This is a one time project, but I'm looking for someone I can work with long term, because I often have these types of projects. So please consider this project a kind of trial to see if a longer term relationship makes sense for us.
The project is to go to [url removed, login to view] and translate one trading system idea I have into code (trading idea is detailed below).
Quantopian is a web app that allows you to write python code to create stock trading systems for free. E.g., you input the rules for when to buy and sell, it applies those rules on historical stock data, and finally it outputs how the system performed in the past according to your rules (i.e., "backtesting").
Experience with backtesting or trading isn't necessary -- if you have good Python skills you should be able to grasp Quantopian quickly. There is good documentation and a forum with plenty of support at Quantopian. If you have experience with Numpy or Pandas, even better, because a lot of the systems use those libraries.
For this project, the trading system idea to turn into Python code in Quantopian is this:
- The securities to potentially buy/sell are: AAPL, REZ, SPY and SHY
- The rules for buying and selling:
--- Once per week, check if each security is above its 200 day moving average. If so, then buy on the open the next day.
--- Continue checking once per week, and if you find the security has fallen below its 200 day moving average, then close the position completely on the open the next day. Else, continue to hold the position.
--- When you buy, each position gets 1/3 of all current capital, whatever that is at the time (even if you only hold 1 stock for a period, it still only gets 1/3 of capital). So if you have 100k, then each position gets 33.3k.
--- Any capital not allocated to AAPL, REZ or SPY (because they're below their 200 day moving averages) goes 100% into SHY (basically SHY is like holding in cash).
--- Each week when you check the 200 day moving averages, you also perform re-balancing: You buy or sell as necessary make each security no more or less than 1/3 of all capital.
That's it. Final deliverable is a copy/paste of the python code that you've already tested and ensured works in Quantopian (just create your own free account).
I'd prefer to bill hourly because these projects can take unpredictable turns, so it's more fair I think if you're paid hourly so you're compensated if the system rules change or things are more complex than expected.
Thanks for your bids!