We Manage Typical Risks of Quant Strategies Carefully
Model risk(Back-testing works but
real trades don’t)
We do not rely purely on back-testing; paper trades
are a norm
We
proactively avoid data-mining and over-optimization
Transience risk(Signals lose edge quickly
with time)
We
track closely all our live signals for signs of
fading
We
watch for early warning signs of loss of edge
Leverage risk(One large leveraged bet
wrecks returns)
We
trade the entire SSF universe of over 150 names –
less than 5% in any one name
Our
net leverage is close to zero and we also avoid
single sector over-exposure
Scale risk(Returns dwindle with
increased size)
We
proactively discard scale-sensitive signals and
constantly watch for impact costs
of actual trading
Diversification
into over 150 stocks also allows us to reduce scale
risks
Regime risk(Change in regime
reduces performance)
We
study signals for their regime behavior a prior
We allow for regime dependent variation in capital
allocation
