There is no freaking way that is a "25 sigma reading". Just looking at the chart you can tell it's not much over 1 sigma (a 1 in six event If you have a distribution that can go down as well as up) event. They must have used as a baseline a period of 3 months or less - and then they outright say the baseline is the 10th percentile of wha…
There is no freaking way that is a "25 sigma reading". Just looking at the chart you can tell it's not much over 1 sigma (a 1 in six event If you have a distribution that can go down as well as up) event. They must have used as a baseline a period of 3 months or less - and then they outright say the baseline is the 10th percentile of whatever that period is. That isn't the way to do it. The baseline HAS to be years and years, or the whole usable time period (if less), and you HAVE to use the mean, not the 10th percentile.
There is no freaking way that is a "25 sigma reading". Just looking at the chart you can tell it's not much over 1 sigma (a 1 in six event If you have a distribution that can go down as well as up) event. They must have used as a baseline a period of 3 months or less - and then they outright say the baseline is the 10th percentile of whatever that period is. That isn't the way to do it. The baseline HAS to be years and years, or the whole usable time period (if less), and you HAVE to use the mean, not the 10th percentile.