Portfolio allocation - when you get different numbers of signals each day

Discussion in 'Automated Trading' started by rmiller3, Jul 7, 2009.

  1. rmiller3


    Many daytrading systems I look at will generate different numbers of signals each day, from 1 to 3 to 20. How do you decide how much to allocate to each trade, when you don't know how many trades there will be that day? I have found two approaches so far. Either put all of your money into the first trade that comes along, or choose to allocate some fixed fraction (say 10%) of your portfolio to each opportunity that comes along. Both are sub-optimal. The first approach (all into the first trade that comes along) leaves you with more risk than you would have if you diversified on days when you have multiple signals. The second (10% into each that comes along) leaves capital underutilized on days when you have only one signal, greatly lowering potential return, though also lowering risk.

    What are your thoughts?

  2. drcha


    My thought would be--a little of both.

    It may be that the earliest opportunities are the best ones. I allocate among the first ten or so opportunities that arise, and hope that the profits generated will make up for the unallocated capital that often results from such a system.

    Never hurts to have dry powder.
  3. This is covered in pretty decent detail by Ganchrow on the SBR forum. I believe he also posted a spreadsheet for the calculations.

    I don't know if I can post a link to a competing forum here, but check out the thread "Another Kelly Conundrum" in the ThinkTank subforum on SBR.