[AUTO_INGESTION][PATTERN_TAXONOMY][PRE_TRADE_GATING][4_FILTER_MODES]

Adaptive Failure Learning Engine

Self-learning failure ingestion, pattern classification taxonomy, and pre-trade memory gating for automated capital defense.

Continuous Failure Learning Lifecycle

Whenever a paper trade closes with a realized loss, Zenth analyzes the context leading to the loss, extracts the root pattern, and writes a synthesized rule into active memory.

CLASSIFIER TAGS
Taxonomy Engine
Auto Root-Cause Analysis
RULE SYNTHESIS
Instant DB Sync
adaptive_learnings
PRE-TRADE GATE
AdaptiveFilter
Blocks High-Risk Traps
TRIGGER AUDIT
Count & Timestamps
Rule Effectiveness Metric

Failure Pattern Taxonomy

Common market traps classified and filtered by the learning engine:

[LOW_VOL_WHIPSAW]

Golden crossover on volume < 80% of 20-period average, causing immediate false breakout whipsaw.

[HIGH_RSI_EXHAUSTION]

Golden crossover with RSI > 70, resulting in overbought momentum reversal before target.

[CHOPPY_SIDEWAYS_TRAP]

Oscillating MA crossover within tight 0.3% price corridor across 10 consecutive candles.

[MACRO_RESISTANCE_REJECTION]

Entry near 24h high with negative taker volume divergence.

The 4 Adaptive Filtering Modes

Control how aggressively active learned failure rules gate incoming trade signals:

ModeExecution Action on Pattern MatchRecommended Purpose
STRICT (Default)Instantly converts [BUY] signal to [SKIP]Production paper trading with maximum capital protection
REPEAT_LOSSESBlocks signal ONLY after pattern caused >= 2 lossesAllows single anomaly while filtering repeating traps
DRY_RUNLogs match telemetry but allows paper trade to executeObserving pattern impact without intervening
DISABLEDDisables memory filter completelyRaw baseline strategy backtesting & benchmarking
Last verified: August 2026Maintained by Zenth Core