Position Sizing in my Risk Premium Tests
Posted on Friday, March 30, 2007 in Uncategorized
One of the readers made a valid point that in my original tests, I was just using a straight $10K per position and that could have impacted the results since my total equity available stayed constant even as the market went down.
I reran the tests from yesterday with a starting account size of $100k, using 1% of the account per position. Note that exposure is not 100% since not all the current Nasdaq 100 existed during this test.
I think the results speak for themselvs:
DayTrading:
Overnight:
-John
Friday, March 30 5:10 pm
Very interesting, going to have to run some backtests of my own on this. Would seem to imply that during a bear market more risk is concentrated during the day that there should be.
Friday, March 30 10:27 pm
BriG,
Please let me/us know the results of your backtests and I’ll post/link to them to provide everyone more information on this topic.
- John
Sunday, April 1 6:26 pm
I ran several tests with SPY and the QQQQ over various time periods going back to 2000, and the results match up similarly to what you and Dr Steenbarger found. The effect seems more pronouced in the QQQQ.
Trying to work out why this is, perhaps that the QQQQ shows more of this is a clue. Working conjecture at the moment is that over selling at closes combined with over buying pre-open, cause bulk of the effect. Over eager buyers in the morning combined with stuck daytraders near the close?
Leave a Comment