Trading algorithms and summary of Jane Street Electronic Trading Challenge
Yuanzhe Liu: CS undergrad, rich experience in software development, VPN business owner
Hao Yuan: Industrial Engineering and CS undergrad, trading algorithms developer on Quantopian, 2 year stock and option trading experience
-
Coded everything in python
-
Concurrent development: Yuanzhe focuses on TCP connection, Hao focuses on strategies
Common stock & ADR pairing: Since common stock is more liquid than its pairing ADR, we use common stock to estimate its ADR's fair value. When ADR is undervalued, we buy ADR, pay a fee to convert to common stock, and sell in strength
ETF Arbitrage: We use the holdings of the ETF to estimates its fair value. When ETF is undervalued, we buy ETF, pay a fee to convert to common stocks of the holdings, and sell in strength. When ETF is over valued, we buy common stocks of the holdings, pay a fee to convert to ETF, and sell in strength.
-
chose a progamming language that was easy to use
-
Divided the roles, stayed focused, and helped each other
-
Started from simple strategies to complex ones
-
Top 4 of the competition
-
Set up TCP earlier, our first test began 3 hours after the pool opened
-
Track pending orders and cancel the ones we don't need
-
Less aggressive algorithms (We were top 2 for 2 rounds and bottom 5 for a few)
-
Implement hedging strategies for risk protection