Performance intelligence
This is the monitoring and analysis layer behind three systematic strategies. It exists to make one claim checkable: that the figures published about a trading operation can be traced back to the exports they came from, and that nothing is presented as measured which was not.
Where data has not been exported yet, the platform says so in the place the figure would appear. That is the design, not a gap in it.
The rules this platform runs on
01
Every displayed number carries LIVE, PAPER, BACKTESTED or DEMONSTRATION. The component that renders figures requires the grade as an argument and throws during the build if it is missing, so an unlabelled number cannot reach a page.
02
Every row in the database references a file with a recorded sha256. The exports are committed to the repository and the database is rebuilt from them, so any published figure can be traced back to the bytes it came from and recomputed independently.
03
A paper-traded result and a backtested one are never summed, averaged or drawn as one continuous line. Aggregates group by grade at the database level, and any figure derived from mixed inputs inherits the weakest of them.
04
Stop distances, targets, thresholds, periods and magic numbers are used to compute results and are then excluded from output by an allow-list. R-multiples and excursions carry the same analytical information without revealing the rules that produced them.
Provenance vocabulary
The database constrains this list to exactly four values, so a fifth grade cannot be introduced by accident and no row can exist without one. Each badge states its grade in text as well as colour, and spells it out for screen readers.
A note on the account behind this platform
The trading account observed on this machine reports as a demo server. Prop-firm evaluation accounts, funded prop accounts and practice accounts are all served that way, and the server cannot distinguish them — so until that is confirmed in writing, results from it are graded Paper rather than LIVE. Paper is the accurate label for real execution against live prices with simulated capital, and it is the one a reviewer can check.
How data reaches this page
The terminal is read only: nothing in this project places, modifies or closes an order, and nothing writes to the MetaTrader directory. Data leaves it as an export, and that export is the artifact everything downstream is derived from.
The export is produced either by the terminal's own report function or by a history-reading script. Neither can trade.
Exports live in the repository with their hashes recorded. The database is git-ignored because it is derived — delete it and one command rebuilds it identically.
Figures are calculated when the site is built, from the database, by functions covered by unit tests. Nothing is transcribed by hand.
Data coverage
Stated before any analysis, because it determines what the rest of the platform can honestly show. A system with no ingested trades publishes no computed performance — its stage below is what the operator states, and it is labelled that way.
X12
0 trades ingested
Stage live is operator-stated. No ingested execution data corroborates it yet, so X12 publishes no computed performance.
V8MR
0 trades ingested
Stage paper forward is operator-stated. No ingested execution data corroborates it yet, so V8MR publishes no computed performance.
ORB
0 trades ingested
Stage validated is operator-stated. No ingested execution data corroborates it yet, so ORB publishes no computed performance.
These are counts of records in the database, not performance statistics — they describe the platform's own state and carry no claim about trading results.
Capabilities
The modules that do not depend on exported trade data were built first, so the platform's structure and integrity model can be reviewed before any performance figure exists. The rest are specified and sequenced; each names the phase that delivers it and what it will contain.
How data reaches this site, what each provenance grade means, and how much data exists per system.
Measured data freshness, ingest history with rejection counts, the source manifest, and the automation event log.
The nine stages a strategy passes through from idea to retirement, with each system at its stated stage.
The stack, the integrity model, the testing approach, and what was deliberately left out.
Portfolio equity and drawdown across all systems, split by provenance grade.
Will contain · Portfolio equity curve, an underwater drawdown chart on a shared x-axis, per-system metric tiles, and a guard summary strip.
Requires an exported trade history
One page per system: lifecycle stage, computed metrics, distributions.
Will contain · R-multiple distribution, MAE/MFE excursion scatter, exit-reason breakdown, per-month heatmap, and win/loss streaks.
Requires an exported trade history
Side-by-side system comparison with a correlation matrix.
Will contain · A sortable table carrying a grade column, plus a correlation matrix that states its overlap count on every cell.
Requires an exported trade history
Every trade, filterable, with a detail view per trade.
Will contain · Filters across system, grade, symbol, direction, outcome, exit reason, session and date, with a summary strip that recomputes from the filtered subset.
Requires an exported trade history
Observed drawdown, streak and tail statistics, each with a minimum-sample gate.
Will contain · Drawdown depth and duration analysis, loss-streak distribution, exposure concentration, and observed tail statistics.
Requires an exported trade history
Balance, equity, daily loss, drawdown and target limits with HEALTHY / WARNING / CRITICAL / HALTED states.
Will contain · A gauge per limit showing threshold, observed value and headroom, a consolidated worst-state banner, and a trading-days tracker.
Requires an exported trade history
Every account, its type, and which systems deploy on it.
Will contain · Account cards stating masked login, broker, server and — crucially — account type, so funded, evaluation and practice accounts are never conflated.
Requires an exported trade history
Configurable exploration of precomputed backtest results.
Will contain · Sanitized axis selectors over a precomputed result grid. Axis labels are opaque by construction, so the lab is configurable without revealing a single parameter.
Requires an exported trade history
Where the results came from, by system, symbol, session and more.
Will contain · Contribution bars across nine dimensions and a gross-to-net waterfall showing commission and swap drag.
Requires an exported trade history
Retrospective analysis of historical results only. No forecasts, no signals.
Will contain · Generated performance reviews, each published with the exact manifest of metrics and trades the model was shown, and a window that cannot extend past generation time.
Requires an exported trade history
A record of past threshold breaches. Never a recommendation.
Will contain · Active rules and the history of alerts raised, with observed values and resolution times.
Requires an exported trade history
The systems themselves — what each one trades, how it is exposed, and the research that validated it — are documented on the systems and research pages.