Coverage reviewed against the production data store on October 2, 2026. Live totals continue to grow.
- The production audit contains more than fourteen thousand recorded signal and outcome observations.
- WTI, Brent, Henry Hub, and selected energy-market instruments retain horizon-specific audit history.
Primary references
Enerlytics links the underlying methodology so customers can distinguish a product interpretation from the source definition.
See the product
A real Enerlytics workflow—not a conceptual mockup.
Weekly research, recaps, and video explainers
See the evidence used in a real market conversation.
Enerlytics publishes weekly market recaps, component explainers, and thesis walkthroughs. For Model Signal, the recurring desk question is: Did the model change because the market changed, or because one evidence family went missing or reversed?
Weekly market notes →Written recaps tied to the same evidence workflow.Public signal reviews →Current states with accountability and limitations.Video recaps on X →Short market recaps, chart walkthroughs, and thesis updates.Explainers on Threads →Component education and the reasoning behind each market read.
Live in Enerlytics
Put this evidence inside the full decision workflow.
Start with a free trial. Review the data, methodology, related evidence, and limitations before making your own market decision.
Start free trial