Blankfein vs Blankfein

Goldman Sachs CEO Lloyd Blankfein gets into an argument with himself in the FT today. Consider this:

In the past several months, we have heard the phrase “multiple standard deviation events” more than a few times. If events that were calculated to occur once in 20 years in fact occurred much more regularly, it does not take a mathematician to figure out that risk management assumptions did not reflect the distribution of the actual outcomes.

And then consider this:

We should resist a response, however, that is solely designed around protecting us from the 100-year storm.

So, having conceded that risk models do a terrible job of predicting the likelihood of outcomes, Blankfein now wants us to buy the idea that those same models can differentiate between 100-year and 1-year-storms. Am I the only one who sees a material inconsistency here?

[via FT]