Initial Portfolio - SP500

Alphathon
#1

Hello


After having started, I came across a question regarding the description of the task in combination with the available data. Unfortunately I didn’t find an answer in the forum so far.


The description states, that at every point in time 01/2007 - 12/2016 we must fully invest into 50 stock, equally weighted. At the same time we obtain the data for the same period. I currently don’t understand the exact implications on the first portfolio weights and see 2 possible solutions:


  • The initial portfolio (as of January 03, 2007) is chosen randomly, since this decision can not be based on historical data.
  • The weights are 0 for every stock until the first rebalancement takes place (for example until April 01, 2007 if rebalanced quarterly).


Is option 2 allowed?


Thank you for a short response!

#2

You can decide to have an initial calibration period. This is fine. Let’s say you need 1 year worth of data, then only start your allocations from 2008. This is fine. In the test data we use, we will use prior data as needed by your payout so that your have allocations across the whole period. E.g. back to the case where you need 1 year worth of data, imagine our test set is N days, we’ll take 1 year prior to these N days so that it works correctly.