| nwcoldfront |
|
Gold User, Member, TeleChart
|
| Registered User |
|
|
|
|
| Unsure |
|
| Thursday, October 7, 2004 |
| Saturday, May 8, 2010 9:04:40 PM |
73
[0.02% of all post / 0.01 posts per day] |
|
Diceman,
You have issues.
Happy trading.
|
Don't take things out of context...
Look at the quote:
"So the sum can never be superior for determining crosses or spreads because it simply does not work...by definition. The formula is wrong."
The sum cannot determine crosses or spreads because the spread is the difference between two indicators. It is wrong by definition. When two indicators cross the difference is zero. The sum will never work.
MACD itself is a spread...a difference between two moving averages.
I guess you can't or simply won't read the posts.
This seems to be a reoccuring theme with you.
You just like to argue and stir up the pot.
TM
|
Remember diceman your response was to my "spread" indicator.
It was mentioned to point out the short comings of divergences that go on "forever" before confirming a reversal. If you read my original post I spoke of the false divergence of using MACD vs its moving average (method #3), and how (method #4) did not create this false divergence in relation to its cross over. It confirmed the entire rally from July to the end of Feb.
The discussion was turned to method #5 to point out how wide historical spreads can signal tops or bottoms.
I have seen other people use the "summing" method you speak for moving averages of different time periods. Some people feel it is more sensitive to turns, etc.
There is nothing wrong with it, but it is not a spread.
Your method probably relates to what I called method #2..and it probably works fine.
TM
|
I do realize what I am looking at. The original discussion that I began was about spreads...Cross overs between these spreads (position above or below) and historically wide gaps between these spreads.
Spreads are a difference not a sum.
In the BRCM example you should really use the formula I provided:
((xAVGC2 - xAVGC30) - (xAVGC12 - xAVGC26))/C
The reference was to method #5 (historical extreme)
The period between 7/22-7/26 does mark a historical low for ZOOM level 3, using this indicator.
Again you should take the time to read the posts...
I'm sure your indicator is fine, but it is simply not a spread.
Take care.
TM
|
A spread is the difference between two indicators not the sum.
If you want to know where two indicators cross you need to take the difference.
So the sum can never be superior for determining crosses or spreads because it simply does not work...by definition. The formula is wrong.
However, the sum of the indicators might be useful in some other way. Just not for crosses or spreads.
TM
|
These are the parameters I have used in TC Net for "the spread" as I call it:
((xAVGC2 - xAVGC30) - (xAVGC12 - xAVGC26))/C
1. You need to take the difference between the two MACDs if you want to know when they cross.
2. Then divide it by the close to get a relative % change that can be used to compare different stocks.
3. Above the zero line is bullish/ below is bearish.
4. You will also see that with this indicator you can use method #5 to evaluate reversal points. Wide historic spreads = reversal.
5. Putting a moving average on this indicator works pretty decent too...to anticipate moves.
One strength of TC Net is the Dynamic Scale Adjustment, so you can see when the spread is reached historical values on a particular zoom level. This works well with Visual Scans.
When you divide by the close you can also see a percent spread. So a value of 0.05 for example is a pretty significant spread, that is prime for reversal.
This "wide spread" (method #5) can also signal the end of long trends.
Take this example:
BRCM got in a strong down trend 3/1 to 7/26/07.
This "spread indicator" helped to pick the bottom.
Plot the indicator and look at BRCM at Zoom 3 ending on 7/20/06. This deep into the down trend and was the day before earnings. The next day the stock gapped down and the "spread indicator" hit a new low for that time frame of about -.22 or -22%. This marked the bottom and the end of the down trend.
The longer the zoom range to better for determining historic lows.
Take another example with a zoom of 1
Look at the QQQQ for May 30, 2006 the next day marks a historic low of about -5% for "the spread" and resulted in "the bottom" of the QQQQ to the exact day (by measure of the close).
Its not always this accurate but its also not a bad indicator at all.
TM
|
Yes that can be a problem with the continuing divergences.
Two types of momentum indicators that are nice to run in tandem are two time frames of MACD. You read them the same way you would read "typical" cross overs.
This works out much better than to just use a moving average.
The only problem is that you can't do this in TC because each time you add a new indicator in TC it creates its own custom scale.
To plot the two together you would need to have a common scale.
There are five basic methods to interpret longer term MACD indicators:
#1. In relation to the zero line (above is bullish, below bearish).
#2. Determining basic short term trend of the MACD to determine divergences or convergences with price trend.
#3. Relation of the MACD to a moving average of itself. Above = Bullish, Below = Bearish.
#4. Relation of the MACD to a different time scale of MACD. Short Term Above = Bullish, Short Term Below = Bearish.
#5. The spread between the MACD signal and its moving average (or another MACD time frame). Wide historic spread either way = Reversal.
------------------------------------------
Let us compare method #3 vs. method #4
Here is an example of the difference :
We will reference the rally from July 25 2006-Feb 26 2007
Using a MACD with moving average (method #3).
(The buy signal is that the MACD is above the moving average 5 EMA)
vs.
Two MACDs of different periods (method #4).
(The buy signal is that the shorter term MACD is above the longer term MACD)
1. They both pick up a bullish cross over on July 25 (so far the same).
2. They both stay bullish until November 1.
3. On November 1 the MACD vs. moving average (method #3) crosses negative and stays below the MACD through February 26. This becomes a false persistent negative divergence for method #3.
4. However, the Short Term MACD stays above the Long term MACD all the way until February 26. A strong trend continuation for method #4.
So there is a big difference.
Method #4 is superior from many vantage points.
Note:
*There is one small false negative cross on the MACD to MACD comparison (method #4) from 8/9 to 8/15. However, a 2 or 3 period smoothing would eliminate this.
**No amount of smoothing will eliminate the MACD vs moving average (method #3) errors.
***February 27 marked our “market melt down day” so all indicators pretty much crossed negative on that day. On shorter term time frames such as 60 minute and 120 minute method #4 showed exit signals on Feb 26.
****It should also be noted that the melt down that occurred on Feb 27 began with persistent selling throughout the day. The main draw down came within a 15 minute period around 3:00 PM. Anyone with a hint of discipline should have been stopped out long before the 3:00 hour. Internals that horrible should never have been attempted to buy into.
TM
|
Its amazing how the market keeps showing last minute "manic fits" to stay alive in the face of negative momentum divergences on virtually every front.
This last minute parabolic peak is reminiscent of the China market mania.
It seems appropriate to have almost a vertical spike to cap off this rally to signal a significant reversal.
|
After a year of two brief fake outs down (may 2006, february 2007)..and a climatic rise to these peak levels...it has been frustrating for the bears.
But...could this finally be the turn down???
It has all the markings of a turn that is can collapse sharply lower.
Third turn down's the charm??
I think so...BIG SELL OFF AHEAD.
|
Re-read my posts Diceman.
I will not keep repeating myself.
All of your points have been addressed.
TM
|
|