Thread by @MacurdyTrader

My personal generalized thoughts and conclusions from #RWRI 12 @nntaleb

-Statistics is about the elimination of noise
-Decision making gets easier in uncertain conditions
-Skin in the game is a filtering mechanism
-Group behavior is not expressed by individual preferences

-Nature is scale invariant
-You create tail events simply by regulating the process - it’s an unnatural constraint
-It’s misconstrued that lowering variability actually lowers risk
-McDonalds exists as an insurance policy
-Everything non-linear is necessarily convex/concave

-Don’t theorize about what people want, just act on what they do
-Have optionality on the convexity of random events
-PTSD vs Post traumatic stress growth. One is talked about because it makes $$$
-Floor your losses
-Zero intelligence trading works when you have convexity bias

Complex systems deal with interactions/feedback loops of agents and becomes computationally irreducible
-Prediction in complex systems is fundamentally beyond our reach
-Think in bounds and limits, not precision
-We use continuity for assumptions in math, not descriptions of life

Reality doesn’t fit the continuous model
-Discrete displays things continuous does not
-Don’t fuck with the tail
-Cut pollution in half and you see non-linear benefits
-The more uncertainty makes the argument stronger for the precautionary principle #antifragile

-Complexity comes from simple recursion
-Scale invariance means adding a 0 does not change the proportion
-A non-linear domain, aka extremistan, has exponential function
-If you cant afford insurance, don’t buy the portfolio

Don’t back test, just look at the distribution. Scale invariance says that going forward we can have large sigma events not captured in empirical back tested history. We only calibrate to the worst that ever happened, not that will ever happen. That’s why options are undervalued

-Complex systems are full of interdependencies, hard to detect, non linear results, cascading effects, and feedback loops
-There is a point where the risk is so immense that it’s beyond worth insuring - were all dead anyway
-% gains vs % losses are asymmetric
Drawdowns memoryless

-The reaction of people to panics is the same kind of response regardless of the initial event. Look for 2nd order effects
-Outlier vs extreme: does it lie outside the range of variation of the distribution
-Extreme value theory says the world is not getting safer. War could be >

-Shadow mean is the way to correct for the visible mean
Ex. 2 people die by Ebola
2000 die by lawn mowers
Shadow mean in Ebola could be 2 or 2 million. Insufficiency of the sample is the problem.
-There is no way to prove me right, only a way to prove me wrong

-Don’t buy insurance on the titanic from someone on the titanic
-When stress testing, plot the acceleration to determine what domain were in
-Endogenous factors far outweigh exogenous ones in bubbles
-Evolution uses extreme value theory for dealing with tail events

So much wisdom shared in a compressed amount of time. Many thanks to the instructors at RWRI @financequant @normonics @trishankkarthik @VergilDen @RaphaelDuoady @ArieHaziza

Great meeting the attendees @coordinatedsync @brenorb @WorkinnDaniel and many more

Awesome meeting the legend himself and he’s much funnier in person