Practitioner course in algorithmic trading
Performance Measurement for Trading Strategies
The statistics used to judge trading strategies and managers: the Sharpe ratio, the Sortino, Calmar and information ratios, drawdown, attribution, and the uncertainty in every one of them.
The Sharpe ratio is the most quoted number in trading and one of the most misread. A strategy with a high Sharpe ratio over two years may be lucky; one that sells insurance may look excellent until the year it pays out. This course teaches the statistics and their failure modes together.
How is the Sharpe ratio calculated?
Take the average return in excess of the risk-free rate and divide by the standard deviation of returns. To annualise from daily data, multiply by the square root of the number of trading days, which assumes daily returns are independent. When they are not, as in many trend or illiquid strategies, the simple annualisation is wrong, and the course shows the correction.
How long a track record do you need?
Longer than most people expect. The standard error of a Sharpe ratio estimate is roughly one divided by the square root of the number of years, for a Sharpe near zero. A true Sharpe ratio of one needs years of data before it is clearly distinguishable from zero. Knowing this changes how you read every performance table.
Who is this course for?
Traders, analysts and allocators who compare strategies or managers. It counts toward the Certificate in Algorithmic Trading, the Certificate in Quantitative Portfolio Management and the Professional Diploma in Algorithmic Trading.
What you will be able to do
- Compute and annualise Sharpe, Sortino, Calmar and information ratios correctly.
- Estimate the standard error of a Sharpe ratio and the track record needed to trust it.
- Measure drawdown depth and duration and explain their link to volatility.
- Attribute returns to market exposure, factors and residual skill.
- Spot the ways each statistic can be flattered.
Syllabus, module by module
Returns done right
Simple and log returns, compounding, annualisation and the risk-free rate.
The Sharpe ratio
Definition, annualisation, its standard error, and non-normal and autocorrelated returns.
Downside measures
Sortino ratio, drawdown, Calmar ratio and time under water.
Relative performance
Tracking error, the information ratio and benchmark choice.
Attribution
Regression on market and factor returns; alpha and its uncertainty.
Gaming the numbers
Selling options, smoothing illiquid prices, and other ways statistics are flattered.
How the certificate is earned
An assessed project: you evaluate several real or simulated return series, rank them with appropriate uncertainty, and submit the code with a report explaining which you would fund and why. On a pass you receive the Course certificate in Performance Measurement for Trading Strategies, issued by the Institute with a verification code. Read how certification works.
Before you start
Statistics to the level of confidence intervals and regression, and working Python.
This course counts toward
- Certificate in Algorithmic TradingSix courses and a capstone that take a strategy from idea to a costed, sized and validated system, with performance reported the way a trading desk would read it.
- Certificate in Quantitative Portfolio ManagementSix courses and a capstone on building portfolios with numbers: mean-variance and its failures, robust optimisation, Black-Litterman views, factors, attribution and risk.
- Professional Diploma in Algorithmic TradingThe top credential on the practitioner path. Eight courses and a 40-hour capstone covering everything between a trading idea and a book you could defend to a risk committee.
Questions applicants ask
What is a good Sharpe ratio?
It depends on the strategy, the costs included and the length of the record. An annual Sharpe ratio above one after costs is often treated as good for a single strategy, but a short track record with that number is weak evidence. The course shows how to compute the uncertainty.
Who created the Sharpe ratio?
William F. Sharpe introduced the measure in 1966 as the reward-to-variability ratio. He shared the 1990 Nobel memorial prize in economics for work on asset pricing.