for one, the author scaled their mean-OLHC values _prior_ to making their train-test split, meaning data from the test set has leaked into the training set. Tim

for one, the author scaled their mean-OLHC values _prior_ to making their train-test split, meaning data from the test set has leaked into the training set. Time series prediction is a bit easier if you understand how the future values are distributed!