all links: [[antidatamining refs]]

project exists since 2005 (rise of Google, beginning of web2.0, no datamining)
-> no wikileaks, snowdon
-> difficult to find good data - created their own tools
-> global surveillance was only a trend

Flashcrash May 2010 = frequency trading
algorithmic crash -> with 'big resonance'
looked for data of this crash, sonorized it as art project
ex. (not very precise dataset, but enough)
difficult to understand what happened, millions of second's timescale
1 second -> billions of data
crash represents 1 trillion dollars - back to 0 few days later ...
happened on 94 markets
technique of 'quote stuffing': sell a lot of fake orders that are never validated, 'pollution'
-> markets began to buy with wrong proces, phenomen of 'algoritmic resonance'

RYBN decided to investigate algorithms (black box - commercial secrets you don't know anything about)
no one analyses robots on the market, responsibility is with the banks

commentator from stream service that provides comments on the market
tried to reproduce moods, keep adrenaline
you can subscribe for free to this service for 1 month
RYBN recorded them

systematic structure: 17000 flash crashes between 2006 - 2011

tried to get into the market, programmed a trading robot
-> you need access to the market
-> you need money (commissions are expensive, what you pay when you sell/buy)

2011: 10000$ from French Dpt of Culture to put on the market
disaster till beginning of 2014, now a lot better
program lives its own life
-> it tweets on twitter (social life, reports on time before bankruptcy) -> transforms system into narrative

Algotrading Freakshow
sources on website (100 articles: scientific papers, reports) & continue to feed
-> introduction of physics into finance (probability), automatons, fiinancial market structure, court reports...
felt like butterfly collectors (of diagrams/anything you can reproduce as recipe)
-> only focus = price

deviations: demantic prediction bots article by Cecolia Porteiro (?)

they reenact trading robots during exhibitions
to understand model

Personal history of finance is mythological system, related ot computer
organise factory to produce methematical computation - Gaspar de Prony
computer = (neo)liberal system, reflects ways of working in factory -> not neutral
Louis Bachelier, markets started to use mathematicla equations to measure risks
W.D. Gann, pioneer of technical analysis (astrotrading), 1909 benefits of 1000% on the market, started writing books to explain his methods
cosmic vibrations that influence the prices - Esoterism (not rational)
-> predicted the crash, made money with it
Robert Krausz, famous loser
-> esoteric part is still very present

Web Bot, comes from rational field 

Burton Malkiel, A random Walk Down Wall Street
-> quants, don't study finance, but mathematicians & physics
a trader just has to learn some rules to mak money, 'turtle program' of Richard Dennis

PNL: Benefit & Loss of an algotrader

text mining based algorithms
most of diagrams are not succesful experiments
-> important
usually you don't find sources for algorithms online, if you find them, they'r eno good
every algorithm is narrative in itself
ex. Twitter Oracle, 2010 < scientific paper (robot in lab), 84% of success - mood of people influence prices - recipe is online
-> 2011 hedgefund with scientists start to implement it (with 5 million $, 3 months later sold in 185.000$); when they launched, they were followed by ayone else & added with self learning machines

Hashcrash 2013
Hackers publish fake tweet 'Two explosions in White house, Obama injured'
1 min after: american markts go in flash crash
-> why they only focus on failures, helps to understand!

RYBN Robot, 2011 in Pure Data
sorting of stocks on market
train robot with history of stock variations
finds vector, choses 'best' performance and buys/sell
-> narrative is not robot, but bankrupcy - metaphor, contribution to self fulfilling prophecy
only lost 300$
has a bank account & algorithmic broker
fake cie, mailinglist (does not work anymore) - only advertising campaigns
-> only making catastrophic predictions - 'psycological war', self fulfilling prophecy
a superstitious investment fund (sentiment analysis, simulation)

markt: the faster you go, the more you win, no big space for chances (fast machines, brokers...)
politicians try to listen to what market has to say - no one knows, rules are tricked (cfr meaning of a bird in the sky)
everyone knows it goes in a crack
systems are not stable