Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Discussion (https://www.forexfactory.com/forumdisplay.php?f=11)
-   -   Auction Market Value Theory & Analytics (https://www.forexfactory.com/showthread.php?t=482744)

stain Apr 21, 2017 12:00pm | Post# 1101

{quote} And how do you "define" the longer term "condition"????
Not sure i got your question
You mean if the market is trending or braketing ?

mzvega Apr 21, 2017 1:13pm | Post# 1102

{quote} Not sure i got your question You mean if the market is trending or braketing ?
Sure, how do you define whether or not the market is trending or bracketing?

"How" isn't as important as "WHY"........
My original question remains.......
What is the purpose of defining the long term condition of the market?

stain Apr 21, 2017 1:23pm | Post# 1103

{quote} Sure, how do you define whether or not the market is trending or bracketing? "How" isn't as important as "WHY"........ My original question remains....... What is the purpose of defining the long term condition of the market?
By observing the number of distribution, one unbroken distribution = braketing. Of course if the longer term dist (let's say the 20d) is braketing, it doesen't mean that the smaller one's (let's say the 3 and the 5D) are doing the same. Only observing the market condition (bracketing /trending) in relation to
1)a specific set of overaly days (3,5,10,15, etc)
2) the behave of the ref. points
we could get a better picture of price behaviour, aka if the 3D s trending into his braket range[assuming is braketing], or there are some early sign of a potential breack (first toward the limit of his range then maybe to the next) etc etc.

mzvega Apr 21, 2017 2:03pm | Post# 1104

3 Attachment(s)
{quote} By observing the number of distribution, one unbroken distribution = braketing. Of course if the longer term dist (let's say the 20d) is braketing, it doesen't mean that the smaller one's (let's say the 3 and the 5D) are doing the same. Only observing the market condition (bracketing /trending) in relation to 1)a specific set of overaly days (3,5,10,15, etc) 2) the behave of the ref. points we could get a better picture of price behaviour, aka if the 3D s trending into his braket range[assuming is braketing], or there are some early sign...
Only observing the market condition (bracketing /trending) in relation to 1)a specific set of overaly days (3,5,10,15, etc) 2) the behave of the ref. points we could get a better picture of price behaviour
We could care less about "price behavior", we are concerned with changes in "Value" not price.

You're trying to give examples of "how" overlays are use.....my question is "why". What is the purpose of using overlays?

#1 Is this overlay bracketing or trending? If it is..... were are the bracket limits?
Name:  1.JPG
Views: 1382
Size:  26 KB

#2 Is this overlay bracketing or trending? If it is..... were are the bracket limits?
Name:  2.JPG
Views: 1360
Size:  28 KB

#3 Is this overlay bracketing or trending? If it is..... were are the bracket limits?
Name:  3.JPG
Views: 1369
Size:  22 KB

Bjorn000 Apr 21, 2017 2:17pm | Post# 1105

3 Attachment(s)
The "Why"

Overlays of multiple timeframes are used to find the market condition.

Only when the market condition is known can an informed trading decision be made.


Finding UL and LL.
Using three TPO's as UL and LL cutoff - This is because i'm assuming this to be a Overlay from a timeframe less than 10 days. If using a timeframe of 10 days or above, then the cutoff should be five TPO's.

Overlay #1 - Bracketing
Click to Enlarge

Name: #1_Bracket.PNG
Size: 162 KB

Overlay #2 - Bracketing
Click to Enlarge

Name: #2_Bracket.PNG
Size: 182 KB

Overlay #3 - Bracketing
Name:  #3_Bracket.PNG
Views: 1253
Size:  135 KB

mzvega Apr 21, 2017 2:36pm | Post# 1106

The "Why" Overlays of multiple timeframes are used to find the market condition. Only when the market condition is known can an informed trading decision be made. Finding UL and LL. Using three TPO's as UL and LL cutoff - This is because i'm assuming this to be a Overlay from a timeframe less than 10 days. If using a timeframe of 10 days or above, then the cutoff should be five TPO's. Overlay #1 - Bracketing {image} Overlay #2 - Bracketing {image} Overlay #3 - Bracketing {image}
The "Why" Overlays of multiple timeframes are used to find the market condition. Only when the market condition is known can an informed trading decision be made. Finding UL and LL. Using three TPO's as UL and LL cutoff - This is because i'm assuming this to be a Overlay from a timeframe less than 10 days. If using a timeframe of 10 days or above, then the cutoff should be five TPO's.


Overlay #1 - Bracketing
correct
Overlay #2 - Bracketing
Not correct....
Overlay #3 - Bracketing
Not correct.......
try again........see if you understand why #2 & #3 are not correct

stain Apr 21, 2017 2:48pm | Post# 1107

1 Attachment(s)
Name:  brack2.JPG
Views: 1267
Size:  40 KB

2 distribution

stain Apr 21, 2017 2:51pm | Post# 1108

1 Attachment(s)
Name:  3brack.JPG
Views: 1268
Size:  30 KB

1 dist but close outside the limite, warning for range expansion or new trend

Bjorn000 Apr 21, 2017 3:23pm | Post# 1109

I forgot about the "Close Rule".

The Close should be Lower than UL and Higher than LL in a bracketing market.

As Stain also said.

mzvega Apr 21, 2017 6:26pm | Post# 1110

I forgot about the "Close Rule". The Close should be Lower than UL and Higher than LL in a bracketing market. .
#3 Would be defined as either "testing for new trend" or "testing for end of trend".
There is not enough info by looking at the chart to determine which one is true. Only by learning how to organize and read your "data" (references) can you differentiate whether the current overlay is "testing for new trend" or "testing for end of trend".

But the proper description for #3 is "testing". "testing for new trend" or "testing for end of trend" requires the ability to read the "data" (analysis)

davethetrade Apr 25, 2017 10:45am | Post# 1111

1 Attachment(s)
Hey its me again. Can someone please help me with this question. And please please dont tell me im missing
the whole picture and that this is just the **** of the tree or what someone told me. The overlay quant analysis
in here is just one way to trade, its not the only. Im not looking to copy Mzvega 100% like most people seem to do.

I have my own way of trading and try to pick the raisin of the cake which i would recomend others to do too, you can
never understand the strategi or spreadsheet as well as the person who made it. Just my 2 cents. Okey anyway.

Click to Enlarge

Name: AD realation.jpg
Size: 149 KB

In this example we dont have any attemped direction but we can see that value is lower. What would you asume out of this
if volume was higher lower or unchanged?

clearly only 1 distribution (more or less) which significe balance
but please tell me if you know this i would be glad
for all inputs, thanks
Dave.

mzvega Apr 25, 2017 11:07am | Post# 1112

Hey its me again. Can someone please help me with this question. And please please dont tell me im missing the whole picture and that this is just the **** of the tree or what someone told me. The overlay quant analysis in here is just one way to trade, its not the only. Im not looking to copy Mzvega 100% like most people seem to do. I have my own way of trading and try to pick the raisin of the cake which i would recomend others to do too, you can never understand the strategi or spreadsheet as well as the person who made it. Just my 2 cents. Okey...
Can you tell me what the huge red letters in the post says?
https://www.forexfactory.com/showthread.php?t=482744

mzvega Apr 25, 2017 11:30am | Post# 1113

And please please dont tell me im missing the whole picture and that this is just the **** of the tree or what someone told me. The overlay quant analysis in here is just one way to trade, its not the only. ................
However, it is the topic of the thread

I have my own way of trading and try to pick the raisin of the cake........
Then that makes you the best person qualified to answer your own question
May I suggest you direct your MP related questions, to a MP thread..

davethetrade Apr 25, 2017 11:39am | Post# 1114

{quote} However, it is the topic of the thread {quote} Then that makes you the best person qualified to answer your own question May I suggest you direct your MP related questions, to a MP thread..
LOL, quantitative methods is not really my area but however i think that even a quant analyst would have benefit of understanding these things, a lot of people in here dont understand that this is one of the most important things in AMT (which is sort of the name of the thread). So i actually dont think its unnecesary at all. But sure i go to another thread.

Cheers.

stain Apr 27, 2017 11:45am | Post# 1115

1 Attachment(s)
Slowly but steadily crafting my instruments

MzVega, if i may ask, which broker do y actually use for your data ?

Click to Enlarge

Name: EXCEL.png
Size: 97 KB

thetail Apr 27, 2017 12:35pm | Post# 1116

Slowly but steadily crafting my instruments
Happy to find new motivated contributors here.

mzvega Apr 27, 2017 4:30pm | Post# 1117

Slowly but steadily crafting my instruments MzVega, if i may ask, which broker do y actually use for your data ? {image}
I now have accounts at forex.com and oanda
My EOD program still "automatically" retrieves data from fxcm
But I can manually tell it to retrieve from either forex.com and/or oanda. There is little or no significant difference in the end result.

thetail Apr 28, 2017 6:28am | Post# 1118

Slowly but steadily crafting my instruments..
Forgot to tell you, since you are at the beginning of the database building, I suggest you export the 2d overlay too, you can filter out if you do not need it, but at least it is present in the database and you can find out more relations in case they exist

stain Apr 28, 2017 11:26am | Post# 1119

{quote} I now have accounts at forex.com and oanda My EOD program still "automatically" retrieves data from fxcm But I can manually tell it to retrieve from either forex.com and/or oanda. There is little or no significant difference in the end result.
Thanks, as an old trader, quality and reliability of the data are concerning me. For instance, i have traded metals for so many years (futures), and use AMVT to trade the xauusd, or the xagusd is quite impossile due the funny playgames that liquidity provider, with the consensus of brokers, play at the close of the spot, or before some important news. The close is often messed up and so the other ref point. (use futures data for reference) Maybe erasing the "funny games" from mt4 history could works ... but is not somethingh i wonna/can focus at the moment.

{quote} Forgot to tell you, since you are at the beginning of the database building, I suggest you export the 2d overlay too, you can filter out if you do not need it, but at least it is present in the database and you can find out more relations in case they exist
Thanks for the advice, but i am not using a database at the moment. I am using a dirty and cheap excel macro to get the job done and nothing else. Probably starting since the beginning with a db is, as per mzvega suggestion, the right choice but .. i can't force myself to learn some programming language at the moment, with 2 jobs [a) run my business b) trading] and a family to run, time is a precious commodity that i have to invest wisely, and i prefer invest my time learning/understanding AMVT instead then focusing on building the instrument to use AMVT. So at the moment, excel is enough, actually i have already some problems with excel since i have never used it and my math skill are belowe the 0, so is already more than enough.

mzvega May 3, 2017 6:13am | Post# 1120

6 Attachment(s)
Serial Correlation........
The lack of day-to-day serial correlation in auction markets is often the starting point for random market theory. Studies as early as Labys and Granger (1970) confirm the day-to-day randomness. Even in trending markets, the expectation of up price or down price for the next day is still about 50-50. Even odds in non-trending markets is easy to understand; in trending ones it is not so obvious. A market that is trending up or down does have an enhanced probability in the trend direction; it is just pretty well masked by the random component that is always present. For example, a trend of one cent per day (on average) is virtually lost in the average daily trading range of, say, ten cents. By grouping, however, the random component can be averaged out and the underlying price movement can be seen. We will illustrate this point with a trend in soybeans.

Comparing one day's data to the previous days data (or comparing one candlestick to the next candlestick, or comparing one price to the next price) following the market in this way, are all examples of "linear measures".

Now if you follow the data, linearly, "day to day" in sequential/chronological order, the market appears to move randomly. There is no "day to day" evidence of a underlying trend.

This is where the random theory conspiracist's see this as some type of evidence that the markets are a random walk. When you look at the data in a linear, chronological , sequential, series (time series) it only appears to be random What this proves is that even in a trend there is no "day to day" serial correlation. Even in trending markets, the expectation of up price or down price for the next day is still about 50-50
Click to Enlarge

Name: example 1.jpg
Size: 145 KB

Rather than "linearly" follow the data price to price. Lets start measuring the same exact data, using "non linear" measures used in auction market analysis. Here we will use a 2 day sample of data. Rather than following the data "day to day" in sequential order. The results look just as random as following the market day to day. There is still no evidence of the underlying downtrend Even with a 2 day sample of data.
Click to Enlarge

Name: example 2.jpg
Size: 114 KB

Let's look at a 3 day sample of data......... Now things start to get interesting..Same exact data.. still using "non linear" measures.Now you start seeing some "evidence" of the underlying downtrend. By using a 3 day sample of data you began to filter out the "day to day" random fluctuations ("random component")
Click to Enlarge

Name: example 3.jpg
Size: 124 KB

It also becomes more apparent with the 4 & 5 day sample of data.
Click to Enlarge

Name: example 5.jpg
Size: 228 KB

So what does this prove?.......... do the math
Click to Enlarge

Name: example 6.jpg
Size: 303 KB

Testing a Trend for Serial Correlation
There is one long downward trend from March 6, 1991, through July 10, 1991. This run covered 88 trading days (127 calendar days) or just over four months. The total price drop is 121 cents or 1.375 cents per trading day. The daily trading-range average over the 88 trading days is 8.17 cents, ranging between a high of 29.25 and a low of 3.25 cents. Table 1-5 has the prices and a list of higher/lower prices on the basis of 1 day, 2 days,. . . through 5 days.

Since a downtrend was Pre-selected, we have inserted a bias to the downside.
Therefore ties those cases where the compared prices were the same-will be awarded to the H, or higher count. That results in Table 1-6

One of the basic fundamentals in AMVT is an understanding that the markets are not linear. Using linear measures to describe & analyze a non linear market is like using a tape measure to measure how many gallons it takes to fill a bucket

If you group your data, group prices over the correct sample of time (price over time) rather than look at the data "linearly", day to day through a series of candlesticks, you see the markets are in no way a random walk.

"A market that is trending up or down does have an enhanced probability in the trend direction"

To test that statement in a downtrend using the data starting with the 1 day sample of data.
If you traded in the direction of the previous day's trend (following price) with the probability of trading with the underlying trend, you would have correctly traded in the direction of the underlying trend 52% of the time and you would have been wrong 48% of the time. Correctly trading in the direction of the underlying trend following price "day to day" is still pretty much a coin toss.

Let's look at the 2 day sample.had you traded in the direction of the underlying trend following the 2 day samples of data rather than following price, you would have correctly traded in the direction of the underlying trend 62% of the time and you would have been wrong 38% of the time.

Now things get interesting..

Let's look at the 3 day sample.had you traded in the direction of the underlying trend using the 3 day samples of data rather than following price, You would have correctly traded in the direction of the underlying trend, 70% of the time and wrong 30% of the time.

4 day sample you would have been correct 73% of the time and wrong 27% of the time.

5 day sample you would have been correct 76% of the time and wrong 24% of the time.etc.

This is what you call "proof of principle"...
Name:  example 7.jpg
Views: 1205
Size:  15 KB


© Forex Factory