CS4646 manual strategy

Overview

In this project you will develop a trading strategy using your intuition and Technical Analysis, and test it against a stock using your market simulator. In a later project, you will use your same indicators but with Machine Learning (instead of your intuition) to create a trading strategy. We hope Machine Learning will do better than your intuition, but who knows?

Template

There is no distributed template for this project. You should create a directory for your code in ml4t/manual_strategy and make a copy of util.py there. You will have access to the data in the ML4T/Data directory but you should use ONLY the code in util.py to read it.

You should create the following code files for submission. They should comprise ALL code from you that is necessary to run your evaluations.

• indicators.py Your code that implements your indicators as functions that operate on dataframes. The "main" code in indicators.py should generate the charts that illustrate your indicators in the report.
• marketsimcode.py An improved version of your marketsim code that accepts a "trades" data frame (instead of a file). More info on the trades data frame below.
• ManualStrategy.py Code implementing a ManualStrategy object (your manual strategy). It should implement testPolicy() which returns a trades data frame (see below). The main part of this code should call marketsimcode as necessary to generate the plots used in the report.
• BestPossibleStrategy.py Code implementing a BestPossibleStrategy object (details below). It should implement testPolicy() which returns a trades data frame (see below). The main part of this code should call marketsimcode as necessary to generate the plots used in the report.

Note that we may not test your code, so we may not know if you didn't organize your code as recommended, but this arrangement will be required for later projects, so it is worthwhile getting it set up this way. The key requirement is that, if necessary, a TA should be able to run your code on a buffet machine and get the same results (e.g., statistics and charts) that we see in your report.

Data Details, Dates and Rules

• Use only the data provided for this course. You are not allowed to import external data.
• For your report, trade only the symbol NFLX. This will enable us to more easily compare results.
• You may use data from other symbols (such as SPY) to inform your strategy.
• The in sample/development period is January 1, 2003 to December 31 2004.
• The out of sample/testing period is January 1, 2005 to December 31 2006.
• Starting cash is $100,000.
• Allowable positions are: 1000 shares long, 1000 shares short, 0 shares.
• Benchmark: The performance of a portfolio starting with $100,000 cash, investing in 1000 shares of NFLX and holding that position.
• There is no limit on leverage.
• Transaction costs for ManualStrategy: Commission: $9.95, Impact: 0.005.
• Transaction costs for BestPossibleStrategy: Commission: $0.00, Impact: 0.00.

Part 1: Technical Indicators (20 points)

Develop and describe at least 3 and at most 5 technical indicators. You may find our lecture on time series processing to be helpful. For each indicator you should create a single, compelling chart that illustrates the indicator.

As an example, you might create a chart that shows the price history of the stock, along with "helper data" (such as upper and lower bollinger bands) and the value of the indicator itself. Another example: If you were using price/SMA as an indicator you would want to create a chart with 3 lines: Price, SMA, Price/SMA. In order to facilitate visualization of the indicator you might normalize the data to 1.0 at the start of the date range (i.e. divide price[t] by price[0]).

Your report description of each indicator should enable someone to reproduce it just by reading the description. We want a written description here, not code, however, it is OK to augment your written description with a pseudocode figure.

At least one of the indicators you use should be completely different from the ones presented in our lectures or examples. (i.e. not Momentum, SMA, EMA, RSI, Bollinger Bands, MACD, MACD Divergence, or anything else I have shown you or given code for). We will be even more impressed if you actually invent your own indicator.

Part 2: Best Possible Strategy (20 points)

Assume that you can see the future, but that you are constrained by the portfolio size and order limits as specified above. Create a set of trades that represents the best a strategy could possibly do during the in sample period. The reason we're having you do this is so that you will have an idea of an upper bound on performance.

The intent is for you to use adjusted close prices with the market simulator that you wrote earlier in the course. For this activity, use $0.00, and 0.0 for commissions and impact respectively.

Provide a chart that reports:

• Benchmark, as defined above, normalized to 1.0 at the start: Blue line
• Value of the best possible portfolio (normalized to 1.0 at the start): Black line

You should also report in text:

• Cumulative return of the benchmark and portfolio
• Stdev of daily returns of benchmark and portfolio
• Mean of daily returns of benchmark and portfolio

Your code should implement testPolicy() as follows:

df_trades = bps.testPolicy(symbol = "AAPL", sd=dt.datetime(2010,1,1), ed=dt.datetime(2011,12,31), sv = 100000) 

The input parameters are:

• symbol: the stock symbol to act on
• sd: A datetime object that represents the start date
• ed: A datetime object that represents the end date
• sv: Start value of the portfolio

The output result is:

• df_trades: A data frame whose values represent trades for each day. Legal values are +1000.0 indicating a BUY of 1000 shares, -1000.0 indicating a SELL of 1000 shares, and 0.0 indicating NOTHING. Values of +2000 and -2000 for trades are also legal so long as net holdings are constrained to -1000, 0, and 1000.

Part 3: Manual Rule-Based Trader (50 points)

In ManualStrategy.py implement a set of rules using the indicators you created in Part 1 above. Devise some simple logic using your indicators to enter and exit positions in the stock.

A recommended approach is to create a single logical expression that yields a -1, 0, or 1, corresponding to a "short," "out" or "long" position. Example usage this signal: If you are out of the stock, then a 1 would signal a BUY 1000 order. If you are long, a -1 would signal a SELL 2000 order. You don't have to follow this advice though, so long as you follow the trading rules outlined above.

For the report we want a written description, not code, however, it is OK to augment your written description with a pseudocode figure.

You should tweak your rules as best you can to get the best performance possible during the in sample period (do not peek at out of sample performance). Use your rule-based strategy to generate an orders dataframe over the in sample period, then run that dataframe through your market simulator to create a chart that includes the following components over the in sample period:

• Benchmark, as defined above, normalized to 1.0 at the start: Blue line
• Value of the rule-based portfolio (normalized to 1.0 at the start): Black line
• Vertical green lines indicating LONG entry points.
• Vertical red lines indicating SHORT entry points.
• Vertical black lines indicating EXIT points (your holdings went to zero).

We expect that your rule-based strategy should outperform the benchmark over the in sample period.

Your code should implement the same API as above for best possible:

df_trades = ms.testPolicy(symbol = "AAPL", sd=dt.datetime(2010,1,1), ed=dt.datetime(2011,12,31), sv = 100000)

Part 4: Comparative Analysis (10 points)

Evaluate the performance of your strategy in the out of sample period. Note that you should not train or tweak your approach on this data. You should use the classification learned using the in sample data only. Create a chart that shows, out of sample:

• Benchmark, as defined above, normalized to 1.0 at the start: Blue line
• Performance of manual strategy: Black line
• Both should be normalized to 1.0 at the start.

Again, you are not penalized for poor out of sample performance. Optimize using in sample data only, then let the out of sample result be whatever it is.

Create a table that summarizes the performance of the stock, and the manual strategy for both in sample and out of sample periods. Do your best to explain any differences.

Hints

Overall, I recommend the following steps in the creation of your strategies:

• Indicator design hints:
  • For your X values: Identify and implement at least 3 technical features that you believe may be predictive of future return.
  • It is usually worthwhile to standardize the resulting values (see https://en.wikipedia.org/wiki/Standard_score).
• Rule based design:
  • Use a cascade of if statements conditioned on the indicators to identify whether a LONG condition is met.
  • Use a cascade of if statements conditioned on the indicators to identify whether a SHORT condition is met.
  • The conditions for LONG and SHORT should be mutually exclusive.
  • If neither LONG or SHORT is triggered, the result should be DO NOTHING.
  • For debugging purposes, you may find it helpful to plot the value of the rule-based output (-1, 0, 1) versus the stock price.

Contents of Report

Describe each indicator you use in sufficient detail that someone else could reproduce it. You should also provide a compelling description regarding why that indicator might work and how it could be used. You should also provide one or more charts that convey how each indicator works in a compelling way. (up to 8 charts).

For the best possible strategy, describe how you created it and any assumptions you had to make to make it work. Provide a chart that illustrates its performance versus the benchmark.

For your manual strategy, describe how you combined your indicators to create an overall signal. How do you decide to enter and exit your positions and why? Why do you believe (or not) that this is an effective strategy? Provide a chart.

Compare the performance of your manual strategy versus the benchmark for the in sample and out of sample time periods. Provide a chart.

Your report should be no more than 3000 words. Your report should contain no more than 14 charts. Penalties will apply if you violate these constraints.

Expectations

• In-sample backtests should perform very well.
• Out-of-sample backtests should... (you should be able to complete this sentence).

Clarifications

• You may use data other than adjusted close if you wish. The functions in util.py already can do this.
• You may only have one position at a time: -1000 shares, + 1000 shares, 0 shares.
• You may trade up to 2000 shares at a time to move from +/- 1000 shares to -/+ 1000 shares in a single day. You may not have a position of +/- 2000 shares.
• You do not need to check your available cash before making a trade. In other words, you can have unlimited leverage, but you cannot exceed the position limits given. This is a detailed way of saying "don't worry about cash".

What to turn in

Turn your project in via t-square.

• Your report as report.pdf
• All of your code, as necessary to run as .py files.
• Document how to run your code in readme.txt.
• No zip files please.

Rubric

Start with 100 points, deductions as follows:

Neatness (up to 5 points deduction if not).

Bonus for exceptionally well written reports or novel indicators (up to 2 points)

Indicators (up to 20 points potential deductions):

• Is each indicator described in sufficient detail that someone else could reproduce it? (-5 points for each if not)
• Is there a chart for each indicator that properly illustrates its operation? (-5 points for each if not)
• Is at least one indicator different from those provided by the instructor's code or lectures? (-10 points if not)
• Does the submitted code indicators.py properly reflect the indicators provided in the report (-20 points if not)

Best possible (up to 20 points potential deductions):

• Is the methodology described correct and convincing? (-10 points if not)
• Is the chart correct (dates and equity curve) (-10 points if not)
• Is the chart correct (dates and equity curve) (-10 points if not)
• Historic value of benchmark normalized to 1.0 with blue line (-5 if not)
• Historic value of portfolio normalized to 1.0 with black line (-5 if not)
• Are the reported performance criteria correct ? (-2 points for each item if not)

Manual rule-based trader (up to 50 points deductions):

• Is the trading strategy described with clarity and in sufficient detail that someone else could reproduce it? (-20)
• Does the provided chart(s) include:
  • Historic value of benchmark normalized to 1.0 with blue line (-10 if not)
  • Historic value of portfolio normalized to 1.0 with black line (-10 if not)
  • Are the appropriate date ranges covered? (-10 if not)
  • Are vertical lines included to indicate entries (-10 if not)
• Does the submitted code ManualStrategy.py properly reflect the strategy provided in the report? (-20 if not)
• Does the submitted code and report reflect an understanding of the subject matter (up to -30 if not)
• Does the manual trading system provide higher cumulative return than the benchmark over the in-sample time period? (-10 if not)
• Did the student use the correct symbol? (-10 if not)
• Did the student use the date periods? (-10 if not)
• Does the strategy obey holding constraints (-5 if not)

Comparative analysis (up to 10 points deductions):

• Is the appropriate chart provided (-5 for each missing element, up to a maximum of -10)
• Are differences between the in-sample and out-of-sample performances appropriately explained (-5)
• Does the submitted code and report reflect an understanding of the subject matter (up to -5 if not)
• Is the required table present and correct (up to -5 if not)

Required, Allowed & Prohibited

Required:

• Your project must be coded in Python 2.7.x.
• Your code must run on one of the university-provided computers (e.g. buffet02.cc.gatech.edu).
• Use only util.py to read data.
• All charts must be generated in Python, and you must provide the code you used.

Allowed:

• You can develop your code on your personal machine, but it must also run successfully on one of the university provided machines.
• Your code may use standard Python libraries.
• You may use the NumPy, SciPy, matplotlib and Pandas libraries. Be sure you are using the correct versions.
• Code provided by the instructor, or allowed by the instructor to be shared.

Prohibited:

• Generating charts using a method other than Python.
• Any other method of reading data besides util.py
• Any libraries not listed in the "allowed" section above.
• Any code you did not write yourself