Neat. For the 25%-35% of the time that it's wrong, how much of that is due to th...

		eglover on Nov 4, 2014 \| parent \| context \| favorite \| on: Show HN: A neural net for predicting Bitcoin price... Neat. For the 25%-35% of the time that it's wrong, how much of that is due to the occasional crazy spikes? That is, what happens when you remove the outliers?