Auction Market Value Theory & Analytics

Unfortunately, I do not……It's probably not an answer you want to hear but I also build profiles & create references points using excel. I used to manually by hand enter the data into a spreadsheet, as my programming skills improved I began to automate the process. I am learning now, to use oandas API where I can call the data into python/excel and build profiles along with amvt references, automatically.

I do this so my "analysis" is not dependent on any one charting package or any particular indicator.
https://www.forexfactory.com/showthr...25#post9341025
Sorry I could not be of any help……

Thanks for your reply. btw, love this thread & info you share. especially cisco research. somehow cisco site is down. its good that you have archived the research in this thread.

Edit: GLScharts.com provides data for Indian Markets!!!

My little contribution to AMVTA team.
As you know there are many files in the MP multi composite profiles in mt4 file section. There are 18 files for each pair as you can see in the left and middle image I attached. I pull out information I want and organize my csv files in 1,3,5,10,15,20 days for each pair I want. I will once program the software pull out information at the push of a button every day and organize my csv files for my data base as seen in the far right. The software is free and search for "advanced renamer" or any other similar software. Enjoy.
JP

Hi is somebody able to program the net flow of all pairs with one click solution (from the extracted data)? I would like to have a table with all the netflows together on one page... Or can someone advise me a program to do that? I think python can do the job but is a lot of work. Maybee there is a better solution.

thanks for helping me out
C

If you already have tables for each of your pairs, you can create one workbook/sheet that has only the "cells" you want to see from other workbooks. Each time you open the workbook it will update itself by the reading cells from other workbooks without having to open them.

here's an example of data from 3 different files on 1 sheet
open the workbook you want data from and the workbook you want to show that data in
select a cell in a blank workbook/sheet type "=" then click the cell you want to display. You can then drag the formula to auto fill the rest

Chris,
Are you looking a kind like this?

Thank you for the information. JP, which program are you using for calc and reading? Would you share the table?
Mzvega, thank you for sharing. So I have to import the data into excel workbooks for reading. Which program are you using for the net flow calc? Excel?

greetings
C

Python/Excel

Chris,
You can search and Google for "data extracting and data transformation application" Now a days you may not need to depend on traditional spread sheet and relational data base if you will, as there are plenty of data transformation applications out there. You can import or read, extract, transform and manipulate your data as you want it to be. Extract, Transform, and Load, ETL Software can do more than what others can do. You may spend sometime reading and watching tutorials though. Since I am using proprietary application, for the sake of integrity of the forum I refrain from posting links. You don't need proprietary software though, many free and powerful applications are out there if you incline to come out from spread sheet and relational databases as alternative. To organize your csv you can follow what I had posted previously at
https://www.forexfactory.com/showthr...0#post10380180
Does this help? I hope.,
JP

If you do not already have knowledge of the C language (and its variants) the way to follow is the one indicated for a long time ago by mzvega: python / excel / db
For the db part you will also have to study the SQL language.
You have to count many hours of studies and tests to get what you want. (months ...)
And if things in trading do not work, the positive side is that you will have learned two of the most important programming languages actually on the scene and that will be useful in many other fields.

thank you Mzvega Iam using python and excel too, but dont realy like it (ps: Iam missing your wonderful entry posts...). And thanks JP for the info, I will have a look at this.

greetings
C

---------------------------------------------

@"tail" nothing new... no answer to my question... I thougth you know me already, doing this not the first time.... In fact Iam doing this longer than you...

Sorry dljonesFan, i quoted the wrong message and i'm not able to modify it after
My message was in response to your prior message

Yes, python can do that in one click solution ... and yes is a lot of work but it' s more flexible than any other ETL proprietary solution.
So, if "better" for you means a faster learning curve, python/excel/db is not the better way

Thank you, but Its not about some learning curve...

It's somewhat disappointing that so many people dismiss AMVT as being "complex", and question it's validity, due to what they call it's "complexity". That's further from the truth

The notion that price changes in period A have something to tell traders about price changes in period B is a hypothesis believed by retail traders, that when tested, proves to be false

Most technical analysis works off an underlying, belief that the market is "day to day" serial correlated therefore, you should be able use previous days prices (candlestick) to "predict" the current day's prices. There are many TA strategies, however they all subscribe to and have the same notion that price changes in period A have something to tell traders about price changes in period B.

The foundation of AMVT is built on empirical evidence that there is no "day-to-day serial correlation"…………..

Let me start by going over a series of tests using over a year of the most recent currency data…………

I will post the actual workbooks & results for 16 currency pairs when I am finished posting……..


I'll start with a test on GBPUSD for day to day serial correlation using the closing price (close)

all you have to do is ask the question of the data, "is today's close higher or lower than yesterday's close H or L ? Then all you have to do to test if there is a serial correlation between yesterdays close and the current days close by asking how often does the previous days close (period A) relays info about the current day's close (period b)? How often was the close used as a predictor for the current day's close? you can count the matches(how often than not) ...........


Here is a test for serial correlation using the "close" ……sampling more than a year of data
Using the old TA notion that price (close) in period A had something to tell traders about price changes in period B. You would be correct 242 times in a sample of 509 days(1.39 yrs), However, you would have been incorrect 267 times. Using the old TA notion that price in period A had something to tell traders about price changes in period B, using that methodology/linear thinking You would have been correct 48% of the time, and wrong 52%.


Here is a test for serial correlation using the "open" ……same sample of data
using the notion that price (open) in period A had something to tell traders about price changes in period B. You would be You would be correct 243 times in a sample of 509 days(1.39 yrs), However, you would have been incorrect 266 times. Using the old TA notion that price in period A had something to tell traders about price changes in period B, using that methodology/linear thinking You would have been correct 48% of the time, and wrong 52%. So far, the methodology proves to be incorrect/false more times than not…………


Let's not stop there, let's continue…………..
Here is a test for serial correlation using the "high" ……same sample of data
using the notion that price (high) in period A had something to tell traders about price changes in period B. Here you would be correct 286 times in a sample of 509 days(1.39 yrs), However, you would have been incorrect 223 times. Using the old TA notion that price in period A had something to tell traders about price changes in period B, using that methodology/linear thinking You would have been correct 56% of the time, and wrong 44%. The results are somewhat better, however minus fees & commission you're still about 50/50.


Here is a test for serial correlation using the "Low" ……same sample of data
using the notion that price (low) in period A had something to tell traders about price changes in period B. Here you would be correct 271 times in a sample of 509 days(1.39 yrs), However, you would have been incorrect 238 times. Using the old TA notion that price in period A had something to tell traders about price changes in period B, using that methodology/linear thinking You would have been correct 53% of the time, and wrong 47%. Again, the results are somewhat better, however minus fees & commission you're still about 50/50................


the notion that the O, H, L, or C prices in period A has something to tell traders about changes in O, H, L, or C prices in period B is statistically false

I know I only showed data for GU, however for all currency pairs the results are the same.

Data on closes for all pairs.........


Data on opens for all pairs.........


Data on highs for all pairs......


Data on lows for all pairs......


There is little evidence when tested that price changes in period A have something to tell traders about price changes in period B, there is only evidence to the contrary. That should make someone question the validity of their trading methodology (TA). The data doesn't support the methodology......

The fundamental difference between Auction Market Value Theory and technical analysis is......

A hypothesis is either a suggested explanation for an observable phenomenon, or a reasoned prediction of a possible causal correlation among multiple phenomena. In science, a theory is a tested, well-substantiated, unifying explanation for a set of verified, proven factors. A theory is always backed by evidence; a hypothesis is only a suggested possible outcome, and is testable and falsifiable.
Pasted from <http://www.diffen.com/difference/Hypothesis_vs_Theory>

The notion that price changes in period A have something to tell traders about price changes in period B is a hypothesis believed by retail traders, that when tested, proves to be false

Most technical analysis works off an underlying, belief that the market is "day to day" serial correlated therefore, you should be able use previous days prices (candlestick) to "predict" the current day's prices. There are many TA strategies, however they all subscribe to and have the same notion that price changes in period A have something to tell traders about price changes in period B.

The foundation of AMT is built on empirical evidence that there is no "day-to-day serial correlation".

The evidence that proves TA methodology/linear thinking to be false, is the empirical evidence on which the foundation of Auction Market Value Theory is built…….

This is what people struggle with, They are "conditioned" to think that price changes in period A have something to tell traders about price changes in period B.

You can't approach AMVT with a TA mindset……...

It simply comes down to this………….you either believe in TA methodology (the notion that price changes in period A have something to tell traders about price changes in period B) despite evidence that proves the contrary.

Or you can choose a methodology (AMVT) with a foundation based on evidence you can prove……

That's what is so difficult for people to wrap their minds around…….
No one wants to admit that everything they "thought" they knew about the market needs to be unlearned.

It's much simpler to dismiss something as "being too complex", than to admit & address the flaws in your thought process/methodology…….

In 2007 I had to ask myself what do I do? Do I ignore the evidence that TA is flawed in it's methodology, and jump from thread to thread looking for a TA "strategy" that I know doesn't statistically exist? Or do I invest my time learning skills to do my own analysis? Unfortunately, the internet/FF is full TA indicators traders "want", (an indicator that removes them from them from a logical thought & decision making process), rather than the tools they need. I can choose to do the work myself, or wait for someone else to do it for me………..

Anyway, here are the worksheets & statistics for16 currency pairs…….

Some might consider this evidence of a "Random Walk"……but wait… I am still not finished…...

EDIT: I posted the wrong "opens" file now updated

Dear Mz Vega,
How about "hour to hour" serial correlated? Will they give the same result as your test? I never made such a test, just curious. Thanks in advanced!!

Nope, no serial correlation there.......

11133 hr bars (463.875 days of data)

Days in Sample days of hour bars
11133 hrs 463.875 days
# Correct 5323
# Incorrect 5810

% Correct 48%
% Incorrect 52%

Thanks MzVega, this fact will shake my wired brain neuron!! I must reconfigure it the right way! What i found interesting is the percentage, they always show near 50-50% (in daily and hours).