Episode 118 ... Manoj Narang (58:47)

Sharp guy, interesting episode.

Has new fund: "MANA Partners"
Formerly of Tradeworx (left July 2015)
Quant trading the way to go
IP has more value than PL you can generate (does that make sense?)
Bringing high frequency trading to asset management investors
Double digit Sharpe ratios, most not comfortable with this
Started at First Boston, 1991, equity derivatives desk
Front office technologist supporting trading desk
8-9 year career on Wall Street
Knew nothing about Wall Street when in college
Never took a statistics class, finance class, economics class
Studied theoretical math and computer science (algorithms, graph theory)
Reasoning quantitatively and thinking logically are the key skills
Algorithmic trading is a quantitative challenge
At Goldman Sachs in late 90s, favorite job, phenomenally smart people
Trading manually back then was exhausting, nothing electronic
Burned out, wanted to start tech firm
Online brokerages to 40% market share in late 1990s, boom
Early adopter of Datek, 1996
Could build tools to help mainstream investors
Started Tradeworx
"Analytical decision support tools"
2001 -- Dot com bubble burst, 9/11 happened, everything changed
Tradeworx became a hedge fund then
Why would a hedge fund commercialize its technology?
2008 a difficult year for hedge funds, Lehman demise
Started high frequency trading in 2008
Eventually became a large high frequency trading firm by volume
Example of valuable IP: intraday trading signals, "MIDAS" bought by SEC to monitor markets
Early adopter of Amazon Web Services, tick data storage
Consolidated audit trail, successor of "OATS" owned by FINRA
Giving regulators investigative powers in modern, fragmented markets
Thesys, an offshoot of Tradeworx, won audit trail contract
Quantitative trading -- hypercompetitive, zero-sum
Paranoid conspiracy theories abound about high frequency trading
Inequality is a natural aspect of any system
Engineer systems to compete after studying the rules (which are complicated)
Questions of fairness, but it's really about discrepancies of skill [one way to rationalize it]
Simplify the regulatory regime, too many rules to understand
Proprietary traders are largest beneficiaries of HFT, then it trickles down
Fragmented markets glued together by HFT
Several dozen trading venues, both lit and dark
Quantitative investing strategies now very crowded
All quants use the same data, same inputs, so use of non-traditional data now key, such as:
Twitter firehose, e*commerce data now mined, satellite imagery
Discretionary managers not worth the fees they charge
If you charge high fees, must empirically show you deserve them, no one can
RenTech, successful quant hedge fund, all employee owned now, no outside money
Secular trend towards mechanization, automation of investing process
Look for patterns in non-traditional data
Renaissance in artificial intelligence now, "machine learning"
Quantitative trading is tolerant of noise
Statistically orthogonal signals
Humans control the capital, not the quants
Quants have super short holding periods, generally
New datasets have very limited histories, so model building from them also limited
Investment success comes from bucking the trend
AQR, DE Shaw, RenTech ... you won't succeed if you try to copycat them
Think for yourself
www.mana-partners.com