API keys and database credentials stay on the server.
MARKET RESEARCH / LIVE MONITOR
Gold scalping lab
--:-- UTC
Research modeCandidate logic is unvalidated. This app does not place trades or guarantee profitable signals.
Latest XAU/USD close—Waiting for market dataData healthCheckingAge unknownLocal API request count—App counter, not official provider usageCandidate stateNot evaluatedAwaiting completed candles
PRICE ACTION
Recent 5-minute candles
XAU/USD · UTC
Waiting for validated market data…
Candles load from stored market data · UTC
Bull candle Bear candle
STRATEGY RESEARCH
Latest candidate evaluation
Pending
System will evaluate completed candles only and fail closed when data is stale or incomplete.
A matching condition is a research candidate, not a validated entry. No order is ever placed.
SYSTEM ACTIVITY
Recent events
No events yet.
AUDIT TRAIL
Candidate log
Research only
No candidate records.
HISTORICAL VALIDATION
Research backtest
No live orders
Runs the selected research hypothesis against stored 5-minute candles. The selected unified hybrid is historical research only: BUY and SELL allowed, compression ATR 1.8, breakout body ratio ≥0.60, close-position ≥0.65, exclude 12:00–17:59 UTC, initial stop distance $4–under $8, and a 2R target by default. This is an unvalidated research hypothesis; all results are historical research only. Entry uses the next candle open and a signal from the prior closed candle. Trades close only at TP or SL; there is no time-based exit. If both levels are touched in one candle, SL is assumed first. Spread and slippage are estimates, not verified Exness fills. Gross R shows the TP/SL outcome before estimated costs; net R subtracts costs converted using each trade's initial risk.
Run a backtest to view results. Recent live-monitor candles may be too few for reliable conclusions.
Controlled filter experiment appears after a backtest.
Matched Candidate A/B/C comparison appears after a backtest.
Forward testing is intentionally disabled for this project. This page runs historical backtests on candles already stored in the database; it does not request new data for testing.
Historical robustness scan compares entry thresholds and target R using stored candles only. It does not change live candidate detection or request new market data.
DATA QUALITY
Historical data audit
This report audits stored candles, timestamp alignment, observed gaps, and whether stored 15-minute OHLC values match three complete constituent 5-minute candles. Gap estimates assume continuous trading and may count normal closures.
Loading audit…
HISTORICAL IMPORT
Backfill older candles
One week per request
Choose a UTC date range. The app fetches one calendar week per click, stores valid 5-minute candles, and derives complete 15-minute candles. Requests are rate-limited and counted against the daily safety cap. You can return later and resume the saved job.
No backfill job configured.
No historical requests will run until you press the button.
STORED DATA ONLY
Mobile candle inspector
No provider requests
Inspect already-stored Twelve Data candles. Tap a candle to read its exact timestamp and OHLC values. Times are displayed in UTC; this screen does not fetch new market data.
No candles loaded.
On a phone, tap a candle body or wick. If candles are crowded, choose 100 candles or filter to a single UTC date.
Selected candle details will appear here.
Data-source limitation: these are stored Twelve Data values, not Exness values. This inspector cannot prove source completeness or broker-price equivalence.