A large use of the TD:=ExtFml("PowerPivots.TrendDirection",1,2) function is made in this forum and in fact, the use of this confirmation tool improves most systems and is consistent with theory (higher time frame confirmation + follow the trend).
Having had some experience with Omnitrader and the problems with that software, I am wondering whether the use of this function could generate the "changing signal" syndrome, i.e. the current week signal is evaluated daty by day and can change accordingly generating changes in the system signal (in other words: the signal confirmed on a Tuesday if no longer confirmed, and therefore disappears, on, say a Friday. Has someby got the answer to this doubt of mine?
Thanks.

Yes, I would be interested in the answer to this.

It would seem signals at turning points (tops or bottoms) would have to wait for a percentage change in direction for verification (introducing lag). Without this verification, all you have is a trend line change.

If PowerPivots uses the peak and trough functions then signals could disappear if not validated by a following "verified" percentage move.

Maybe one of the forum leaders can shed light on how Major and Minor pivots are produced and when they become "valid".

Thanks.

I Started out with a chart loaded with 2000 bars EOD, attached the expert w trend direction.
I then loaded another chart, but to the date of the long entry,
I continued to load Charts but I added a date to each of them.
My results showed no disappearing signals, no hanging chads, no funny business.

I did this with all types of entry; long, short ,long exit, and short exit

I will tell you that I have done a lot of testing against the power pivot plug in.
My result is that a few of the experts, in very rare case for some, will give false signals. But also know that functions built in the dll are good.

So what I am trying to say here is that henry's formulae on the forum are most certainly accurate and valid but that's beyond the point. Read this and make sure you remember this:

ANY system you copy or purchase from someone should be tested and tested and tested again by you. There is no exception to this rule :twisted:

In the training video for formula 2 I show how to adjust the data loaded to check signals. And henry explained it just above.

Patrick :mrgreen:

---

My Site

Well people, we certainly didn't mean to get anyones blood pressure up. I thought it was a valid question/point regarding the use of the Power Pivots external formulas to determine trend.

After reading the manual, it appears to me that PP+ determines trends by waiting for a validated takeout of a pivot. The direction of the pivot takeout determines whether the trend is up or down.

I haven't tested, but in looking at the signals there does appear to be lag before a trend is confirmed. Therefore, I assume PP+ does validate the pivot is there to stay before returning a value to you in the external formula. Again, I don't use PP+ calls to determine trend, so test to make sure you are getting expected results.

You are correct that end users need to test, test, and test more before risking money on a system in the market. By the same token, formula writers should be held to a higher standard to DOCUMENT their code in these forums, hanging chads or not.

:D

hey minnamor and topcat...... that is an excellent question and concern..... and that was my number one concern when putting the power pivots plugin through the paces.... just looking at the math almost forces you to believe something might change..... any chart with a formula relying on data not yet here, how can it not.....

believing that was the case, my test was the inverse of henry's.... i took days off the chart a few at a time to see if any signals changed..... obviously this works best with a chart having many signals.....

i was more interested in if the money signals changed or moved, le, se, and such..... but i watched every symbol.... i eventualy gave up trying to find any money symbol that changed.... that was with a heck of a lot of looking...... so those formulas, many of which are beyond my abilites, must be accurate and stand on the chart as posted......



the absolutely perfect pivot labelings are a different story.... as mentioned before, with today's math that relies on tommorrow's data how could there not be a lag..... even without considering the formulas, just look at the chart, how can that 100% dead on labeling be....

the pivots , minor low 'r', intermidate low 'i', minor high and so on are labeled a couple days after the fact..... specificaly, if you look at a chart of 'sze' with the powerpivots plugin 't-fade' expert attached, the minor low labeling on 4/27/05 will not appear on the chart till 3 market days later.... the minor high label on 4/22/05 will not appear for 3 market days...

so taking things one at a time here,

1. in a very clinical sense, yes some of the symbols change..... appearing on the chart 3 days after the fact is a change, or is to me.....

2. i could never find any symbol that vanished from the chart going forward or changed in the way of elliot wave truncations...... once their placed they stand.... still looking though ....

3, the money symbols, in my experience, never changed are were posted on the day, with no lag.....

i had written up the steps taken but never got around to finishing it, this has convinced me to go forward.... patrick has said many times, test and retest, we could put with that question and requestion ..... thanks minnamor and topcat for your questions, they help us all......h

hayseed... this thread really has two different issues.

One is the labeling of the pivots (major, minor, intermediate...) and the other is the call to Power Pivots using the external formula call.

Our concern was that the external call in the formula to determine trend direction be CONFIRMED as a true pivot before returning a trend direction. From what I have seen, that appears to be the case.

Topkat, you did not raise anybody's blood pressure and it was a valid question :D

Sorry If it seemed otherwise :)

Patrick

---

My Site

hey topkat..... my experience with power pivots has only been through the plugin itself and the systems provided with it or equis....

your concern with, """Our concern was that the external call in the formula to determine trend direction be CONFIRMED as a true pivot before returning a trend direction'''' , requires some deep thought....

only a small part of which would be , hopefully any signal would be vaild or confirmed before its returned even if that requires a 3 day wait.... and of course a signal today might not be tommorrow, thats with all systems.....


after carefull rereading minnamor's post let me add this...... given that primary trends do change, if we add 'highlights' to the label the primary trend in the expert editor, yes those hightlights will change when the primary trend changes..... so whats red today might be green tommorrow.... all bars will change color back to the prior trend change..... the labelings of pivots does not change.... only highlights.....


as an example, use 'aa' and highlight the primary trends.... using the downloader remove the price data back to 5/13/05.... all labelings stand but the highlights change due to primary trend change......and again, thanks for posting your concerns, we all gain.....h

I think we are re-hashing what we already agree on, except for the pivot labels.

I believe the last leg of the pivot label is dynamic until it is validated by whatever method Power Pivots uses to cement it into place (a percentage opposite move, or whatever). Until such time as validation takes place, it could disappear on you if price movement doesn't let it validate.

We just wanted to make sure in the systems posted that the External Formula (ExtFml("PowerPivots....)) calls are using validated pivots for determining trends. According to the manual, that is the case.

Here is the main problem with PowerPivots Plus,
They Are Password Protected the new user hasn't experemented with the code and is just using the canned systems alone.

The highlights of the pivots will change as they confirm up or down

If you are referring to me, I have played with the external formula calls to PP+.

We're just trying to get a simple answer to our question???

It's good to trade with the trend, but in addition to trading with the trend, wouldn't you like to trade stocks outperforming the market as a whole?

This expert is based on the idea of going against the prevailing trend and taking trades when prices cross pivots.
It uses the most sensitive Trend Direction setting 1,1 "minor pivots on a daily time frame" and thus it's fault.

If I'm not mistaken, there is no external formula function. Let me know if I am wrong.

topKat, The Contrarian system uses both Nth Pivot Price and Trend Direction external functions

Thanks, I'll try combining those two to see how it works, though I don't think it's something I would want to trade.

On TrendDirection formulas that only check for 1 PivotStrength and 1 TFactor (i.e. "PowerPivots.TrendDirection",1,1) you are in essence getting a lot of Contrarian plays.

There is no check for a higher level trend, thus the higher level trend(s) as defined by PP+ could be going with or against you.