What Are The Best Settings For Technical Indicators

. I often close say short position even if MACCI of 1, 3 ,5 min are 70 ,90 95

Click to expand...

Trader333 said:

Grey1,

A good post, Well done and Thanks and as usual I have a question. You say:



My question is why ?


Paul

Click to expand...

Grey1 said:

This post is the reply to Glen's question about what settings .


We are technicians which means we beleive ,,

1) CYCLES DO EXIST in stock market
2) Cycles repeat them selves

This is the heart of TA.

The cycle in market is measured by calculating the distance between two LOW, LOW or HIGH HIGH . ( Now go and look for few cycles in 1 min time frame ).. can you see any ? well they are not all that clear cut are they? if they were then every body would BUY LOW SELL HIGH .

Two ways of measuring the CYLES

1) mathematical
2) empirical

Mathematicians use various techniques to measure these cycles, Guys like Ehler have worked hard to be accurate in finding the beginning and end of the cycle.. donot bother too much as he is still working on it lol

Others use guesstimate and they call it heuristics , or rule of thumbs , or empirical to measure the cycle

Once you have the width of the cycle then you can use indicators such as RSI or CCI to find out how far you are from the top or bottom of the cycle .

The problem is empirical values of the cycles are nothing except guestimates . A guy called Lambert who designed CCI said he thinks market is going through cyle every 60 Time frames ( depending on what time frames you are using .) and then he uses 1/3 of this figure to design his famous CCI . so he recommends 20 time frame for his oscillator . i.e 20


One could argue market is seasonal and 60 time frames means nothing . I would say , I agree
Once could argue any number could be used until it proved wrong I would say , I agree

Therefore there is not a lot out there where we could technically justify these settings .

Now
What next then ?

Well.,, I recommend you should start with the CCI( 6) smoothed by 5 and up date these settings every few month to cater for seasonal changes

Is there any way we could have a better approach ?

Yes .. if you apply CCI in different time frames technically you are chasing the NODE ( one of the highs or lows ) and eventually one of the time frames will hit the node and you have a better estimate of the cycle width .

Ideally you need to have as many time frames as possible like CCI of 1,2,3,4,5,6, time frames ……… to give you a better estimate of cycle width

Now, what should I use in my trading grey1 ?

You can use what I have sent you which is CCI of 6 smoothed 5 time but if you used any different figure you are still not wrong as long as you smooth it accordingly .

Remember , none of MACCI reading are exact and they are at best guestimates , however if they used in different time frames they become much more reliable and technically acceptable ..
You can also force fit the CCI reading on actual chart to see if there is an immediate reversal when MACCI approaches 100, -100 and if it did not adjust your settings to get a perfect match

The bottom line is DONOT WORRY ABOUT THESE SETTINGS. WHEN ENGINE GETS NEAR 100 ish you know where you stand with the market .. I often close say long position even if MACCI of 1, 3 ,5 min are 70 ,90 95


Grey1

Click to expand...

lwebb said:

Grey1

Another insightful post. Quick question about the smoothing number of 5 (against the ma of 6). Is there a relationship between the two lengths. If you use a ma length of say, 12, what should the smoothing length be?

I've read elsewhere of people advocating identifying the cycle length and then using half or a third of that length for indicators. Also of stepping up timeframes by factors of 2, 3, etc. Is there any validity to this?

Thanks.

Click to expand...

lwebb said:

I've read elsewhere of people advocating identifying the cycle length and then using half or a third of that length for indicators. Also of stepping up time frames by factors of 2, 3, etc. Is there any validity to this?

Thanks.

Click to expand...

This is correct. Lambert recommends this too .

Click to expand...

lwebb said:

Pardon my ignorance here but who is Lambert. Do you recommend any authors or books on TA for background reading?

Click to expand...

Grey1 said:

Lambert it the designer of original CCI

grey 1

Click to expand...

millsy500 said:

Heres a simple system i'm going to be trying out on the dow over the next couple of months. . CCI setting 48 trading off 15 min chart. When CCI passes back down through 100 sell. Up through -100 buy.

Click to expand...

millsy500 said:

Hi Nastrader

Thanks for that explaination and chart. Do you use this method? Do you change the settings for different TF's?

Millsy

Click to expand...

Grey1 said:

Two ways of measuring the CYLES
1) mathematical
2) empirical
Mathematicians use various techniques to measure these cycles, Guys like Ehler have worked hard to be accurate in finding the beginning and end of the cycle.. donot bother too much as he is still working on it lol

Others use guesstimate and they call it heuristics , or rule of thumbs , or empirical to measure the cycle

Once you have the width of the cycle then you can use indicators such as RSI or CCI to find out how far you are from the top or bottom of the cycle .

The problem is empirical values of the cycles are nothing except guestimates . A guy called Lambert who designed CCI said he thinks market is going through cyle every 60 Time frames ( depending on what time frames you are using .) and then he uses 1/3 of this figure to design his famous CCI . so he recommends 20 time frame for his oscillator . i.e 20

One could argue market is seasonal and 60 time frames means nothing . I would say , I agree
Once could argue any number could be used until it proved wrong I would say , I agree

Therefore there is not a lot out there where we could technically justify these settings .

Now What next then ?

Well.,, I recommend you should start with the CCI( 6) smoothed by 5 and up date these settings every few month to cater for seasonal changes

Is there any way we could have a better approach ?

Yes .. if you apply CCI in different time frames technically you are chasing the NODE ( one of the highs or lows ) and eventually one of the time frames will hit the node and you have a better estimate of the cycle width .

Ideally you need to have as many time frames as possible like CCI of 1,2,3,4,5,6, time frames ……… to give you a better estimate of cycle width

Now, what should I use in my trading grey1 ?

You can use what I have sent you which is CCI of 6 smoothed 5 time but if you used any different figure you are still not wrong as long as you smooth it accordingly .

Remember , none of MACCI reading are exact and they are at best guestimates , however if they used in different time frames they become much more reliable and technically acceptable ..

The bottom line is DONOT WORRY ABOUT THESE SETTINGS. WHEN ENGINE GETS NEAR 100 ish you know where you stand with the market .. I often close say long position even if MACCI of 1, 3 ,5 min are 70 ,90 95
Grey1

Click to expand...

Nastrader said:

Hi Grey1,

I hope I'm not beating a dead horse - LOL.
But I would like to continue discussions on using different MACCI values, since I did some anlysis and would like to share my findings with everyone.

This analysis shows the Peaks and Valleys in Multi-TF’s (MTF), and the use of different CCI & smoothing MA values. From my observations, you are very correct in saying that using a MA to smooth the CCI is very important, and can be seen clearly in the MTF charts attached to visually see these findings for today (1/12/07).

The different MACCI values used in the attached charts are as follows:

Ø CCI-6, 5MA, color coded White (Only used in 5 & 15-Min Charts)
Ø CCI-12, 5MA, color coded Yellow
Ø CCI-14, 6MA, color coded Green
Ø CCI-20, 6MA, color coded Lt Blue (NDX) & Purple (DOW)

Therefore, the total MACCI’s at the bottom of the 1, 3, 10, & 30-Min Charts are 5, while the 5 & 15-Min Charts have 6 (yes, you can see them all). The MTF’s charts used are 1, 3, 5, 10, 15, & 30-Min, which are all shown on the attached charts.

The findings:
1) The interesting thing was that all of the MACCI’s with different values have their Peaks & Valleys occurring approximately at the same time in matching TF’s (as listed below). The oscillations in their respective TF’s, all cycled along side each other, this finding was a surprise also. The main differences found are the Peak & Valley amplitudes are different within each different MACCI value.
2) The matching Peaks & Valleys are as follows (all times are in ET):
a. Peak at 9:52/10:00 = 5, 3, & 1-Min - Strong matches.
b. Valley at 10:50 = 15, 10, 5, 3, & 1-Min – Strong matches.
c. Peak at 12:50 = 15, 10, 5, 3, & 1-Min – Strong matches.
d. The 30-Min had a strong Peak at 13:30 where all of the different MACCI values met, but had no matches.

To summarize the Market direction for today, it can be best seen on the attached 5-Min chart, where Lt Blue Trend Lines were drawn along side of the DOW (in the price window). The market opened with the usual whipsaws and formed a Peak at 10 AM, then fell to form a Valley at 10:50, then pushed up and stayed O/B the reset of the day, trending. You’ll notice in the 10-Min Chart I used Lt blue horizontal lines to indicate the HOD & LOD at 10:00. This helps to give a visual aid to see the if market is either going below, above or trending through the 10am HOD/LOD lines as it did all 3 today.

Also there are Lt Blue lines drawn vertically at the best matching Peaks & Valley on all of the Charts as well.

For the time being, I plan to use the CCI-12, 5MA and CCI-20, 6MA as my two MACCI’s in my MTF charts.

Would like other traders to share their findings or their use of MACCI values, so that we may all benefit and find the most appropriate one. Also, someone may have found the seasonal MACCI values referred to by Grey1, which would be interesting to be aware of and use.

Cheers,

NasTrader

Click to expand...

Glenn said:

NasTrader
The use of CCi as the underlying indicator means that the MACCI is already harmonised for volty (different velocity stocks) because it uses the mean deviation over the selected number of periods.

As regards seasonality, I assume that this means varying cycle times. As to whether this refers to variation throughout a year, a month, a week etc I don't think matters.
What seems to me to be important is to capture the cycle period of the moment and to employ it as a variable.

One way to adjust for this would be to use the Homodyne Discriminator to find the underlying cycle period and to use 1/3 of this period in the CCI Calculation. This would produce a CCI which is dynamically adjusted for 'seasonality'.
However the Homodyne takes time to adjust to a change in cycle period, such that in a simple case of moving from a cycle period of 40 to say 20 bars could take around 40 bars adjust.
To refine this to some extent it would be necessary to use something like the rate of change of the Homodyne as an indication of whether the cycle period is increasing or decreasing, and as a variable for estimating the new cycle period in advance.

Similarly, although Ehler points out that the Dual Differentiator exhibits some overshoot error, it does nonethless respond more quickly than the Homodyne and might in the overall scheme of things prove to be better.
Whether either of these would produce a useful improvement to the MACCI I can't say because I haven't tried them.
And a lot depends on whether the MACCI is used purely for indicating exhaustion or in addition for identifying turning points.
In your chart examples you recognise that the Market is bullish and so I assume that you are looking for the MACCI to work in such conditions.
Turning point identification need to be accurate to minimise risk of stop-out.
When the market is trending, the MACCI or it's deriviatives will be attempting to find the new, much longer cycle period, but limited in its adaptation by the number of CCI periods in use at the time, and this will continue into the time when the market goes back into oscillation, meaning that the MACCI is corrupted during the beginnings of the oscillation.
The same applies to when the market moves from oscillation to trending.
This is why (I guess) Grey1 recommends the MACCI for oscillating markets only, because they last longer than trends and the MACCI has time to settle down.
imho.
Glenn

Click to expand...

Nastrader said:

Hi Glenn,

Very informative description on cycles, and a bit over my head when discussing the “Homodyne Discriminator.” I’ve reviewed some of Ehlers theories and his presentations from his web site, but these were too complicated for me to become involved with for my trading purposes. But please don’t misunderstand me; I really do appreciate your descriptive reply. I like to keep my strategies a bit simpler, and have become somewhat comfortable with Grey1’s use of MACCI’s as an oscillator, although I need to gain more experience in their use. Your explanations on seasonality of cycles have made this topic a bit clearer.

My purpose in my previous post was to investigate various different CCI periods and MA’s (MACCI’s) when applied to cycles. The differences, which I did find were not alarming, since the peak/valley cycles for all of the different MACCI’s were relatively the same, and thus made me, feel more comfortable.

My main purpose in using MACCI’s is to find the turns & exhaustion, and not ID Trends, since the MACCI’s provide invalid turns during trends. The invalid turns during trends, is one condition I’m trying to find a better way to ID. Yes, and I have the same understanding that Grey1 considers the use of MACCI’s for oscillating markets only (Grey1 calls it Strategy # 3). Grey1 has also stated that when the Dow is > + / - 50 points, then the market is in a trending mode, which has held to be true. But to me there needs to be more verification when a turn is valid or the market is just oscillating in a strong trend – I have found that this concern only occurs when the market is O/B, not O/S recently (but this will change at some point). Currently, I am using codes to ID when the Market breaks thru lower MA’s in MTF’s to coincide with the MACCI turn & then decide (manually) if the turns are valid or not. For the time being they have kept me from entering during these false turns. But the negative issue I have with this is that I lose valuable time in deciding if I have a true oscillation or trend. I’ve also noticed while the Dow was breaking to new highs, the lagging NASDAQ is now breaking out to new highs creating divergences in oscillations between the two. So I am monitoring both indexes to keep me aware.

I do have a question though, since you are well versed on this subject, do you use MACCI’s as an oscillator? And if so, which MACCI values do you use?

Thanks for sharing your informative knowledge on cycles,

NasTrader

Click to expand...