Tag: alpha-research
Concepts
- Alpha Correlation and Effective Breadth
- Estimating Alpha Half-Life
- Alpha That Is Really Beta in Disguise
- Blending Alpha Signals with Machine Learning
- The Break-Even Transaction Cost of a Signal
- Estimating the Capacity of an Alpha
- Choosing the Target: Raw, Excess or Residual Returns
- Choosing the Prediction Target: Return, Rank or Direction
- Choosing What to Forecast
- Combining Many Weak Signals
- Conditional IC by Volatility and Liquidity Buckets
- Conditioning a Signal on Liquidity
- Conditioning a Signal on the Volatility Regime
- Putting Trading Costs Into the Training Objective
- Designing a Signal With Costs In From the Start
- Decomposing Where a Signal's Turnover Comes From
- Document Embeddings as Alpha Features
- Equal Weights vs Optimised Blends
- Designing a Cross-Sectional Feature Set
- Framing Alpha Research as a Prediction Problem
- From Raw Data Field to Tradeable Signal
- Blending by Family, Then Across Families
- IC Decay Curves and Signal Half-Life
- Where Alpha Ideas Come From
- Informational vs Analytical Edge
- Measuring a Signal's Decay Curve
- Monitoring Live Alpha Model Decay
- N Versus T Tradeoffs in Panel Signal Tests
- Nonlinear Feature Interactions That Actually Pay
- Warning Signs of an Overfit ML Strategy
- Overfitting the Cross-Section
- Neutralising a Model's Output Against Risk Factors
- Rank IC Versus Pearson IC
- Raw, Excess or Residual Return as the Target
- Regime-Conditional ML Models
- Researching Intraday vs Multi-Day Alpha
- Sequence Models for Multi-Horizon Return Forecasts
- Short-Horizon ML Alpha Under Cost Constraints
- Shrinkage of Noisy IC Estimates
- Signal Autocorrelation vs Return Autocorrelation
- Structural vs Behavioural Edge
- Subsample Stability Across Universe Slices
- Supply-Chain Graphs for Alpha
- Testing Whether Two Signals Have Different IC
- The Capacity-Sharpe Frontier
- Turnover Control in ML Strategies
- When an ML Signal Becomes a Strategy
- Who Is On the Other Side of the Trade?