Why sample size changes the strength of a pattern
A result observed three times and a result observed one hundred times should not carry the same weight. ASD reports sample size, uses minimum gates, separates training from chronological holdout evidence, and avoids treating small streaks as laws.
Practical interpretation
Recent windows remain useful for describing current form, but their volatility is disclosed and compared with a broader baseline.
What ASD avoids
Certainty language, unsupported causality, hidden missing data, and a recommendation created only because a price looks attractive.