Forge

Your strongest research candidates, with the evidence, potential levels and confirmation still needed. A ranked opportunity is only qualified when all checks pass.

Best opportunities now

Waiting for research

Loading recommendations…

Two daily research windows

Checking exchange calendar…

Research overview

—

Shown—

Saved data—

Research mode—

Early setups—

EARLY today—

READY today—

FADED—

INVALIDATED—

False + (4h)—

n / lead—

Waiting for first refresh…

TRADING REPS

0 COMPLETED

VISUAL PRACTICE

Historical candles only up to the practice timestamp. Future data stays hidden until you lock.

ADVANCED · MANUAL REP JOURNAL

TRADING SKILL SCORE

visual reps · sample-aware

PATTERN PERFORMANCE

descriptive only

WHAT AM I GETTING BETTER AT?

sample-aware

REP JOURNAL

0 reps

REAL-TIME STREAMING = NOT CONFIGURED

Loading scalp shortlist…

Market map

Research context, not a buy. Equal-weight and median returns. Stale quotes excluded. Shared names across themes are not independent confirmation.

Session

Red: negative return · Neutral: ±0.4% · Green: positive return · Grey: insufficient data

Loading rotation from cached quotes…

Cycle research (gated lag study — not a next-sector forecast)

Tests whether one group’s strength precedes another over 1 / 3 / 5 / 10 sessions. Holdout sample sizes and failed signals are required. Cached daily bars only.

Not run. This desk never publishes a “next sector” pick without evidence metrics.

Forge cached candidate tiles · subset, not sector strength

Filters & search

Metric

Size by

Sector

Minimum Forge Score

Recommendation State

Search ticker

Loading heatmap…

More research views

Explore intraday setups, the market map and saved outcomes.

Downloads

Download a sanitized desk pack. Manual only — nothing is uploaded.

Include paper performance summary

Captures current board, coverage, brief, evidence and runtime — no secrets, no orders.

How this desk works

A guide to research states, scores, setup plans and trading practice.

Feedback & improvements

Saved to your local project inbox for the next development review. Submitting feedback does not automatically change the app.

Calibration

Win rate—

Target—

Stopped—

Expired—

Outcome marks from the ledger.

Research Quality

Signals—

Trigger rate—

Target-first—

Stop-first—

Correct rejects—

Missed winners—

Buzz failures—

Avg R—

Avg lead time—

Top current issue: Insufficient data.

Measured cohorts only.

Cohorts: awaiting observations.

What can we learn? Insufficient data.

Paper engine · OPS controls

Risk limits and simulation status. Full trade review lives in History → Paper.

Start£10,000

Equity—

Cash—

Open P&L—

Realised P&L—

Return—

Open—

Closed—

Open risk—

Rejected—

Forward recommendation replay · no broker orders.

Waiting for an eligible recommendation whose entry zone is actually touched.

Strategy Lab

Loading separated backtest, shadow and active-paper evidence…

Promotion is manual and one stage at a time. No live orders.

Thresholds

Configured values load from /v1/admin/thresholds — they are not blank defaults.

Desk controls

Paper fills are simulated within Forge. Live orders and Telegram recommendation sends stay off.

Buzz Edge

Events—

Lead n—

Real buzz events only. Empty means none recorded — stats are not invented.

Vendor budget

Loading shared API limits…

Shared across Alpha / Tango / Forge. Cache hits do not spend.