The Motley Fool Discussion Boards
Investing/Strategies / Mechanical Investing
|Subject: Re: criteria to sort ranks||Date: 3/7/2013 7:57 AM|
|Author: mungofitch||Number: 242133 of 247039|
Personally, when testing screens with different liquidity filters, the
only criterion I look at is the average daily dollar volume in the last 3 months.
(not price, not share volume, not separate tests on those two, not market cap)
When I use a small minimum liquidity number I do my backtest with a high friction assumption,
and when I have strict liquidity test I use a smaller friction allowance.
Run the test with several different combos and you can see where the best tradeoff might be.
e.g., if you require a million dollars of daily dollar volume your
trading friction might be only .4% round trip, but if you are allowing
stocks of $75k daily dollar volume you might want to simulate with 1.5% or even 2%.
Run both of those tests and see which one gives the better returns after friction.
Of course, just because you're allowing very illiquid stocks to be
eligible doesn't mean they are getting picked all that often.
Use extra high friction numbers for the not-too-strict liquidity tests if a large
fraction of the picks are actually small cap, which you get an idea of
by seeing how much a screen changes without friction using different minimum liquidity cutoffs.
Big difference = lots of small caps, so the average friction cost is going to be higher.
I much prefer to have the liquidity filter in the screen definition and
then test to see how the whole screen works (with appropriate friction)
rather than trying to skip over individual low-liquidity picks from the
output of a screen definition that doesn't have a meaningful liquidity test.
|Copyright 1996-2013 trademark and the "Fool" logo is a trademark of The Motley Fool, Inc. Contact Us|