RIT Exercise F1. Trading on public information

Transcription

1 RIT Exercise F1 Trading on public information Timing The F1 case will be 24/7 on server jhasbrou.stern.nyu.edu. For exact times see NYU Classes announcements. Play at least three times. Your profit on each run is Realized P&L. You can play up to 30 times. Your overall score = average profits (averaged over the top half of your runs). 2018, Joel Hasbrouck, All rights reserved 2 1

2 Overview You are a news trader trading on macroeconomic news before others can trade. RIT stock index is RTX. At beginning of run, RTX = 1,050. Traded security is a futures contract based on the RTX. The size of the contract is $250 RTX Initial notional value is $ = $262,500 You will be trading one of ten identical futures contracts, designated RTX1F, RTX2F,, RTX10F. A run lasts 360 seconds (six minutes). 2018, Joel Hasbrouck, All rights reserved 3 And in the real world October, 2017: current S&P composite index is SPX 2,500 Size of CME E-mini S&P futures contract is $50 SPX Notional value = $50 2,500 = $125, , Joel Hasbrouck, All rights reserved 4 2

3 At random times, there will be random news. If news is positive, buy; if negative, sell (short) Automated traders also get the news, and react Some are fast, some are slow. You should try to trade along with the fast ones. Every run has a different random draw of the news. You have a position limit of 20 contracts. 20 $262,500 = $5.25 Million In a typical class, most players do around ten runs. Average of top three usually $200,000 to $300, , Joel Hasbrouck, All rights reserved 5 The news window A news item contains text. Your firm has an algorithm that filters the text and produces a target estimate for RTX. Some news is trivial small change in estimate. Some news is substantial large revisions. Next A run with the first news item at 36 seconds. 2018, Joel Hasbrouck, All rights reserved 6 3

4 7 Starting at 36 seconds, the automated traders start selling, driving the price down to (about) , Joel Hasbrouck, All rights reserved 8 4

5 At 90 seconds, the traders start driving the price down to (about) , Joel Hasbrouck, All rights reserved 9 The slow automated traders Slow traders are patient. They use limit orders and behave like market makers. They are slow to react to news. When there is news, they don t cancel their bids and offers. Their bids and offers are stale: hit/lift them to make profits! 2018, Joel Hasbrouck, All rights reserved 10 5

6 The fast automated traders The fast traders get the news at the same time you do. They buy on good news, sell on bad news. They try to hit/lift the stale bids/asks of the slow traders. Within seconds, they drive RTX to the new value. During that adjustment period, they are your competitors. You need to react when the price adjustment is just starting. Compare: SPY reaction to 8:30 announcement. 2018, Joel Hasbrouck, All rights reserved 11 Hints When there is news announcement, use market orders. You need to buy/sell quickly, while the price trend is just beginning. To trade quickly, configure your RIT client for one-click orders. In the Book Trader panel: Toggle the lightning bolt on/off to activate. Set the quantity to 10 or 20 (contracts) Then: To buy, right-click on the ask side of the book. To sell, right-click on the bid side of the book. 2018, Joel Hasbrouck, All rights reserved 12 6

7 Workspace windows The portfolio panel (which will report your profits). The book trader panel (to enter orders). The news panel (to view incoming news items) The security charting panel (to plot the price) 2018, Joel Hasbrouck, All rights reserved 13 Why not use limit orders? Suppose there is good news then bad Ask (Your) bid 2018, Joel Hasbrouck, All rights reserved 14 7

8 Start by using just market orders Once you re experienced with market orders Maybe experiment with limit orders. Use limit orders only when the price has fully adjusted and there is no major trend under way. 2018, Joel Hasbrouck, All rights reserved 15 Warning This is a simulated market a slow market The price adjustment occurs over seconds. Real markets are much faster Many price adjustments are complete in under a second. 2018, Joel Hasbrouck, All rights reserved 16 8