Home Deadlines Constraints Submission Sign up details
× Home Deadlines Constraints Submission Sign up details
NUS Investment
Golden Horse

NUS
Golden Horse Fund Management

proudly presents

algorithmic trading competition 2018

Always wanted to put your trading knowledge to the test? Here is your chance to win some big prizes!
1stplace
$3000SGD worth of prizes
2ndplace
$2000SGD worth of prizes
3rdplace
$1000SGD worth of prizes

The Timeline

Drag to your left to see more!
  • 19/1/19: Official Start of Competition
  • 28/2/19: Submission of Backtest
  • 1/3/19: Commencment of Walk Forward Test
  • 1/5/19: End of Walk Forward Test and Judgment Day

Constraints & Guidelines

Asset Universe

  • Forex
  • US equities
  • US Futures (CBOE/CME)
  • ICE Futures / London Metal Exchange
  • Other instruments subjected to request and confirmation

Rules

  1. Asset traded must be sufficiently liquid, trading position equal or less than 10% of daily liquidity
  2. Max DD less than or equal 50%
  3. Starting Capital of $10m, assuming transaction cost 0.02% of trade size (in & out)
  4. Backtest 01/01/2003 – 31/12/2018
  5. Backtest written in Python / R / Matlab / Excel
  6. Calculations based on daily returns

Judgement Criteria & Submission

winner
Back-tests:
  1. Max Drawdown
  2. Sharpe Ratio (Return / Vol)
  3. Months to recovery
  4. % Positive month
Out-of-Sample:
  1. Return
  2. Volatility

Submission Requirements

  • 1) Data / Data Source
  • 2) Code for verification
  • 3) 1-2 page report explaining logic of algorithm, and including performance measures

Is the deadline for the registration 19 January 2019, as per the date designated for the official start date of the competition?
The deadline will be 19th Jan, and teams are allowed to start work on a strategy before that.

Will NUS Invest be involved in the submission of the entries?
On the 19th Jan, after the registration is closed, GHFM will send an email to all participants giving them the official instructions for the competition – which will include the email address to submit their entries to. Entries will be vetted by GHFM, and NUS Invest will not be involved as NUS Invest members are allowed to participate in the competition as well.

Is the forward testing done by GHFM or the teams themselves? If it's done by the teams, are they expected to send a report similar to the submission requirements for the backtest?
For the forward test, GHFM will run it. Teams will be required to write their backtest code/excel such that it will work for the forward test. If there are any issues, GHFM will contact teams to do the forward test and submit it additionally, and will verify that the strategy has not been changed.

Are you allowing any packages to be used? Can I assume that these packages are either accessible to the public (eg. Quantopian, QuantConnect, etc.) or can be sent by the teams?
Packages that are open-sourced can be used. For best coding practices, put all imports at the top of the code.

FAQs by Participants

Will there be profit-sharing for winning teams?
The purpose of the competition is not to connect crowd-sourced algorithms to capital. It is to promote interest and research in algorithmic trading and the hedge fund industry among students, and to identify talents for potential recruitment. As a result, there will be no profit-sharing. Nonetheless, as GHFM will not claim IP over the submissions, teams are free and encouraged to submit their submissions to crowd-sourcing algorithmic trading competitions.

Will there be any mentorship for the competition?
To be fair, GHFM will not provide one-to-one mentorship / help to any teams. That being said, feel free to email us if you require some form of direction or guidance and we will be glad to help. We will make any correspondence available in this FAQ to ensure that information is fairly distributed to all participants.

What would the input format of the forward test code be? Do we have to link our code to any financial data library or can our code be generic and the forward test data would be inputted manually?
The code should be generic such that it can take in different data with different lengths for the forward-testing phase. You do not have to link it to a financial data library, unless you want to.

Do we have to take into account the actual bid ask prices or do we assume it to be part of the implied 0.02% transaction cost in the rules?
You are right, you do not have to take into consideration bid/ask spreads, as the 0.02% transaction costs is used as a proxy to include it. You can however, choose to use open or closing prices.