The Formula That Killed Wall Street

[Unimatrix11]

So world war 1, was not predictable?! Yeah, maybe, i, along with most historians, fall for the third property of ´black swans´, but AFAIK, everybody saw it coming. What nobody did predict, was the course of it and that it would last that long. But when he says, that everybody thought the world had been civilized and realized the need for peace during the century after napoleon, he is mistaken, to a point that is very close to the opposite of the facts. War was in fact something many people (or shall i say ´men´?) were looking forward to, as crazy as it may sound today. War fiction was what science fiction is today, kinda. All the fictions, even the truest to reality, were widely off though. So, it is clearly the course of the war, that seems to have been easily predictable before it started, but was not at all - not the war itself.

[TCO]

I think there is a lot of silliness in the tendancy to think that a particular formula hurt WS or that a correction would help it. It is part of the quant misconception. I'm not dissing the interestingness of looking at stats in finance. But think that there are many fundamental insights that can be gained from reading Brealey and Myers (especially the boxed parts and sidebar discussions) and thinking about them from a fundamental micro econ point of view.

It's like when I told my screwed up and inscrutable 8 type of product, 3-4 convertor step set of downstream businesses...eff you. We are going to figure out how much product left the docks in fiber form. And how much got sold in converted form. And I am going to go to shipping reports for the one...and sales reports for the second. And screw the messed up in between stuff.

Or when the Master's degree student asked Nobel prize winner at LTCM if there really were market inneficiencies enabling their strategy to work...and was cursed for it and told that the reason for said nickles to be picked up was because of jerks like him...but the Nobel prize fooker was the one wrong in the end...

[TCO]

Oh...and too much damned black swan crap. Puleeze. Just make the speculators pay their own bills with no bailout...and Bayesian derived market forces will lead to the proper allocation of capital...eventually.

[Naked Gents Rut]

I don't get the importance of the Black Swan concept. Yeah they exist, but what the **** are you supposed to do about them? Unknown unknowns are pretty much impossible to predict or prevent.

[Impaler[WrG]]

I think the real thrust of this Black Swan concept is that people foolishly behave as if events that are know to have very low probability will never ever happen and then get whacked upside the head when said event dose in fact happens. It all boils down to a recommendation for individuals and society to spend more time preparing for the consequences of rare events.

[Naked Gents Rut]

How do you meaningfully prepare for a fundamentally unknowable event?

[TCO]

I don't get the importance of the Black Swan concept. Yeah they exist, but what the **** are you supposed to do about them? Unknown unknowns are pretty much impossible to predict or prevent.

I don't care for the Black Swan concept. But I do think that unknown unkonwns are not completely un dealable with. In essence the important thing is to kmove unknown unknown to known unknown. One way can be knowledge of your limitations. For instance, if you are all over launching a new product from the technical aspects, but realize that you lack perspective on marketing communications, this can drive getting help in thast area. However if you are cocky, you will not even know or acknowledfge the gap. I read somewhere that the duymbest people are the most confident in their opinions and the smartest most nuanced and caveated. Similarly, there are times when definite knowledge may be lacking but a hunch may be helpful. IU think these faculties can be developed and honed, especially in particular areas.

[Naked Gents Rut]

In essence the important thing is to kmove unknown unknown to known unknown.

True. I'm not sure how easily this can be done in practice, however.

[TCO]

I'm thinking more aboiut business situations or even military strategy...but I think there are processes that are conducive to this. Heck, I often do internal (and exgternal) interviews for business analysis. I genreally have hypotheses I'm testing and have worked out qeustions, to get max benefit from the time. But I already know from doing this a lot, that not only will I learn things that move known unknown to known known...but I can predict that I will surface some unknown unknowns.

[KrazyHorse]

Moving unknown unknowns to known unknowns is not an unthinkable process even in a model-based field.

For instance, some of the most important work being done in theoretical HE physics has to do with estimating/bounding uncertainties in physical observables that come from not being able to fully do the calculations. In a more empirical sense you can examine your assumptions one by one and look for evidence of situations where these assumptions failed. How often did these failures occur? Are there any other associated signals?

[Zkribbler]

The whole purpose of scientific and other investigations is to chnage unknown unknowns to known unknowns. Take for example, Columbus. He sailed off into the West, not really knowing what he'd find. Sure, he fell off the edge of the Earth where he was devoured by monsters, but at least he found out!!

[Naked Gents Rut]

Take for example, Columbus. He sailed off into the West, not really knowing what he'd find.

No. Columbus thought he knew what he would find, ie. an Asia that was far closer to Europe than it actually is. He was saved from death by an unknown unknown. Columbus was lucky.

[TCO]

Look at it from his patron's point of view, though. They might have realized that sending explorers out would return value regardless of whether the specific hypothesis was validated or not. More like doing botanical study on a new continent. It's actually unknown unknwon moving to to known known.

I also think with experience and consultation that a single person with some unknown unknowns can get them surfaced and at least have known unknowns, if not known knowns. This is a good reason for having other people review your work or plans.

Also, one person may have a lot unknown unknowns...because he is a blockhead who doesn't even appreciate his limitations. But his sponsor, manager should be aware of this and watch out for it and compensate.

[Unimatrix11]

You guys lost me a long time ago - so i just blurb out something fitting to my avatar: Expect the unexpected...

[TCO]

I'm losing people because I'm unclear even to myself. I think your statement is pithy and actually possible (to some extent).