The lazy story
A common framing is that signals decay because the market gets smarter and forecast accuracy fades. We have not seen that as the dominant pattern. In our book, the forecast quality holds longer than people expect.
What actually goes first
Execution cost rises before forecast accuracy falls. As more participants size into the same window, slippage and fill quality degrade. The signal still predicts direction; the trip to capture that prediction just gets more expensive.
Implication for live monitoring
We track the realised cost-to-capture ratio per signal, not just hit rate. A signal whose hit rate is intact but whose cost-to-capture has doubled is on the way out.