Performance Measures
Return
The average upside/downside
Risk
The variability in the Return
Risk adjusted Return
Accounts for both
of the above
Risk Adjusted Return
There are some special measures
which help account for both the
Return and Risk of a trading strategy
Sharpe Ratio
is a pretty famous Risk adjusted return measure
Average Return
Risk-free rate
of return
Sharpe Ratio
Risk i.e. Standard
deviation of Return
Sharpe Ratio
The Sharpe Ratio is a very
standard measure to evaluate a
trading strategy
Sharpe Ratio
There are a few conventions
to how it is calculated
Sharpe Ratio
Often, the risk-free
rate is assumed to be 0
Sharpe Ratio
Often, the risk-free
rate is assumed to be 0
Information Ratio
Sharpe Ratio
In the US, the risk-free rate of
return (from treasury bonds) ~ 0
Sharpe Ratio
For futures? - add
explanation here
Sharpe Ratio
The Sharpe ratio is normally
calculated using Annualized
Returns
Sharpe Ratio
First, lets consider the
relationship bet ween daily and
annual returns
Here is a time
series of price
data for the
NIFTY
We can compute the daily
returns from the prices
Return = Ptoday/Pyest-1
Lets consider a simple
trading strategy
where we hold a long
position on the Nifty
for the entire period
Then these would be
the returns we would
get as a result of the
trading strategy
We can compute
the average and
Standard deviation
for this series
Then we know that this
trading strategy has
Average daily return = r
Risk =
The return on each day is a
random variable
with mean = r
standard deviation =
Average daily return = r
Risk =
There are, on average, ~252 trading days in a
year in the US markets
Annual return = R1 + R2 + R3 + + R252
Each of these represents a daily return
Annual return = R1 + R2 + R3 + + R252
Each of these daily returns is a
random variable
Annual return = R1 + R2 + R3 + + R252
It is safe to assume that these random variables are
1) Independent
2) Identically distributed
Mean and SD for
each Random
Variable
Average daily return = r
Risk =
Annual return = R1 + R2 + R3 + + R252
It is safe to assume that these random variables are
1) Independent
2) Identically distributed
IID
Average daily return = r
Risk =
Annual return = R1 + R2 + R3 + + R252
If each of these random variables has
Mean = r
SD =
Annual return = R1 + R2 + R3 + + R252
Mean = 252 *r
Mean = r
SD =
SD = 252 *
Going back to the Annualised
Sharpe
Going back to the Annual Sharpe
r
Assuming the
Daily Sharpe =
risk free rate =0
252*r
Annualised Sharpe =
252*
Going back to the Annual Sharpe
r
Daily Sharpe =
Annualised 252*r
Sharpe =
Annualised Sharpe =
252*r
This factor will depend on the trading
frequency - daily, weekly, monthly
Performance Measures
Return
The average upside/downside
Risk
The variability in the Return
Risk adjusted Return
Accounts for both
of the above
Quant itative Trading
with the help of
Financial Markets
Trading Strategies
developed using
Mathematical Models
involves trading in
Quant itative Trading
with the help of
Financial Markets
Trading Strategies
developed using
Mathematical Models
involves trading in
Now, we come to the heart of
the matter
Mathematical Models
A Quant trader
Studies Historical Data
Mathematical Models
A Quant trader
Studies Historical Data
Identifies patterns in
security prices
Mathematical Models
A Quant trader
Studies Historical Data
Identifies patterns in security prices
Develops mathematical
models that capture these
patterns
Mathematical Models
A Quant trader
Studies Historical Data
Identifies patterns in security prices
Develops mathematical models
that capture these patterns
Uses these
mathematical
models to
develop trading
strategies
Mathematical Models
There are generally 2 steps involved in
developing a trading strategy
Building a model
Testing the model
Backtesting
Building a model
This step answers one question
Should the trader go long or
short on a given security/index?
Building a model
The objective is to build a model
Historical
data
Model
The model inputs could be historical
data for the security, for the
market, macroeconomic factors etc
Long/
Short
Mathematical Models
There are generally 2 steps involved in
developing a trading strategy
Building a model
Testing the model
Backtesting
Backtesting
Backtesting evaluates how the model
would have performed in the past
The return, risk, Sharpe Ratio are
calculated by applying the trading
strategy to past data
Backtesting
Backtesting is a standard way to
evaluate how well a trading
strategy might perform in reality
