Qoala moolj applies stochastic modeling to latency-optimized market data, producing decision-support signals with a verifiable, daily-published performance record. Built for traders who want the reasoning behind every recommendation, not just the recommendation itself.
Illustrative predictive variance band plotted against realised intraday price action, refreshed on each model cycle.
The core engine ingests tick-level market data through a latency-optimized pipeline, applies a stochastic model to estimate short-term predictive variance, and outputs a ranked signal set before the next quote cycle completes.
Model parameters are re-fitted on a rolling basis rather than left static, which keeps the output responsive to changing volatility regimes without requiring manual recalibration by the trader.
Tick-level ingestion
Normalisation, cleansing
Predictive variance estimate
Ranked, confidence-scored
Each trading session produces a Daily Transparency Report covering the signals issued, the model's stated confidence at issue time, and the outcome once the position window closed. Reports are retained in an audit trail that can be reviewed session by session, not summarised into a single headline figure.
| Field | Description |
|---|---|
| Session ID | AEST date stamp, model version |
| Signals issued | Count and instrument breakdown |
| Confidence band | Model-stated variance range at issue |
| Realised outcome | Logged against closing reference price |
| Feed latency | Median ms for the session |
Illustrative report structure shown for layout purposes only. Field values are not actual performance data and do not represent guaranteed or expected results.
Risk controls are built into the model rather than applied afterward. Three mechanisms work together to manage exposure as conditions change intraday.
Position size recommendations are derived from a rolling drawdown estimate, so exposure contracts automatically when the model's recent predictive variance widens beyond its typical range.
Signal thresholds shift with detected volatility regime changes, reducing signal frequency in choppy conditions rather than issuing recommendations at a constant rate regardless of market state.
Before a signal is surfaced, correlation against currently open positions is checked, limiting concentration in instruments that tend to move together during the same session.
Qoala moolj is positioned as a decision-support layer for traders who retain full control over execution. The platform's role is to process volumes of market data that would be impractical to review manually, and to present the resulting analysis in a format that can be checked against the trader's own read of the market.
The methodology is documented rather than presented as a black box. Data sources, model versioning, and report retention policies are available for review by any active account holder.
Raw tick and quote data is pulled from licensed market data providers and normalised into a consistent schema before entering the model, with feed integrity checks run continuously.
Each model cycle is validated against out-of-sample historical data before its output is surfaced live, and any cycle that fails validation thresholds is withheld rather than published.
Validated output is formatted as a ranked recommendation with confidence band and suggested position sizing, delivered to the dashboard and, where enabled, via API for review.
Signal output is available via a REST endpoint that returns the ranked signal set, confidence band, and timestamp in JSON. Integration typically involves polling the endpoint or subscribing to a webhook, depending on the account tier and existing execution stack.
Latency depends on the underlying feed and instrument class, but the pipeline is designed to deliver signal output within the same quote cycle for supported instruments. Latency figures for the prior session are included in each Daily Transparency Report.
Market data is sourced from licensed providers covering ASX equities, index futures, and major FX pairs. Feed status, including any interruptions, is logged in the audit trail and disclosed in the relevant daily report.
Incoming data is checked for gaps, duplication, and timestamp consistency before it reaches the model. Sessions with feed anomalies are flagged in the report rather than silently smoothed over.
Access is structured around account tiers that determine instrument coverage, API call limits, and historical report retention. Full tier details and current terms are provided during account setup.
No. Qoala moolj produces decision-support signals and analysis. Execution decisions and order placement remain with the account holder or their existing execution system.