Is the market structure worth following?
DAILY DECISION BRIEF
Today
What needs your attention now, what is developing, and what can wait.
MULTI-USER SUBSCRIPTION BETA · PAPER TRADING · SERVER-SIDE SECRETSAdvanced research workspace.These diagnostic, provider, audit, and validation tools are intentionally separated from the daily subscriber workflow. The canonical server decision still governs Radar, the journal, and paper execution.
Market trend
Loading completed-daily SPY, QQQ, and IWM regime data…
Provider and data health
Shows whether the model is using live or cached data, how much history was analyzed, and whether the newest bar is complete.
Provider——
Request transport—header preferred
API credits—reported by provider when available
Data freshness—latest completed bar
Analysis coverage—aligned / requested bars
Completed-bar policy——
Authoritative daily model
The same completed-daily-bar model drives the shared Radar, journal, and automated paper-trading eligibility. Browser research views may add detail, but they do not override this decision.
Thesis state——
Cycle state——
Entry state——
Execution layer——
Loading the canonical Cloud Run model…
Walk-forward exhaustion evidence—
Browser/server consistency—
Decision objective
Separates where the stock is, where the evidence points, and whether the current price is an acceptable entry.
Market state—trend, range, or transition
Directional forecast—scenario range, not a guaranteed target
Should enter now?——
Estimated time to entry—analog-based range
Preferred entry zone——
Entry trigger——
Payoff limit—worst entry price that preserves minimum R:R
Entry timing evidence—
Conditional outcome after trigger—
Plan expiration / invalidation—
Event risk—
Lower-timeframe entry confirmation—
Liquidity / gap risk—
—
Entry readiness and trigger checklist
Shows how close the setup is to becoming actionable and the exact conditions still missing.
Trigger proximity
—
—
Timing confidence——
Next best action——
Primary setup——
Alternative setup——
Completed-bar change watch
Compares the current locked model state with the previous completed bar for this ticker and interval.
Waiting for a completed-bar comparisonThe browser stores only a compact local state snapshot.
Forward scenario map
These estimates are overlapping evidence views, not mutually exclusive calibrated probabilities.
Continuation scenario——
Entry-opportunity scenario——
Invalidation scenario——
Opportunity queue
Batch-scans a small private watchlist and ranks symbols by entry readiness, evidence quality, and estimated time to trigger.
Browser watchlist scans use the secure Cloud Run proxy. Twelve Data requests are serialized to one uncached symbol per minute; use the Cloud opportunity queue for broad scanning.
| Rank | Symbol | State | Bias | Action | Readiness | Bars to entry | Entry zone | Evidence | |
|---|---|---|---|---|---|---|---|---|---|
| Run the scanner to build a ranked opportunity queue. | |||||||||
Past → Present → Cycle-adjusted future
Actual adjusted market bars with real dates.
LONG cycle remaining (green)
SHORT cycle remaining (red)
Bar height = cycle remaining; 100% fills the entire price plot
RSI(14) overlay · 70 overbought · 30 oversold
Cycle map
Estimated from confirmed pivots, leg duration, price distance, momentum, volatility, and volume.
Last pivot—
Move since pivot—
Next phase—
Benchmark edge since pivot—
Completed long periods——
Completed short periods——
Current period elapsed——
Estimated remaining range——
—
—
Indicator-family confirmation stack
Correlated indicators are grouped into trend, momentum, participation, relative-strength, risk-regime, and outcome-evidence families.
ADX / directional movement——
Chaikin money flow——
EMA structure——
NATR volatility regime——
Higher-timeframe proxy——
Historical analog outcomes——
—
Cycle-adjusted risk vs reward
—
Trade side—
Entry—
Structure-aware stop—
Scenario target zone—
Conservative reward : risk—
Technical-model probability—
Analog probability—
Conservative decision probability—
Calibration status—
Break-even—
Decision expected value—
Position size—
RSI influence—
Family confirmation gate—
Analog evidence—
Stop distance—
R:R quality—Mechanical payoff geometry only
Predictive confidence—Evidence quality, not win probability
Trade readiness—Final entry-readiness score
—
Model evidence
Independent model auditor
This module does not vote on direction. It checks data integrity, arithmetic consistency, sample sufficiency, look-ahead risk, calibration, and contradictions after the model runs.
Audit score—not yet evaluated
Analysis / displayed bars—indicator warm-up
Benchmark alignment—exact interval matching
Probability uncertainty—provisional outcome range
Provider indicator cross-check is optional and consumes four Twelve Data API credits. It verifies local RSI, ADX, ATR, and EMA20 calculations but does not replace them automatically.
Rolling-origin research test
Rebuilds signals only from information available at each historical decision point, enters on the next bar, uses non-overlapping horizons, and compares against a simple EMA50 baseline.
Model test trades—TAKE-qualified historical signals
Model average outcome——
EMA50 baseline——
Model Brier score—probability quality
Model max drawdown—in cumulative R
Evidence verdict——
Planned-entry triggers—wait-plan validation
Post-trigger average—conditional result after planned entry
Actual bars to trigger—median realized timing
Friction stress—selected execution cost applied
This is a leakage-controlled historical simulation, but it is still research on the same rule family used to design the app. It is not independent proof of future performance.
| Measure | Model | EMA50 baseline | Interpretation |
|---|---|---|---|
| Run the test after loading at least 300 completed analysis bars. | |||
Locked walk-forward validation log
Forecasts are stored locally with their original completed bar, entry, stop, target, horizon, and probability. Locked records are never revised by the model.
Locked forecasts0non-overlapping per ticker/timeframe
Resolved forecasts0target/stop outcomes
Probability calibrationUnavailablerequires 200 resolved forecasts
Brier score—lower is better; descriptive until sample is large
Locking a forecast creates an auditable local record. The app will not claim calibrated probability until at least 200 resolved, non-overlapping forecasts are available.
| Locked | Symbol | Side | Bar used | Entry | Stop | Target | Decision p | Outcome |
|---|---|---|---|---|---|---|---|---|
| No forecasts locked yet. | ||||||||
EdgeLens account
Sign in to your isolated EdgeLens workspace. Journals, signals, alerts, rules, execution history, and brokerage state are scoped to your account.
The existing internal account remains test-user. New customers can create a Starter account and upgrade from the Account & plan screen.