“An edge in analysis of social media is much more feasible than in analysis of news – there are only six traders globally who are successful trading off news analysis, ” Rob Passarella, DataSift.
“There is a huge leap to move from getting interesting signals to a viable investment strategy,” Leigh Drogen, Estimize.
On using sentiment data in financial models: ” A very specific model will give a lot of false positives,” Peter Hafez, RavenPack
Richard Edwards, HED Capital Management, “Fear is even stronger than greed in moving markets.”
“The third time a trader hears a story from the market he should fade the trade,” Richard Edwards.
“Commentators and regulators have to distinguish between institutional algorithmic execution and high frequency trading, ” Tommi Vuorenmaa, Valo Research and Trading.
“By creating the category of non-obligated prop trading firms, regulators and exchanges have removed the cross-subsidy of trading in small cap shares by those trading in large cap,” Natan Tiefenbrun, Turquoise.
“A lot of the alpha from quant managers has been negative recently, but it is still pure alpha,” John Godden, Sciens Asset Management.
“We were asked by an investor to start an HFT strategy, and they were a big client of ours – so we did,” Giovanni Belliossi, FGS Capital.
“The best fund managers in quant are self critical, ” Rob Moore, FRM.
“There is performance persistence in manager returns,” Prof. Robert Korajczyk, Kellogg School of Management Northwestern University.
“Machine Learning is just a statistical method for addressing markets,” Dr. David Surkov, Egham Capital.
“Modern machine learning is so powerful that you can throw any data at it – including random data – and it will generate “a good model”, but it will break down spectacularly after a period,” Dr. David Surkov.
“Neural networks have been a disaster when used in investing,” Dr. Marco Fasoli, Titian Global Investments.
“Dumb artificial intelligence systems tend to out-perform complicated A.I. systems – it is very important not to overfit!” Anonymous Panellist.
On using machine reading textual analysis: “Not since Caxton has text been so sexy,” Rob Passarella, DataSift.
Does John Godden go on a comedy roadshow? That line about negative alpha was solid gold and I almost fell out of my chair laughing.(MODERATED)
people actually paid to listen to these pearls of wisdom?
The concoction of someone that follows machine based learning and finds Caxton sexy is a recipe for disaster! Would be interesting to see how he has fared 3 years hence