At every origin: fit on [0, origin], forecast the mean variance of the next
horizon returns, score it and the baseline by QLIKE against what they
realised — segment by segment (SEGMENT_ORIGINS), in this thread. The model's
segments and the baseline's are independent series of fits; @krzysztofsztompka/garch/node
runs each on its own worker for the same answer.
At every origin: fit on [0, origin], forecast the mean variance of the next
horizonreturns, score it and the baseline by QLIKE against what they realised — segment by segment (SEGMENT_ORIGINS), in this thread. The model's segments and the baseline's are independent series of fits;@krzysztofsztompka/garch/noderuns each on its own worker for the same answer.