Apr

16

In 2007 some pairs trades moved 25 standard deviations from their historical norms. If markets were truly random walks, such events would have had miniscule probabilities of ever occurring even had the stock market been in existence since the Big Bang.

As Augustin Lebron said in his excellent book The Laws of Trading, the more market participants assume some event can't happen, the more certain it becomes that the event will happen. Remember when house prices never went down?

William Huggins responds:

that 25 st devs quip just shows you what the spread of their calibration set was, and where the real problem was imo - assuming recent stability was some established fact (one of the problems with exponential weight decay)

Nils Poertner recalls:

Italian friend of mine (big hitter at US bank) had a burn out in 2007, and his wife checked him into a monastery (ora and labora Benedectiner type). Ora is contemplative - it is receiving (not thinking, thinking, thinking) it is what we don't know. Whether you have "AI" or "science models" it is like standing in a room full of mirrors and putting more mirrors into them and thinking one knows the future.

William Huggins writes:

My favourite way out of echo chambering one's self (and to avoid the lure of gurus as this august list taught me forever ago) comes from 17th century Japanese poet Matsuo Basho: "Do not seek to follow in the footsteps of the wise. Seek what they sought". In any competitive domain (markets, chess, etc), you have to learn how to think like a master rather than proverbially copying some smart kid's homework.

An overreliance on historical data instead of trying to imagine what could (quite reasonably) be, is one of the contributing factors that led the 2007/8 crash. In their August 2007 conference call (see pg 21), AIG's Joe Cassano told analysts (regarding their CDS business), "it is hard for us, and without being flippant, to even see a scenario within any realm of reason that would see us losing $1 in any of those transactions".

25 st devs indeed…

Larry R Williams adds:

People talking and over talking their book is a massive clue. Like tom lee this and last year.


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search