|
PureBytes Links
Trading Reference Links
|
Among the many things I don't understand, this one has
been bothering me of late:
A while back, I decided to check one of my assumptions
and test the higher-high, higher-low/lower-high, lower-low
definition of trends. The easy way was to create a system
using peak() and trough(). It worked beautifully. Virtually
any contract I ran the system past, it made money. This
I took to confirm the validity of the trend definition.
Then the obvious dawned on me: Why not see whether there
was enough of the move left, on average, to make a buck from it
after the peak or trough was far enough behind us to get the
signal in real time? I wrote another system that included a delay
factor, so that one would enter or exit a trade only when the
price had retraced from the peak or trough by the appropriate
percentage. Again, it worked just fine. In historical testing, it
made money like magic on anything from 5-minute to daily bars.
Problem: When I put it on real-time data, it gave a lot of bad
signals. Then it suddenly recalculated things, decided that the
minor up and down trends of the last few weeks--this was
on smallish intraday bars--had really been a long up trend,
gave a new set of signals, and declared itself a winner.
Does anyone understand these functions well enough to
explain this behavior to me? I knew that peak() and trough()
backdate their results by putting their signal several bars
before it was possible to receive it; that is what I was trying
to correct with the delay factor. Now it seems that they
also recalculate their old percentages by comparing against
the latest data rather than limiting themselves to the data
that available in real time.
No doubt this is a real beginner's mistake (despite having
played with this for years), but it would have seemed
reasonable to assume that a change of X% three weeks ago
should remain X%, even if we looked at it later. This sort
of thing has to be seen within its context, or it's useless.
Is there some reason the functions have to be written this way,
which I'm completely overlooking, or did someone just
butcher this piece of code?
Many thanks.
Owen Davies
|