Signal Decorrelation
Many weak INDEPENDENT bets beat a few correlated ones. Which of your scanners actually diversify — and which just duplicate each other?
Why it matters:a scanner's win-rate tells you nothing about whether it addsto the book. Two signals that fire on the same names and lose in the same weeks aren't two bets — they're one bet held twice. This measures the correlation of realised return streams and reweights toward genuine diversification. Research, not advice; correlations on thin data are noisy and shift as trades resolve.