Why Are We Unable to Forecast Extreme Values/’Outliers’?
There’s a complete lack of pushing stats to be the best it can be.
As traders we understand that a single rare occurrence of a stock changing value within a day of 40% can happen. We also for some reason are satisfied with completely accepting the fact that this is an “outlier” I believe in statistics we have completely failed at answering the question of “how do we forecast these” if we want science and statistics to be the best and we think math can answer all questions in our universe why do we accept this limitation?
There are many a resources on extreme value theory and how to fit an extreme value distribution and even how to calculate what an extreme value would be for the distribution. here’s a video describing what is extreme value theory and how it fits into quantitative risk management — also used in the insurance industry.
But as a whole we just solely accept that we cannot forecast these changes. And in most cases we are actually misclassifying them as “outliers” rather than accepting them of extreme values within in a distribution. Here’s a great video depicting the delineation between extreme values and outliers:
I’m very curious y’alls thoughts as I’m going on a journey to try to answer these questions and solve a way to forecast extreme values first in finance but I think as a whole whatever is discovered can apply to many fields.
ps: please message me or leave a comment if interested in exploring this or send me an email firstname.lastname@example.org
Also follow me on substack if you want to embark on this journey: