Apparently what he learned was not "not much", as evidenced by this line: There was a similar quote from the CFO of one of the big banks (David Viniar, Goldman Sachs), who said, "We were seeing things that were 25 standard deviation events several days in a row" during the crisis. Experiencing 25-stdev events is, as any mathematician can tell you, a statistical impossibility, let alone seeing many of them. So what this DB who wrote the Nautilus piece, and the math-inept CFO are really telling us is "our risk calculations were non-sense, because impossible things don't happen." Let's think about mortgages to highlight what is wrong with risk calculations. When a bank makes a bet , it goes on their balance sheet, and there is a risk weighting factor assigned (for the purposes of capital to asset ratio). Government bonds, for example, have a risk factor of 0. That is, they don't really count against a bank's risk, because everyone assumes that the treasury, for example, isn't going to go under. Mortgages get a risk weighting factor of 0.5 independent of the type of mortgage. This was the case before 2008 and continues to be the case now. So, a doctor making $250,000/yr takes out a 60% loan-to-value mortgage on a $500,000 house: she gets a 50% risk weight on the bank's balance sheet. A deadbeat takes out an interest only ARM at 105% LTV and guess what he gets? A 50% risk weight adjustment. Fortunately those types of loans don't really exist anymore, but the risk weight system still does, and it is still fundamentally the same as it was ten years ago. So, to circle back to the original point, when something that your model says in nearly impossible happens, then a bunch of other things that are statistical impossibilities as calculated by your model also happen, then your model is useless. Full stop. If suddenly I dissolved into the aether only to reappear on the other side of the wall, which quantum theory says is technically possible although so unlikely as to be actually impossible, then you also did the same thing, we wouldn't try to recalibrate our theory; we would scrap it entirely and look for a different model that could explain what we now know to be true. For whatever reason finance and economics are impervious to empirical evidence. The risk exposure of the biggest banks have only grown since 2009, as the biggest banks now count balance sheets in the tens of trillions of dollars. We will see another calamity in the offing, and you will see once again people saying that things that couldn't happen did. Hopefully next time our leaders will have the courage to face and to stop what they didn't in 2008-10.But until 2008, I’d never really experienced the extreme tail of a probability distribution firsthand.