Technical Indicator
A combination of a fast-reacting moving average (e.g., 5-10 period EMA) crossing above a medium average (20-50 period EMA), explicitly filtered by a higher-timeframe regime trend or ADX threshold. The filter ensures entries only happen in the direction of the dominant macro trend.
Highly susceptible to parameter instability and repeated whipsaw losses in low-volatility, ranging markets.
Fast Moving-Average Crossover with Regime Filter
Technical Indicator
Breakout strategies that buy a new high of a defined n-period range (Donchian) or a close beyond an ATR-based channel (Keltner). Frequently paired with a volatility contraction squeeze to filter for explosive moves.
Suffers from clusters of small losses during false breaks and highly volatile chopped ranges.
Donchian and Keltner Volatility Breakouts
Technical Indicator
Utilizes the MACD histogram trough followed by two rising bars or a transition above zero as a trigger, typically combined with independent trend-strength filters like RSI > 50 or price > 200 SMA.
Prone to severe whipsawing in consolidating markets if traded without a longer-term trend alignment or volatility filter.
MACD Histogram Acceleration
ML/AI Model
An AI architecture that fuses a Gaussian-Process changepoint-detection module into an LSTM Deep Momentum Network. It computes online probabilities of regime breaks (severity and location) as dynamic features.
Computationally intensive for live streaming; heavily dependent on historical regime analogies.
Gaussian-Process Changepoint Deep Momentum Network
ML/AI Model
A decoder-only Temporal Fusion Transformer incorporating a Variable Selection Network and LSTM, using volatility-normalized returns across multiple time horizons (daily to annual) to dynamically size momentum.
Significant risk of backtest overfitting on non-stationary financial series; opaque decision logic.
Momentum Transformer
ML/AI Model
A deep neural network combining convolutional blocks, inception modules, and an LSTM to model the last 100 states of the Level 2 Limit Order Book (top 10 bids/asks volume and price).
Requires low-latency, sub-second HFT infrastructure; edge is heavily eroded by exchange taker fees and spread slippage.
DeepLOB CNN-LSTM Order Book Classifier
Volume/Microstructure Signal
Measures aggressive buy-minus-sell executed volume (Cumulative Volume Delta) and order flow imbalance across venues, filtering out passive, spoofed, or resting limit orders.
Decays rapidly (seconds to days); signals can be negated if passive resting limit walls absorb all aggressive taker flow.
Executed Order Flow and CVD Imbalance
Crypto On-Chain/Derivatives Signal
On-chain metric analyzing balance increases in high-net-worth non-exchange wallets (e.g., >$1M or >50 BTC). It clusters addresses by entity and isolates real accumulation from exchange reshuffling.
Heavily reliant on accurate address clustering/labeling; delayed confirmation of OTC deals; wallet splitting can create false negatives.
Entity-Adjusted Whale Accumulation
Crypto On-Chain/Derivatives Signal
Monitors the net flow of stablecoins (USDT, USDC) into identified exchange deposit wallets, acting as a real-time proxy for sidelined fiat entering the market.
Short-lived predictive horizon; deposited stablecoins may be designated for yield staking, derivatives collateral, or margin, not necessarily aggressive spot buying.
Stablecoin Exchange Net Inflows
Forex Macro/Positioning Signal
Sorts currencies based on relative yield curve curvature (e.g., 2-year vs 10-year differentials) between nations to capture expected trajectories in monetary policy rather than simple nominal rate spreads.
Performance varies dramatically across central bank regimes (e.g., quantitative easing vs tightening cycles).
Yield Curve Curvature (Curvy Trade)
Forex Macro/Positioning Signal
Utilizes the term structure of option-implied standard deviation, skewness (25-delta risk reversals), and kurtosis derived from broker quotes via the Vanna-Volga method.
Mapping implied skew to expected returns is highly regime-dependent and prone to structural breaks triggered by central bank interventions.
Options-Implied Moments Term Structure
Multi-Factor Ensemble System
A volatility-scaled architecture that aggregates 1-month, 3-month, and 12-month time-series-momentum models, inverse-volatility weighting positions across a broad universe of assets.
Relies heavily on statistical diversification across uncorrelated assets; highly correlated crypto pools reduce its effectiveness.
Multi-Horizon Trend Ensemble
Multi-Factor Ensemble System
A hybrid framework that equal-weights fundamental macro trends (inflation, growth forecasts, real carry) around theoretically neutral baselines, subsequently enhanced by technical market price confirmation.
Susceptible to look-ahead bias in backtests if unrevised, publication-vintage macro data is not strictly used.
Macro-Trend Conceptual Parity System
Multi-Factor Ensemble System
A layered portfolio approach synthesized from top quant funds (Man AHL, Aspect, Graham). It opens a 'probe' size (20-35% risk) on a fast trigger (order flow, fast EMA), adding size iteratively as independent fundamental, ML, or slow-trend families turn positive.
Execution heavy; demands rigorous mathematical constraint of correlation weights to prevent multiple variations of a price signal from triggering a full allocation.
Staged-Confirmation Multi-Factor Architecture
Ask a followup
Sign in to run · $20 free credit, no card · every claim cited