L2Azimuth

DEX Array identifies the moment a trader places a high-contract size limit order within the bid/ask spread, indicating a clear signal of directional conviction. The scanner validates whether that intent translates into real price movement, filtering out noise through multi-layered confirmation: no spoofing contamination, high-probability directional forecasting, and aligned volatility signals. Trades execute when all conditions align, capturing momentum from informed aggression before it fully materialises.
Detection triggers when an order price satisfies P(bid) < P(order) < P(ask), indicating aggressive placement between the bid and ask. The scanner then verifies whether subsequent bid/ask updates absorb this order and drive directional movement. Three validation gates must align: Spoofing Score below 0.4, Microprobability Forecast exceeding 0.82 in the same direction, and the proprietary HF Omega metric returning positive. Only when all conditions hold does the system execute, ensuring each trade captures genuine aggressive intent rather than transient noise.

Trespass reads structural tension in the order book by aggregating bid-ask depth across multiple levels, isolating authentic liquidity shifts from transient noise and quote stuffing. High-frequency execution meets precision filtering: trades trigger only when imbalance strength, directional probability, and spoofing absence all align. Trespass delivers consistent, volume-driven alpha across changing market conditions, handling execution latency effectively while adapting to different market regimes.
The core mechanism employs tick stacking, aggregating volume across L levels (typically 3-5) on both sides to compute the imbalance ratio (total buying/selling pressure divided by total volume). This raw signal gives us the relative normalised imbalance ratio, which then passes through proprietary noise filtering to remove quote stuffing and transient pressure. Trades execute when the filtered imbalance breaches configurable thresholds (typically >0.5), Microprobability score exceeds 0.8 in the same direction, and Spoofing probability remains under 0.2. Exit logic is risk-profile dependent, closing positions early if market conditions deteriorate during the execution of a trade.

A machine learning-based forecasting model that predicts the next 2 ticks of price movement using top levels of order book data. Included Exclusively in Forecast.