|
PureBytes Links
Trading Reference Links
|
--- Chuck Rademacher <chuck_rademacher@xxxxxxxxxx> wrote:
> As I mentioned in one of my earlier posts, there aren't enough AB users who
> appreciate the problem and are prepared to do the work to eliminate it.
Conclusion:
Perhaps it is based on my own limited experience, but I am more concerned about
removing bias caused by survivorship (ignoring inactives) and by splits (using
adjusted prices in filters and ranking formulas).
My (untested) hypothesis is that, in most cases, including/excluding dividends may
make a difference, but not enough of one to make a good system look bad (or a bad
system look good). It might, however, affect results enough to lead one to pick a
good system over a better system.
Thus, ideally, one would like to use data that includes dividends. I will explore
what form that data might take in another post.
Detailed Reasoning:
What follows is a bit of thinking out loud to clarify my own understanding. It might
also help others get a handle on the issue. Here is my expanded list of possible
problems that might arise for long (and short) systems using data that ignores
dividends:
1. Entry problems for long systems: if the system buys beaten down stocks, a large
30% to 50% dividend could drop the price enough to get a false buy signal. Probable
result: likely only a loss of opprotunity (could have had money in another stock)
since the stock is no more likely to drop further than it is to rise.
2. Exit problems for long systems: a large dividend and the resulting price drop
would likely trigger an early exit. Probable result: an early exit with a profit,
and perhaps an opportunity cost if the stock goes up after the dividend payout.
3. Exit problems for long systems: a medium dividend (%2-4%) might also trigger stop
loss exits but only if the stock had been doing poorly to begin with (or if the
system uses very tight stops). Probable result: loss or opportunity and the
slippage/commission cost to enter a new trade to replace the stopped out one.
4. Profit calculations for long systems: Ignoring dividends would either not affect
profit calculations (if there were no dividends during the holding period) or would
underestimate profits (because adding the dividends in would increase the gain).
5. Profit calculations for long systems with multi-year holding periods (not likely
many such investors are reading this board) which happen to mainly pick dividend
stocks might so underestimate returns that the system would be discarded when it
might have been kept if the data had included dividends.
6. Short systems might have bigger problems with dividends: Ignoring dividends in
testing could underestimate profits since any price drops related to dividend
payments would be illusionary. Also a short system that enters on downside breakouts
could get, form time to time, a false entry signal.
b
__________________________________
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes
http://hotjobs.sweepstakes.yahoo.com/signingbonus
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Yahoo! Groups Links
To visit your group on the web, go to:
http://groups.yahoo.com/group/amibroker/
To unsubscribe from this group, send an email to:
amibroker-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|