Every trader believes more data makes his estimate safer. It does not. If the market changed, a million observations can make the wrong answer feel certain.
his name is Joe Blitzstein. Harvard statistics professor, and he put the two most famous theorems in probability on a blackboard in a free lecture.
the law of large numbers says an average moves toward the truth as the sample grows. one clause hides underneath: the observations must be independent and identically distributed.
markets promise you neither. volatility shifts. correlations break. regimes turn. mix the old world with the new and the number gets smoother while the estimate gets worse.
the central limit theorem shows the price of precision. half the noise needs four times the data. one tenth needs one hundred times more. certainty gets expensive fast.
that is every backtest polished with another decade of history. more rows did not prove the edge. they made the old regime harder to see.
the edge was never having more data. it is knowing whether the data still comes from the same world.
$BTC Yesterday's trade plan canceled. Bitcoin first decided to take that $661.k level so that makes the bounce probably strong.
So what we do is wait for fresh plans.