Signal Decay
Definition
Signal decay refers to the deterioration of an investment signal's effectiveness over time. As markets evolve, what once delivered a strong edge may lose its predictive power due to crowding, structural changes, or shifts in macro conditions. In quantitative and systematic investing, signal decay is a critical factor in maintaining long-term strategy performance.
A decaying signal may still show strong historical results (in backtests), but its real-world utility diminishes as the market adapts.
Why It Matters to Investors
- Signals with strong historical performance may underperform going forward
- Overused or widely adopted strategies become less effective (crowding)
- Structural shifts (e.g. monetary policy, market structure) can erode signal reliability
- Forces regular model review, revalidation, and potential replacement of inputs
- Helps avoid overfitting to past data and false confidence in stale strategies
The TiltFolio View
TiltFolio Adaptive is built on signals that are inherently resistant to decay. Rather than relying on fragile statistical edges, the system uses:
Market internals, such as how defensive stocks behave relative to cyclical ones, are difficult to game, deeply rooted in investor psychology, and less likely to decay because they reflect real-time risk-taking behavior. These signals are harder for large capital flows to neutralize or distort.
TiltFolio Balanced does not rely on signals that can decay. Instead, it maintains its diversified allocation (50% bonds, 30% stocks, 20% gold) regardless of market conditions, relying on strategic diversification rather than dynamic signals.
TiltFolio Adaptive prioritizes durability over complexity. Its signals are chosen for their intuitive link to actual investor behavior and their persistence across regimes. While performance is monitored for breakdowns, the system is designed to minimize reliance on signals prone to overfitting or rapid obsolescence.
Real-World Application
• Momentum signals can decay as more investors pile into the same trades
• Overfitted machine learning models perform poorly out-of-sample due to decay
• Some macroeconomic indicators lose power after regime shifts or policy changes
• Hedge funds regularly retire or replace signals that show declining outperformance