Financial Crisis and Lessons for Engineers

Whilst your content may be admirable the title is rather daft..
I think the financiers should be taking lessons from engineers and scientists not giving them.
They are a bunch of charlatans and best and crooks at worse. Ant Engineer or Scientist who operated the same way as the banks and the money men would be fired damn quick.
Del

No doubt that many of these financiers were neglegent at best and criminal at worst. Engineers get people killed when they screw up like this - the plane or bridge fail catastrophically.

I had a friend in the industry who warned of exactly this scenario last New Years. And his major point was that everyone relied on model predictions. It sounded like anyone who questioned the models' predictions was treated as a heretic.

Maybe I should have titled it "Models are fallible too"?

Yeh, people confuse maths with reality...maths is just a tool like any other and it can easilly be missused.

how does that saying go;

figures don't lie, but liars can figure.

and there nothin worst than an abused tool.

Measurements of the fingerprints - my term - of systems over long periods of time always provide the information asserting the non-randomness of disturbances underpinning systems failures; and the root-causes of these disturbances in some cases can be determined through a decoupling of the Fingerprint Time Series into harmonic patterns (or if you prefer patterns harmonics) which can effectively be related to each of the systems degrees of freedom and therefore failure modes and failure forecast capability.

Just my approach to correlated forecasting.

Good reflection though on your blog.

I was told a famous mathmetician composed a group for his financial service. All of them are composed by a group of mathematicians and physicians without anyone of financians. they loook down up Well str. and win billion every year.

probability and statistic is important base on modeling. games theory and information theory( of cause include prediction encode) play important roles. a good modeling can win but, otherwise will be fail heavily.

not everyone is able to understand verious of distribut and sampling.

Can anyone out there introduce me more useful control theory forum or bbs?

I run a website which has a wiki, forum, and blog devoted to Control Systems. Start here if you are interested http://wikis.ControlTheoryPro.com. The google groups on control (sci.engr.control) is reasonbly active with a lot of good people on it but also a lot of spam. Admittedly my forum, while still online, is largely dead due to lack of interest by anyone but me.

How would you work this into your model ?

http://www.prisonplanet.com/thousands-of-troops-are-deployed-on-us-streets-ready-to-carry-out-crowd-control.html

and this

http://www.prisonplanet.com/half-a-million-plastic-coffins.html